In this research there has been made an effort of bringing light into the shadows of mobile services pricing and formation of price perceptions. At the moment several Finnish and foreign mobile operators are wrestling with problems of pricing. Problematic has been deciding how and what to charge in mobile services, especially as mobile service users are having problems in accepting the prices above the fixed Internet connection. Challenges are also brought by fast emerging new technologies and services. Due to these characteristics of mobile services pricing and business models used in other industries are not always applicable for mobile environment. Also the urge to develop value in mobile services business creates a need for further research for aggregating more detailed knowledge on pricing of mobile services.

The special features that differentiate mobile services from their e-service equivalents concern the ability to locate people through their handheld device; the availability of services regardless of time and space; the immediacy of receiving information and service at hand, etc. The wireless environment has not only special features but also special peculiarities, like wireless network, constraints of mobile terminals, and usability implications. Mobile services bring challenging opportunities for all actors involving in mobile business. Its role can vary from very simple and passive to very active and dynamic by being strategically positioned in the value network that is developed.

To bring more insight into the problems of pricing there is studied how mobile service users perceive the prices, how they assess the prices, and what are they ready to pay for mobility. Especially challenge has noticed to be in high differences between different user segments. Mobile service customers are thus divided into three differentiated segments which price perceptions are studied segment-specifically. And in order to obtain some market specific knowledge, there has been concentrated on studying Finnish mobile service markets.

1.2. Research rationale

Price has been observed as an important element of affecting to diffusion process of new mobile services. Furthermore, price in this study is not seen as an absolute amount of money but rather as a perception of a customer. The perceived price is formed from the bases of a customer’s experience about mobile services and in comparison to prices of other service channels.

Even though it is known that price is an integral part of diffusion enhancement activities, there is very limited knowledge on its actual effects on diffusion of mobile services. It is more or less unknown, how users of mobile services perceive the charged prices and what are the dynamics affecting to perceived prices. Therefore we must acquire better understanding on users themselves; what are they characteristics, how they observe prices, and how the price and mobile service acquisition is connected.

The probability of adopting mobile services by private users may depend on, not only the characteristics of marketing activities, but also the characteristics of consumers who see mobile services as a viable option for fulfilling their service needs. This research focuses on detailing and examining the price related factors and variables that have an impact on mobile service users’ perceptions. The research objective of this study is to define factors affecting to price perception of mobile services customers and bring out the importance of pricing in adoption of new mobile services. There is also given a proactive solution, in a form of bundle pricing, which is expected to positively affect the price perceptions of mobile service customers; that way inducing the diffusion of mobile services.

Users of mobile services are divided into three segments according to amount of usage of mobile services: heavy users, moderate users, and prospective users. The factors measured in this research are expected to have an impact on customers’ price perception: satisfaction to operator’s services, users’ innovativeness, price sensitivity, price transparency, and investment readiness on mobile services.

Even though there are several (co-)operating parties in mobile services business - operator/service providers, technology provider, and content providers perspectives – the concentration in this study is on relationships between service operator and (private) end users. As a service operator is operating on the customer surface it is also responsible for the pricing (and other marketing) activities. From the customer view-point publicly presented prices should be transparent, unambiguous, and thus risk reducing that it would enhance the phase of adoption and create trust amongst potential customers. From these bases there is constructed following research questions.

Research questions of the study are:

RQ1: Which factors affect to price perception of mobile service customers?

RQ2: How does price and pricing methods affect to price perception of mobile service?

RQ3: Can there be influenced price perception through bundle pricing strategy?

As a ground-hypothesis we have expected that it is useful that customers of mobile services can be divided into three segments. The segmentation into three segments was derived from the notion that consumers in electronic services business can be divided into three basic groups according to price elasticity (Kollmann 2000). And as there was also hypothesized that price elasticity affects strongly on perceived prices of mobile services it was useful to apply the findings of Kollmann for segmenting purposes.

Special interest in this study was to examine an operator’s ability to affect to perceived prices of mobile services which is seen as a main factor inducing or preventing the diffusion of mobile services. Moreover, as the area of pricing it-self is extremely wide and multidimensional it is useful and practical to examine the price perceptions of mobile service customers rather than to go for examining absolute prices.

By answering to research questions there can be obtained better insight on how mobile service users perceive prices of mobile services and on what grounds they assess them. As there are only few studies concerning the dynamics behind price perception and its influence on consumers’ readiness to invest on mobile services this study brings important knowledge to both actors of mobile services business and theorists.

1.3. Outline of thesis

Structure of the thesis is divided into four main entities. In chapter 2, there is examined the theoretical backgrounds of the research topics and gone through the findings of the published studies. Chapter two contains also a model formation for empirical study. Chapter 3 contains insight on the research design and research methodology. There has been brought insight into the data collection, methodological choices, and reliability and validity of this research. Thus chapters 1, 2, and 3 are concentrating on creating a basic knowledge for understanding better the purpose of this study and for better interpret the results of the empirical findings.

In chapter 4 there has been examined the findings of this research on price perception of mobile services. There is created factors that have an influence on respondents’ price perceptions and examined variables that are affecting to these variables. In chapter 5 there has been concentrated on examining respondents’ interestedness in acquiring mobile services in bundles and factors explaining the variance in the interestedness. This chapter also contains an examination of a proactive pricing method bundling in order to answer to needs for affecting positively to consumers’ price sensitivity. In this chapter there is also formed mobile service bundles that are most often used or preferred. This knowledge is especially targeted for business actors. And finally in chapter 6 there has been summed the results of the study up and discussed the findings of this research and its outcome from the viewpoint of mobile operators and theories of mobile service pricing.

2. Literature review on pricing of mobile services

Amongst the private providers of telecom services there have been recognised that customers of telecommunication, who are willing to swap providers, will make their decision on price first. Innovation management for telecommunication products/services is thus a question of pricing. (Kollmann 2000, p.8) New services and additional features of mobility and personalization over mobile devices enable also new types of mobile services (Jonason&Eliasson 2001, p.341). Quality of mobile services is of great concern as there are multiple parties operating in connection together in mobile services production. The required co-operation with parties involved to service production and service delivery lays great challenges for service quality and pricing practices.

These new features and development of wireless technologies and business practices have generated a need to develop and invent new methods of pricing. Therefore it is necessary to inspect the consumer perceptions of mobile services prices and satisfaction to services to be able to create more effective pricing schemes. We have taken special emphasize on factors related to price perception of mobile services, how different elements affect to the emergence of low or high price perceptions.

2.1. Background of mobile industry

Wireless carriers have been generally successful in gaining considerable revenue from customers and have, hence, often been very profitable. This is the Achilles’ heel for wireless operators although some operators have gained considerable revenue additions from short message services (Jonason&Eliasson 2001, p.341).

Customers in telecommunication industry have preconceived notions about the price and value of telecommunications services. Customers have historically complained about the level of local charges, more than they have about long distance, although local service is frequently offered at a price lower than actual cost. When long-distance service is priced well over cost, and local service is generally priced well under cost, customers expect to pay very low prices for local services and apparently do not mind that long-distance could be less expensive but is not. Perhaps these customer reactions are based partly on their ability to control the expenditures on long-distance calling because the charges are usage-based. (Strouse 1999, p.187)

The features of mobile business compels, as stated earlier, close co-operation of different parties to produce high quality mobile services. For players interested in launching new applications and services, alliances can present an opportunity to test the market by shifting the rules of the game. However, the difficulty with alliances and partnerships is their maintenance and management over the long term. (Nordström 2001)

In mobile environment the providers of content and applications contribute value to services, but rely on the network operators to charge the end-user (Jonason, 2002). Thus, operators are to be more or less service integrators in m-commerce as seen in figure 1. The content providers bring the richness while the operators bring the reach (Jonason&Eliasson 2001). By providing technical capabilities and environment for the use of service and content providers, the operator plays an important role but is not the value provider perceived by consumers (Mylonopoulos et al. 2002). E.g. DoCoMo’s I-mode is successful Japanese operator which users are charged for surfing Web pagers and sending or receiving e-mail 0.3 yen per packet of 128 bytes of transferred data. The billing of these services is exclusively handled by NTT DoCoMo costing the content provider 9 per cent of the revenue from the end-user while the remaining 91 per cent of the content charge goes to the content provider. (Jonason&Eliasson 2001)

There is also noticed that operators tend to be myopic in their pricing contracts towards the providers and thus reduce their incentives to continue to contribute value. The spill-over in the market, coupled with the pricing base, makes the roles in the market endogenous and subject to dynamics. The operators’ decision of how to price their mobile Internet presence will therefore not only be a tool of competition but also a major determinant of which companies they will be able to regard as partners and which will become competitors. (Jonason 2002, p.191)

The most common arrangement is that operators handle the billing towards the end-users, while the providers add value to the end-users in the form of branded services and applications. It is thus debated, how end-users revenues are to be divided between these third party providers and the operators. (Jonason&Eliasson 2001) In general, one of orientations for creating innovative pricing schemes is the aim that consumers would become so used to the particular features of their products/services that they would find it hard to change to another system. Ross (1984) has suggested in his studies that effective price changes are based on anticipated reactions of customers and competitors, rather than just the firm’s own costs and circumstances. (Finch et al. 1998)

As operators must do close co-operation with other parties, it lays challenges for pricing structure to support the production of high quality output, and the diffusion of usage of mobile services. In e-and software businesses penetration pricing strategies have been regularly used. Argument for using low prices in new services or products is to accelerate the diffusion for achieving the critical mass. Acquiring the critical mass is equally important also in mobile services business. According to observations mobile business tends to start expand fast when achieving the penetration rate of 30 per cent (Poropudas 2000). The critical mass brings along also positive externalities for operators (e.g. extra sales in other services). Laffont et al. (2000, p.63) have stated that the positive externalities are prominent especially in mobile services in telecommunication sector.

The special characteristics of positive externalities in telecommunications business is the declining marginal benefit along with the increase of usage and data volumes. The positive externalities expected from adding more users to the network may be reversed to negative if the marginal user causes congestion. Thus, mobile services cannot be distributed without a charge and moreover this possible congestion should be managed somehow. The optimistic law of Metcalfe stating that the benefit of users grows with the square of the number of other users in the network is therefore only true as long as the marginal user does not cause congestion (Jonason et al. 2001, p.344).

The other special characteristics of mobile services business is the problem of making products chargeable. The most common arrangement has been that the mobile operators handle the billing towards the end-users, while the providers add value to the end-users in the form of branded services and applications. Jonason states that chargeability becomes acute if decision models become subject to uncertainty and continuously change as a result of market dynamics. This situation arises when products are multidimensional and the producer has difficulties determining which dimensions are in demand and which dimensions can be charged for. (Jonason 2002) In such a situation the product or output (Hayek 1954) is not well defined and, hence, neither is price.

More explicitly, the problem of chargeability arises when the owner of the valuable entity and the owner of the chargeable entity are two separate firms and, therefore, need to agree upon a pricing contract. Technically more complex products are faced with this type of pricing problem such as the delivery of content and applications to mobile devices, typically news, and banking and entertainment services. (Jonason, 2002, p.186)

To deliver more accurate about the business practice in mobile service industry, next is presented an example case about the Japanese mobile operator NTT DoCoMo’s business model:

The subscriber’s bill depends on usage of the I-mode function on the phone, and the number of fee-based I-mode content services he/she subscribes to. There is a basic fee of 300 yen per month to access the I-mode service which is paid via the subscriber’s phone bill to NTT DoCoMo. Since I-mode is based on packet-data transmission, users do not pay for the time they are connected to a service. Instead they are charged according to the volume of data transmitted. The most common arrangement is that the operators handle the billing towards the end-users, while the providers add value to the end-users in the form of branded services and applications. It is debated, however, how end-users revenues are to be divided between these third party providers and the operators. (Jonason&Eliasson 2001, p.342)

Using I-mode to surf Web pagers and send or receive e-mail generates a charge of 0.3 yen per packet of 128 bytes of transferred data. The billing of these services is exclusively handled by NTT DoCoMo costing the content provider 9 per cent of the revenue from the end-user while the remaining 91 per cent of the content charge goes to the content provider. (Jonason&Eliasson 2001, p.344)

2.2. Satisfaction to services

Chase and Tansik (1983) have classified services based on consumer contact. The extent of consumer contact with the service organizations was used as a means of differentiating types of services. Three types of services identified are:

(i) Pure service – organizations in which the customer must be present for service production (e.g. fast food restaurant, nursing home, etc.)

(ii) Mixed service – organizations in which there is both face-to-face as well as back office contact with the customer (e.g. commercial airline).

(iii) Quasi-manufacturing service – organizations in which there is no face-toface contact with the customer (e.g. credit card, a long-distance phone company, etc.) (Chen et al. 1994)

According to this classification of service types, mobile services can be located into the latter service category as there is practically no face-to-face customer contact. This notion of no face-to-face contact with customers creates the peculiarity of mobile services and compels researchers to produce more industry specific insight. In this study we have concentrated on price perception and factors affecting to it. Perceived service quality and value factors are esamined in this study in a form of service satisfaction. And service satisfaction is expected to possess a strong influence on emergence of perceived price of mobile services.

Service quality is the central issue facing the service industry. In Western-world basically all countries are called service economies as over half of their gross national product becomes from service sectors. Measured in this scale the U.S. economy is leading as service sector in its economy accounts for approximately three-quarters of the gross national product and nine of the ten jobs the economy creates. (Kotler 1997, p.466)

And when observing services or products, quality is of great importance. As Zeithaml et al. (1990, p. 9) has stated, improving service in the eyes of customers is what pays off. Excellent service pays off because it creates true customers – customers who are glad they selected a firm after the service experience, customers who will use the firm again and sing the firm’s praises to other. When service improvement investments lead to perceived service improvement, quality becomes a profit strategy.” And Kollmann (2000, p.9) has continued this to value aspects related to mobile services industry by noting “that the potential subscriber will arrive at a maximum price at which he/she will be prepared to buy into the telecommunication system, in the sense of “connecting up”. This price represents the value of the telecommunication system, which the potential subscriber hopes to materialise through the purchase and continued use.”

When discussing quality, value, price, etc. of a service it is always dealing with a perception of a customer of the service. Kotler and Armstrong (2001, p.192-193) have processed perception through motivation. A motivated person is ready to act. How the person acts is influenced by his or her perception of the situation. All of us learn by the flow of information through our senses. However, each of us receives, organizes, and interprets this sensory information in an individual way. Perception is the process by which people select, organize, and interpret information to form a meaningful picture of the world.

In addition to features of a service, a strong influencing factor to perception of a customer is price. Especially in case of services price plays an important role in creating expectations towards service quality and functionality. As Nagle and Holden (2002) have noted, generally price represents nothing more than the money a buyer must give to a seller as part of a purchase agreement. For a few products, however, price means much more. Such products fall into three categories: image products, exclusive products, and products without any other cues to their relative quality.

2.2.1. Perceived service quality

“When service improvement investment lead to perceived service improvement, quality becomes a profit strategy” (Zeithaml et al. 1990, p.9). E.g. Chen et al. (1994, p.24) have defined perceived service quality by differentiating it from actual goods quality with two dimensions: (i) it involves a higher level of abstraction rather than a specific attribute of a product; and (ii) a judgement is usually made within a consumer’s evoked set. And Parasuraman et al. (1988) identify five quality dimension that linked specific service characteristics to consumer expectations of quality: (i) tangibles – physical facilities, equipment, and appearance of personnel; (ii) reliability – ability to perform the promised service dependably and accurately; (iii) responsiveness – willingness to help customers and provide prompt service; (iv) assurance – knowledge and courtesy of employees and their ability to convey trust and confidence; and (v) empathy – caring, individualized attention provided to customers.

Among the several factors that influence buying behaviour, two important ones are “price” and “quality of service”. Price is the cost of making the purchase, while quality of service refers to how well a customer is being served, including the extent to which the server helps the customer, the manner of the server, and son on. Usually, a low price, while contributing positively to product/service selection, contributes negatively to quality of services expectations (Olson, 1977). Hence, a trade-off exists between price and service quality. (Tse, A. 2001)

One of the major ways to differentiate a service firm is to deliver consistently higher-quality service than competitors. The key is to meet or exceed the target customers’ service-quality expectations. Customers’ expectations are formed by their past experiences, word of mouth, and service-firm advertising. The customers choose providers on these bases and, after receiving the service, compare the perceived service with the expected service. If the perceived service falls below the expected service, customers lose interest in the provider. (Kotler 1997, p.476)

Achieving a competitive advantage based on providing outstanding service quality (e.g. Lewis 1989) is a common strategy used by marketers. However, simultaneously maximizing consumers’ perception of service quality and minimizing costs is often difficult. So far, no research has been done looking into the exact nature of the trade-off relationship between price and level of service. (Tse, A. 2001) Though, it is important to obtain insight about the interaction between service quality and price.

Attempts to validate the relationship between price and product quality have proceeded along two different approaches. In one approach, researchers have attempted to verify that buyers do perceive a positive price-quality relationship. Levit (1954) found in his studies that the subjects were more likely to choose the higher priced brand for a product when the price differential was large than when it was small among the different brands of a certain product. These findings were later supported by Tull et al. (1964), McConnell (1968) and Monroe and Krishman (1985), who found that buyers do perceive a positive price-quality relationship. In the second approach, there has been studied whether there is a positive correlation between actual product quality and price. Even though there can be generally draw a conclusion that higher price products have higher demand, which might be as result of a superior quality Gerstner (1985) did not found any support to this reasoning. He assessed the degree of positive correlation between quality and price for 145 products and concluded that the relationship between quality and price appeared to be product specific and generally weak. His findings suggest that some products display a positive quality-price association in the marketplace, but others do not. (Chen et al. 1994)

2.2.2. Perceived service value

According to e.g. Kotler (1997, p.10); Nagle&Holden (2002, p.74) value is the consumer’s estimate of the product’s overall capacity to satisfy his or her needs. And furthermore, the value that is key to developing effective pricing strategy is not use value, but rather what economists call exchange value and what marketers call economic value to the customer. This value is determined first and foremost by what customers’ alternatives are plus the value of whatever differentiates the offering from the alternative. (Nagle&Holden 2002, p.74) And in this phase there is explicitly shown that value is of utmost extent dealing with perceptions of consumer. Perceived value has its origin in utility and psychic space and the relationship is depicted in figure 2. In this study we have concentrated on the perceived utility value and have thus excluded the effects of psychic space values.

To deliver compelling value for both buyers and sellers, mobile service providers must capitalize on the inherent differences that distinguish mobile business. New service model opportunities introduced by mobile service business means that it will be insufficient for service providers to simply meet the current benchmark of the landline Internet experience. Consumers will expect more from their mobile service providers.

According to Monroe the perceived value of adopting any product is dependent on the relationship between the perceived quality and the perceived monetary price (Monroe 1990, p.46). Thus, the perceived value can be increased either by quality improvements or by price reductions. Obviously, an increased perceived value tend to increase the willingness for adoption of an innovation. Moreover, potential adopters might form adoption thresholds with respect to the perceived value, i.e. when the innovation reaches – either by quality improvements or price reductions – a certain threshold level, actual adoption takes place. (Link 1997, p.104)

Customers will pay more for a service that provides more value to them. When the service in question is telecommunications, customers are hard pressed to disclose what that value is. And in comparison to that, consumers of telecommunication services have been very sensitive to the prospect of paying more for telecommunications services and consider the low cost to be a right. (Strouse 1999, 188) This notion is especially important when pricing mobile services – are users of mobile services comparing the prices to telecommunication services.

E.g. the value to the end-users of the I-mode portal is created by the providers of content. The operator does not offer any content services within the portal and although the billing of the services, provided by the operator, is clearly valuable for the content providers it brings little real value to the end user. The increase in revenue can therefore be derived from the value created by the content providers. The ability to charge for this content creates an incentive for the providers to offer these services to the end-user via the operator, since the alternative, to offer the services on a regular Web page or on a unofficial I-mode page, takes away this opportunity to charge. (Jonason&Eliasson 2001, p.347)

2.3. Pricing and price perception

Pricing determines significantly the bottom-line figure on the operating statement. Pricing determines what products will and will not be sold, in what volumes, and with what profit. In more detailed terms, pricing specifies which equipment will be operated, what inventory commitments are to be made, what cash flow can be expected, where sales efforts should be applied, which markets should be approached and penetrated, and, ultimately, what return can be expected on the firm’s invested capital. Therefore, pricing should not be given short shrift as an annoying but somehow necessary detail. (Tucker 1966, p.3)

Dean (1950) introduced the concept of skimming pricing as the strategy of initially setting a relatively high price on an innovation to “skim the cream” of demand. This strategy rests on a general assumption that buyers, who will potentially adopt an innovation (i.e. new product or service) relatively early, are individuals who perceive a very high value in the innovation and thus are prepared to pay a high price for adoption. A number of reasons for a high willingness to pay by the earliest adopters have been put forward in the literature. Early adopters have been characterized as high in social status, venturesome, and eager to try new ideas (Rogers 1983, p.251) and consequently less price sensitive. There are apparent difficulties in evaluating the benefits of an innovation with respect to its price and inherent qualities, as well as in relation to competing alternatives. This also seems to favour a high willingness to pay by the earliest potential adopters of an innovation (Nagle 1987, p.139). E.g. Kollmann (2000) has found in his research of telecommunication industry that there is clear difference in price elasticity between consumer segments.

2.3.1. Price sensitivity

At an aggregated level, price sensitivity is often equated with price elasticity (Link 1997, p.36). Elasticity or sensitivity of demand refers to how volume-sensitive a product or service is to change in price. If the percentage change in quantity sold is greater than the percentage change in price, the demand for that service is considered elastic. If it is less, it is inelastic. Elasticity of demand is a valuable strategic tool of pricing. Astute managers are aware of the demand elasticity of their various services and take advantage of this factor in their pricing policies. Demand information requires a constant policing of markets and ultimate consumers. (Tucker 1966, p.12)

At the level of an individual potential adopter of an innovation, price sensitivity is equivalent to the degree to which a potential adopter is unwilling to pay a certain price for adoption of the innovation. Accordingly, the willingness-to-pay construct conceptualizes an individual’s insensitiveness with respect to a high price for adoption. In fact, price sensitivity at the individual-adopter level appears equivalent to the concept of price consciousness for a potential buyer of any product, defined as “the degree to which he or she is unwilling to pay a he or she is unwilling to pay a high price for a product and willing to refrain from buying a product whose price is unacceptably high” (Monroe 1990, p.51). Price consciousness is strongly (negatively) related to the price acceptability level as well as to the width of latitude of price acceptability (Lichtenstein et al. 1988). Thus individuals who are price conscious are generally not prepared to pay a high price for the product in question. Additionally, the range of acceptable prices is relatively narrow for price conscious individuals. (Link 1997, p.36) For this study, the importance is not on calculating the exact price elasticity percentages of mobile service, but rather acquiring evidence on differences between user segments and how it affects the price perception.

The proactive pricing approach recommended by Ross (1984) focuses on issues related to price increases and decreases. His main thesis is that effective price changes are based on anticipated reactions of customers and competitors, rather than jus the firm’s own costs and circumstances. Under some conditions the consumer is not price-sensitive enough to react to adjustments in price prior to the expiration of the service. Consumers tend to be more price-sensitive at higher price level; the marketer has the opportunity to make price adjustments that may be viewed as more significant amounts. (Finch et al. 1998, p.473-479)

On the other hand bundle pricing is seen by Jonason very promising method for electronic and mobile commerce. When consumers have similar average valuations for the information goods, profits are highest from selling only a single, complete bundle. When consumers have different average values users will prefer to purchase individual items. (Jonason et al. 2001, p.345)

In the figure 3 and 4 there is presented Kollmann’s (2000) study of telecommunication usage and findings of relationship between price, level of usage and charge; and degree of acceptance. According to the study there can be defined three user segments which have different perception/elasticity of a service price and willingness to use the service at that price. He found that in both ends of pricing (high price/low price) the price-elasticity is substantially lower (inelastic). Thus, influencing to these two market segments on pricing is some what non-effective. And therefore, it is possible (and more effective) to pursue a quality-focused marketing strategy (e.g. improvement of service/speech quality). The effect on consumer behaviour is here determined by the necessity of a new purchase. (Kollmann 20000, p.9) In this study we consider charges relating more to traditional voice business not so much to new mobile services business. Therefore we will concentrate more on observing price elasticity. Applying this model into our case, we may expect that users, whose level of usages are very high or low, are not sensitive for price. While at the same time the middle segment is relatively the most sensitive for changes in price.

Tekstikehys: a = highest price at which there is no demand

In the case of charge/acceptance function Kollmann found that there is opposite effect compared to price/acceptance function. Thus, by reducing charges of telecommunication there can be increased acceptance as well. But elasticity is opposite so that segments are most charge-elastic in both ends of the charge scale; and in the middle is the inelastic segment. (Kollmann 2000, p.11) In this study low and high prices are indicators of the usage frequency of mobile services and transmitted data. The middle segment in the figure 3 and 4 is price elastic segment where miner changes in price have substantially high effects on acceptance/usage. Thus price-based is best strategy is best suited for this segment according to the theory.

2.3.2. Perceived price

In the complex pricing environments of services, it is often difficult to use objective price for determining its role. For this purpose there is often proposed to use perceived price. Perceived price can be defined as the customer’s judgment about a service’s average price in comparison to its competitors. The notion of perceived price is based on the simplistic nature of competitive-oriented pricing approach. The guidance available to customers consists of information about whether they are charged more than or about the same as the competitors charge. Perceived price does not eliminate objectivity; rather it adds some subjectivity with the goal of achieving greater organized pricing structure. (Chen et al. 1994, p.25)

In the complex pricing environment of services, it is difficult to use objective price for determining its role. Most of the services Chen et al. considered for the study offer a wide variety of products. The price of these products varies widely within a particular type of service industry. To eliminate this difficulty, Chen et al. (1994) proposes to use perceived price. Perceived price can be defined as the customer’s judgment about a service’s average price in comparison to its competitors. The notion of perceived price is based on the simplistic nature of competitive-oriented pricing approach. The guidance available to customers consists of information about whether they are charged more than or about the same as the competitors charge. Perceived price does not eliminate objectivity; rather it adds some subjectivity with the goal of achieving greater organized structure. (Chen et al. 1994, p.25)

Even though this pricing model could be one of the best methods, its complexity and difficulty of obtaining needed information of customers’ price perceptions is hard to manage. The proactive pricing approach recommended by Ross (1984) focuses on issues related to price increases and decreases. His main thesis is that effective price changes are based on anticipated reactions of customers and competitors, rather than jus the firm’s own costs and circumstances. (Finch et al. 1998, p.473)

Based on the upper mentioned notion, the concept of perceived price is useful for our study as, in addition to insights to personal customer characteristics, there can be obtained generalisible insight to formation of high or low perceived prices.

2.3.3. Pricing of mobile services

Companies with new products for which they are trying to build cash flow often make the mistake of building the start-up cost of acquiring and servicing a new customer into a large, up-front fee. Because high uncertainty undermines perceived value, such companies lose potential sales and win sales only at lower prices than they otherwise could. By absorbing the up-front cost in higher monthly fees, the seller communicates confidence that customers will be satisfied and enables customers to pay as they enjoy a known value from product usage. Consequently, the seller should close more sales and, assuming that the product or service delivers the promised value so that the customer continues to buy it, the seller can ultimately expect a greater cash flow and a higher net present value (NPV) per customer acquired. (Nagle et al. 2002, p.90-91)

The unique characteristics associated with services compared with products include intangibility, inseparability of production and consumption, heterogeneity, and perishability. Since services themselves are intangible, one of the most tangible aspects of the offering is the price of the service. In a study of 323 service firms by Zeithhaml et al. (1984), the average responses across all firms showed that cost-oriented pricing strategies were more often used than either competition or demand-oriented pricing strategies. One of the overriding concerns among service firms is covering costs, as service costs are difficult to calculate. Of additional interest was the finding that most of the service firms in the study do not reduce prices to increase business during slow demand periods. (Finch et al. 1998, p.474)

One special feature, concerning pricing in telecom industries, has been brought up by Laffon & Tirole (2000, p. xv): “marginal-cost pricing for all services is not viable in telecom industries. Price discrimination may be the prerequisite for the viability of certain investments.”

Features and development of wireless business makes it necessary to concentrate on pricing in order to develop and invent new methods of pricing that best induce the adoption and development of mobile services business. As operators must do close co-operation with other parties of mobile services production – services and content providers – it lays great challenges for pricing structure to encourage all parties to produce high quality output.

As additional features of mobility and personalization over mobile devices enable new types of e-commerce applications (Jonason et al. 2001, p. 341) it also compels close co-operation of different parties to produce high quality mobile services. For players interested in launching new applications and services, alliances can present an opportunity to test the market by shifting the rules of the game. However, the difficulty with alliances and partnerships is their maintenance and management over the long term. (Nordström 2001)

Through pricing there can be obtained significant effects as Kollmann in his studies has discovered. Kollmann’s notions are important in creating a true image about the importance of pricing in mobile services business. First of all he found out that customers of telecom services are generally very price sensitive and make their decisions mainly based on price. Secondly there was discovered that there exists clear segments that posses different price elasticity. These two characteristics enable a telecom company to use pricing actively in competing markets. (Kollmann 2000) In pricing of telecommunication services there must also take under considerations the effects of positive externalities.

There are two obvious shortcomings to single price strategy. First, the firm clearly is leaving excess money on the table for many buyers who are willing to pay more. These high-end buyers may perceive significantly greater value from purchasing this product, relative to other buyers. This excess value that consumers receive by paying a lower price, when in fact they are willing to pay a higher price is referred to by economists as “consumer surplus”. The firm would be better off if it could capture some of this surplus by charging higher prices to these buyers. The second problem with a single price strategy is that the firm leaves nearly half of the market unsatisfied, even though it could serve it at prices above the unit variable cost. In industries with high fixed costs, serving those additional customers is very tempting, and possibly very profitable. Moreover, by ignoring these buyers, the firm leaves wide open an opportunity for low-cost competitors to enter the market and establish a competitive presence. When those competitors won a large enough installed base at the low end to establish a reliable service network, they were then able to attack higher-margin markets. (Nagle et al. 2002, p. 228)

Rather than the monthly fixed charge, it is income from the number of call minutes which determines basic commercial success for network providers (Kollmann 2000, p. 8). One major improvement of packet-based technology over circuit-switched technology is that users can pay for transport volume rather than time of transport (Jonason et al. 2001, p. 342). The network provider can charge the end-user per volume of data transmitted or received, greatly reducing the cost of real time applications. As these two major technical innovations redefine the wireless services that will be offered in the near future, they also create a pricing problem for the operators. (Jonason et al. 2001, p. 341)

Subscription plans also make it easier to develop close relations with customers. Subscription pricing lends itself to finding out what the consumers need, and to customization of offerings. (Fishburn et al. 1997, p.6) For example, in software business there has been employed so called “functionally based tired” –pricing. Clients can select from three tiers: Select, Gold, and Platinum. Select has basic functionality while Platinum has full functionality. This tiered structure comes from traditional marketing practice, which assumes that if a buyer is given a choice of three service levels, the buyer will choose the middle package. This has been a common marketing tactic in the US cellular service market. (Bontis et al. 2000, p. 251) This pricing method is closely related to Paris-Metro pricing. In that there is created two or three service classes but they do not differ in service functions but only in charged prices and thus in expected service quality.

There have been tried and suggested many other charging models for mobile and internet usage. For example following have been mentioned by Cushnie and Hutchison (2000) and Reichl & Haidegger (2002): (i) usage based charging, in which consumer is charged according to realized consumption; (ii) fixed price, is non- or partly metered pricing scheme which does not change according to the usage; (iii) bundle pricing, is used for pricing/billing of bundles of services or products with a lump sum; (iv) Paris-Metro, is a pricing method especially for congestion management purposes, the pricing scheme is mostly self-regulating the congestion with two-piece pricing model; (v) packet charging, is used in new e- and mobile services which uses packet-based technology; and (vi) edge pricing, is characterised by concentrating charging functionalities at the edges of the network (e.g. access routers) and has become one of the central paradigms for charging Internet services.

These methods will be merged into a combination which would best support the business goals of the parties in a question. We will discuss more detailed Paris-Metro pricing and bundling in coming chapters.

One of the major incentives for developing new revenue and pricing models in mobile business is the fact that mobile services cannot be distributed without charge as Jonasson et al. (2001) has noted. However there is clear evidence in other businesses that pricing is a powerful tool in inducing new services or products. Low prices or even “free-ware” is used especially in software business to boost sales of new products. A most common argument for using low prices in new products is to accelerate the diffusion to achieve the critical mass which after the diffusion process is self-enhanced.

2.3.4. Bundling of services

By definition bundling of telecommunication services is the packaging of different telecommunications services for customers. Usually telecommunications bundling simply refers to local and long-distance services, but it is also used to describe the combination of either or both of those service with Internet access, cable television, wireless, and any other service broadly described as telecommunications. (Strouse 1999)

According to the Stigler’s, Adams and Yellen’s (1976) analysis the reservation values for the components of the bundle were negatively correlated. That feature made it appear that bundling serves much the same purpose as third-degree price discrimination. A surprising result of Schmalensee (1984) analysis was that bundling can be profitable even when demands are uncorrelated or even positively correlated. His explanation for the profitability of bundling was that it reduces the effective dispersion of reservation values and thereby makes it possible for the seller to extract a greater fraction of the potential surplus. (Salinger 1995)

Carroll et al. (1999) has discovered a set of conditions under which bundle pricing is profitable strategy. Carroll discovered that a company should cover the three necessary conditions for price discrimination to occur. First, the firm must have some market power[1] (Jeitschko 2001) which is often the case as many of the operators are old state monopolies. But it is important condition to be estimated especially in a case of new virtual and pure mobile operators. To cover this condition an operator must have a prominent market share which excludes many young mobile and virtual operators. Second, there can at best be imperfect arbitrage opportunities for consumers[2] (Jeitschko 2001), which is highly important in case of mobile services as e.g. video-clips, music, pictures, etc. could be easily forwarded to other users with lower fees if not technically restricted. It is thus essential that the resell opportunity is hindered technically or the costs that enable it must be prominent. And third consumers must have different price elasticity of demand[3] (Jeitschko 2001). Price elasticity in telecom business has been studied e.g. by Kollmann (2000) who discovered three consumer segments which have different price elasticity. These segments were differentiated according to the amount of the usage of telecom services. In mobile services business consumers could be segmented according to the amount of transferred data. As Kollmann’s findings were also empirically proved, there can be expected that also in mobile services business there exists different customer segments that differ in the amount of transferred data (service usage), and in price elasticity. Even though this is the case Kollmann’s findings cannot be directly converted to mobile services business as there exists some prominent differences in the nature of these business and consumer characteristics might also differ notably. The segmentation by price elasticity and service usage must be thus empirically studied case-by-case and market-by-market.

To be able to state the basic conditions for the usage of bundle pricing strategy in mobile services business, we have united together the separate findings of Adams&Yellen’s (1976) and Carroll’s et al. (1999). Even though there can be stated that bundle pricing strategy can be profitable when above discussed three general conditions are covered, the condition must be empirically studied and evaluated and market-by-market.

In generally, according to Schmalensee, a seller faces three alternative strategies to offer her or his products or services (Venkatesh & Vijay 1993):

1. Pure components: the seller prices and offers the component products/services as separate items, not as bundles.

2. Pure bundling: The seller prices and offers the component products/services only as a bundle and not as individual items.

3. Mixed bundling: The bundle as well as the individual component products/services are priced and offered separately.

Adams and Yellen (1976) argued that mixed bundling at least weakly dominates pure bundling. Moreover, McAfee, McMillan, and Whinston (1989) have shown that, while mixed bundling virtually always strictly dominates pure bundling, the optimal bundle price is sometimes greater than the sum of the prices of the individual goods (Salinger 1995). When valuing pricing strategies for new mobile services, there should be laid few goals, which are expected to be covered. These goals are attendance maximization, service usage maximization, and surplus maximization (Ansari et al. 1996). The optimal number of items to be included in a service bundle for a profit-maximizing firm that uses pure components, pure bundling, or mixed bundling strategies is determined by Venkatesh & Vijay (1993).

The context of bundle pricing approach covers is complex and there are several aspects that should be considered when choosing right strategy (see above). The bundle context includes following areas: price segmentation, price discrimination, and product range restriction, reduction in classification /processing costs, scope economies, consumers’ search economies, and risk reduction. (Venkatesh & Vijay 1993)

Some examples of practices of using bundle pricing (Strouse):

Wireless providers have bundled services as a competitive differentiation tactic. Cellular providers bundled airtime minutes into their service offerings to serve as a competitive weapon against their local competitors. PCS providers, eager to attract the customers of cellular providers, have leveraged the digital capabilities of the PCS technology to bundle news and stock quotes, messaging services, and other enhanced services in the basic price of the wireless service. Paging services frequently offer stock quotes and voice mail in their subscription prices.

Local exchange carriers have discovered a synergy between their high-speed digital subscriber line (DSL) offerings and Internet access. IXCs have found markets by bundling long-distance and other services such as Internet access. MCI offers its long-distance subscribers a lower-than-market price for Internet access. In another example of bundling, in the United Kingdom, cable operator Nynex Cablecomms Ltd. offered its customers free local calling within its franchised area if they would subscribe to two premium-priced cable television channels. Nonetheless, bundling of local exchange and long-distance services is the most anticipated form of bundling, and it is not yet widely available to consumers.

The arguments in favour of bundling are thus strong, and suggest that à la carte or unit pricing will not be the dominant mode of commerce in information goods. Observation of Adams and Yellen shows that mixed bundling items are offered for sale separately, (and in combination, but with the price of the individual items higher than they would be otherwise) is always better than pure bundling (items are available only in combinations). Furthermore, bundling is most appropriate for producers with an established brand (Fishburn & Odlyzko 1997), i.e. market power.

In online service area, it is common for customers to pay for larger blocks of time than they used. The reason is that there exists a need for insurance (predictable costs), overestimation of usage, and a hassle factor (whether each call is worth the money or not). In addition to the consumer preference for flat-rate and bundle pricing, there are reasons for producers, especially in areas where network externalities are important, to also like these plans. These factors are part of a general preference by consumers for simple and predictable pricing. (Fishburn & Odlyzko 1997)

Information goods are characterized by negligible marginal costs, and therefore arguments in favour of bundling are stronger for them than for physical goods. Bundling arguments show that producers can obtain more revenue by combining disparate items, since that allows them to exploit uneven preferences that consumers have for different goods. In most situations bundling is advantageous to the producers. (Fishburn & Odlyzko 1997)

When using a flat-rate or bundle pricing, there should not be used pure single price strategy as there are related obvious shortcomings. First, the firm is clearly leaving excess money on the table for many buyers who are willing to pay more (i.e. consumer surplus). These high-end buyers may perceive significantly greater value from purchasing this product, relative to other buyers. Second, the firm leaves nearly half of the market unsatisfied, even though it could serve it at prices above the unit variable cost. In industries with high fixed costs, serving those additional customers is very tempting, and possibly very profitable. And by ignoring these buyers, the firm leaves wide open an opportunity for low-cost competitors to enter the market and establish a competitive presence. (Nagle & Holden 2002)

Bundling pricing is seen by Jonason very promising method for mobile business. When consumers have similar average valuations for the information goods, profits are highest in selling only a single, complete bundle. (Jonason&Eliasson 2001) Bundle pricing is a way to present an unambiguous price for the usage of a bundle of services. Bundling also increases the amount of pricing tools and choices for service producers. E.g. Laffont argues that it is optimal to lower the price of a service if doing so raises the demand for a complementary service on which the utility charges a mark-up (Laffont & Tirole 2000).

Cable companies are providing voice and Internet access services over the coaxial cable in their networks. High-density television broadcasts will be digital, presenting opportunities to merge services. Convergence represents the trend under which formerly separate services are merging, and bundling will represent the packages of service. (Strouse 1999) One of the big attractions of Java and the Network Computer to the software industry seems to be the possibility of charging consumers according to their usage of a particular product. However, while there are obvious attractions to per-use pricing, the basic economic arguments based on utility theory are not as clear as for bundling, where those arguments strongly support the idea of selling combinations of items. The simple utility maximization argument might favour per-user pricing in a substantial fraction of cases, what we observe in the market are repeated failures of à la carte pricing. (Fishburn & Odlyzko 1997)

The arguments in favour of bundling are strong, and suggest that à la carte or unit pricing will not be the dominant mode of commerce in information goods. E.g. in the United States, customers wants bundling and so do carriers (Strouse 1999). However, even if bundling does dominate, as Fisburn predicted, there are likely to be niches for fixed-fee sales. One of the few general results about the economics of bundling is the observation of Adams and Yellen that mixed bundling is always better (with proper price) than pure bundling. (Fishburn & Odlyzko 1997)

In mobile services business to be able to employ the bundle pricing, there must be, in general, realized three, basic conditions: the firm must have some market power; there can at best be imperfect arbitrage opportunities for consumers; and consumers have different price elasticity of demand.

Mobile services business, as mentioned, is yet to be defined and services are thus difficult to value and compare. Therefore the perceived risk is relatively high for consumers, but also for service and content providers. By bundling services into larger blocs, consumers need to evaluate only one bloc of services rather than several items separately (see figure 1). Hence total costs are more transparent and total value is easier to perceive. From the basis of these observations, there can be come to a conclusion that through bundling of mobile services there can be modified service bundles which have higher cost transparency, greater perceived value, and lower price elasticity. In the end, consumers make their decisions according to their own personal perceptions of expected values and perceived costs of mobile services with often contradictory information.

Also for service and content providers perspective there exists great uncertainties related to mobile services business. By creating blocs of services, there is created better defined and easier to value mobile services environment. This means less perceived risk and increased calculable of the service environment. The bundling is also a clear way of positioning the services and content in respect to competitors. Bundles may thus be considered as an indicator of service/content quality and used for segmentation purposes. The definition problems can thus be radically diminished through bundle pricing approach.

Bundling in mobile services business where are two or more co-operating actors, pricing decision and the follow-up of service usage gets easier as services are included into bundles. This supports also strategic perspective of positioning services into chosen bundles. In mobile environment billing would be co-ordinated by operator which would further distribute the earnings according to usage of services in the bundle.

The objectives for mobile pricing should enable fixed prices and price differentiation, and increase the efficiency of pricing (by capturing part of this consumer surplus). From customers perspective mobile services are shown in a simple and transparent way whence purchase of mobile services is enhanced. And from mobile services producer’s perspective bundling is considered as price discrimination which offers practical and profitable tool for defining emerging business; reducing perceived risks of consumers and service providers; and aggregating the modest value dimensions into a bloc of higher perception of total value.

In addition to above mentioned arguments for the bundle pricing strategy, also Strouse (1999) found several benefits on favour of bundling: the preference of customers for a single bill, and for simplicity which bundled services offer; for providers that are currently prohibited from entering certain markets, bundling represents the opportunity to enter new markets without requiring large investments (bundling requires smaller investments in existing facilities); bundling creates an excellent cross-selling opportunity (it is many times easier to sell a new product to an existing customer than to sell existing products to a new customer).

2.3.5. Innovativeness

Consumer innovativeness has remained a field of enquiry beset by confusion over basic terminology and an inability to come to terms with the weak evidence on which its generalisations about the personality profiles of “consumer innovators” are based (Foxall 1995, p.269). Innovativeness is the degree to which an individual adopts an innovation relatively earlier than other members in the system do. The construct is measured in one of two ways in marketing studies. The first measure is based upon time of adoption, such as for those individuals who purchase in the first x weeks, moths, and so on. The second is a count of how many of a pre-specified list of new products a particular individual has purchased at the time of the study. (Engel et al. 1986, 542) On the other hand, Goldsmith and Hofacker (1991) have defined innovativeness as “the tendency to learn about and adopt innovations (new products) within a specific domain of interest”. Whereas the most common definition is the definition by Rogers (1983) in which it is defined as relative-time-of-adoption. (Link 1997, p. 33)

According to findings of Kirton (1976 and 1989) the innovator prefers to think tangentially, challenges rules and procedures and is less inhibited about breaking with established methods and advocating novel perspectives and solutions. The innovator is easily bored by routine and seeks novelty and stimulation in discontinuous change; he or she tends towards risk-taking exploration and trial. (Foxal 1995, p. 272) In addition, according to Engel et al. (1986, p. 544) Income is almost always useful in profiling innovativeness with higher-income people having the ability to take the risk of trying new products. For low-priced products, this relationship may not be as important, however.

3. Research design and methodology

As both Cooper&Emory (1995, p.114) and Kerlinger (1981, p.300) have stated that research design is the plan, structure, and strategy of investigating conceived as to obtain answers to research questions and to control variance. Research design has two basic purposes: to provide answers to research questions and to control variance. Therefore it is important to emphasize the discussion of these areas and more accurately depict the research outlines.

In this chapter there has provided a closer look at the areas of research design relevant for this research. First there is discussed the research methodology in order to classify this research under the prime research designs. Second there has been brought a glance to data collection procedures. And third part of this chapter consists of discussion on reliability and validity from the point of view of this research.

There can be classified three broad categories through which a phenomenon can be understood: experience, reasoning and research. The experience does itself subsume a number of sources of information that may be called upon in a problem-solving situation. It is most immediately at hand for all people. The second category reasoning consists of three types: deductive reasoning, inductive reasoning, and the combined inductive-deductive approach. And the third category research has been defined by Kerlinger as the systematic, controlled, empirical and critical investigation of hypothetical propositions about the presumed relations among natural phenomena. (Cohen&Manion 1980)

As methodological choices have played a critical role in scientific research of all sciences, also Peirce has sorted four general ways of knowing or, more accurately, fixing belief. The first is the method of tenacity were men hold firmly to the truth, the truth that they know to be true because they hold firmly to it, because they have always known it to be true. A second method is the method of authority. This is a method of established belief. If a noted physicist says there is a God, it is so. A third method is a prior method. It rests its case for superiority on the assumption that the propositions accepted by the “a priorist” are self-evident. But the fourth method is the one of interest method of science. The scientific approach has one characteristic that no other method of attaining knowledge has: self-correction. A scientist does not accept a statement as true, even though the evidence at first looks promising. He insists upon testing it. (Kerlinger 1981 & Christensen 1997)

3.1. Methodology of thesis

Research methodology literature has traditionally separated three basic types of research design: exploratory, descriptive, and causal research. Exploration is particularly useful when researchers lack a clear idea of the problems they will meet during the study. Through exploration the researchers develop the concept more clearly, establish priorities, and improve the final research design. Exploration may also save time and money if it is decided the problem is not as important as first thought. (Cooper 1995, p.117-123)

The purpose of descriptive research is to provide an accurate snapshot of some aspect of the market environment, such as consumer evaluation of the attributes of some products versus competing products (Aaker et al. 1995, p. 73-74). Descriptive research is used when the purpose is as follows: to describe the characteristics of certain groups; to estimate the proportion of people in a specified population who behave in a certain way; and/or to make specific prediction (Churchill&Iacobucci 2002, p. 107-108). Whereas Best has defined descriptive research with conditions or relationships that exist; practices that prevail; beliefs, points of view, or attitudes that are held; processes that are going on; effects that are being felt; or trends that are developing (Cohen&Manion 1980, p.68).

Descriptive studies require a clear specification of who, what, when, where, why, and how of the research. (Churchill&Iacobucci 2002, p.107-108) In descriptive research, hypothesis often will exist, but they may be tentative and speculative. (Aaker et al. 1995, p.74)

A causal research is concerned with determining cause-and-effect relationships (Churchill 1995, p. 145) meaning that in causal study there is explained relationships among variables (Cooper&Emory 1995, p.116) and their effects to each other. It is of little value to know the variables that affect an outcome unless we know how they affect it. Research projects designed to investigate these relationships are know as causal studies (Tull&Albaum 1973, p.10).

In addition Goldberger (1973) has noted that in methodological terms, the models have been referred to as simultaneous equation systems, linear causal schemes, path analysis, structural equation models, dependence analysis, test score theory, multirait-multimethod matrices, and the cross-lagged panel correlation techniques. Behind this diversity of subject matter and terminology, several common features are identified. One relates to the analysis of non-experimental data; the absence of laboratory conditions demands that statistical procedures substitute for conventional experimental controls. A second one concerns hypothetical constructs; many of the models contain latent variables which, while not directly observed, have operational implications for relationships among observable variables. A third common element relates to systems: the models are typically built up of several or many equations which interact together. (Bagozzi 1980, p.83-84)

Bagozzi (1980) has also depicted “levels of understanding in explanation” in causal studies. These levels are presented in table 1 which contains four different levels of understanding. Meaning that the higher the level of understanding the deeper insight is produced from the research subject.

TABLE 1. Levels of understanding in explanation.

Levels of understanding in explanation	Explanation
One	Phenomenon z exists in state Q.
Two	The phenomenon is of the nature Q and is produced by factors x₁, x₂,…, x_n.
Three	Factors x₁, x₂,…,x_n are interactive or have interacted in manner y₁, y₂,…, y_n to produce in some past or present time a phenomenon of the nature Q.
Four	Factors x₁, x₂,…,x_n interact in a manner y₁, y₂,…, y_n for reasons w₁, w₂,…, w_n, thus producing a phenomenon of the nature Q.

The subject of this thesis is to study the effects of pricing and usage related factors on price perception of mobile services and interestedness in mobile service bundles. There is measured, in this research, relationship of factors affecting to price perceptions and variables explaining these variables. From these bases this research can be defined as descriptive study as it fits well to its definitions previously discussed. But also causal research would be highly useful for studying this topic especially as it would useful to acquire the knowledge from the factors causality towards price perception.

3.2. Data collection

For collecting empirical data there can be used two basic methods, monitoring and interrogation processes. The former includes observational studies, in which the researcher inspects the activities of a subject or the nature of some material without attempting to elicit responses from anyone. In the survey mode, the researcher questions the subjects and collects their responses by personal or impersonal means. The data may result from i) interview or telephone conversations, ii) self-administered or self-report instruments sent through the mail, left in convenient locations, or transmitted electronically or through another means, or iii) instruments presented before and/or after a treatment or stimulus condition in an experiment. (Cooper&Emory 1995, p.115)

On the other hand Curchill&Iacobucci (2002, p.454, 458) have classified data collection techniques into two basic groups, non-probability samples and probability samples. Non-probability samples involve personal judgment somewhere in the selection process. Sometimes this judgment is imposed by the researcher. The fact that the elements are not selected probabilistically precludes an assessment of sampling error. Whereas, in probability samples one can calculate the likelihood that any given population element will be included in a probability sample because the final sample elements are selected objectively by a specific process and not according to the whims of the researcher. The objective selection of elements allows the objective assessment of the reliability of the sample results which is not possible with non-probability samples.

According to upper mentioned classifications of sampling techniques, this research can be considered as probabilistic. The sample data was randomly collected through stratified sampling technique from the operator’s customer data base which consists of a large amount of mobile services users. Though, the randomness was limited by two ways: first, the sample data was collected from one customer data base; and two, samples were collected from three different segments according to customers’ usage level of mobile services.

This level of categorization of sampling technique was still inadequate. As mentioned, the target in this research was to describe three different user segments of mobile services: heavy users, moderate users, and prospective users. This characterisation was used as a criterion for more elaborated classification under the probability sampling technique. Churchill (1995, p.605;610) notes that through stratified sample there is enabled the investigation of the characteristic of interest for particular subgroups. It is important that studied subgroup is adequately represented in the sample for which purpose; stratified sampling is one way of ensuring adequate representation from each subgroup of interest. And more preciously, there is used a proportionate stratified sample in which the number of observations in the total sample is allocated among the strata in proportion to the relative number of elements in each stratum in the population. From this basis there was taken equal sample sizes from each customer segments to confirm the adequate representation of each segment.

In this research there is applied a survey method through which there are well designed tools for conducting descriptive study in mobile services business. And as there is targeted to produce closer insights to effects of price perception on adoption of mobile services, there is needed a data collection method for gathering the needed information on characteristics of mobile services pricing. For this purpose there has been chosen a mail questionnaire survey method. Though keeping in mind the recommendation of both Parten’s (1950) and Kerlinger’s (1981) who have advised to avoid mail questionnaires if possible as “most mail questionnaires bring so few returns, and these from such a highly selected population, that the findings of such surveys are almost invariably open to question”. “The best advice would seem to be not to use mail questionnaires if a better method can possible be used”.

Despite of the previous statement, the empirical data was collected through quantitative postal survey in which there was sent 3000 questioners (1000 questioners for each customer segment) to operator’s mobile service customers during the spring 2003. Characteristics of each segment were following: ‘prospective users’ were those who had broad-band internet connection (which they paid by themselves) and basic mobile phone connection but no WAP (wireless applications protocol)-connection. They were using basic mobile services like short messages, ring tones, etc. ‘Moderate users’ did also have a broad-band Internet connection but instead of basic mobile phone connection they had WAP-connection. They were using some data intensive services and transferred more data through mobile connection than prospective users. It was also required that they must have usage of data and voice services during the past six months. ‘Heavy users’ were differentiated from the two previous groups by not having analogue broad-band Internet connection when all internet usage was conducted through mobile Internet. There was also required them to have high volume data transfers and heavy usage of data intensive services. And customer must also be active users of mobile services. And they had to have also gprs (general packet radio service) or regular data transfers during the past six months.

Respondents were asked to complete a 7-point scale (Likert scale) on each question or proposition indicating its importance in defining their beliefs, attitudes, and intentions toward mobile services and pricing of mobile services. Every question had also a choice of “I don’t know”. Three types of questionnaires were prepared and sent. Even though most of the questions addressed the same issues, there were some differences in terms of experience about using the mobile services and attitudes towards pricing practices. Response rate after three mailing rounds was 25.9% (778) which was a bit lower than expected 30 percent.

3.3. Validity and Reliability

The concepts of reliability and validity concern the degree to which the measuring instrument is free of measurement error. Reliability and validity are essential criteria for developing trustworthy information about consumers and their behaviour. It has been argued that most marketers fail to ascertain either the reliability or the validity of the measuring instruments they use. Literature also suggests that the best resolution to the problem of reliability and validity is to verify research findings by quantitative methods whenever possible (Runyon & Stewart 1987, p.44).

3.3.1. Reliability

Reliability is the “accuracy or precision of a measuring instrument” (Kerlinger 1980, p.443). Reliability refers to the degree to which a measure is free of variable error. In other words the less error the greater reliability. Reliability refers to the accuracy, consistency, stability over time, and reproducibility of a measurement instrument. Reliability has been identified as a necessary but not a sufficient condition for validity (Nunnally 1978).

Reliability is concerned with estimates of the degree to which a measurement is free of random or unstable error. It is not as valuable as validity determination, but it is much easier to assess. Reliable instruments are robust; they work well at different times under different conditions. This distinction of time and condition is the basis for frequently used perspectives on reliability – stability, equivalence, and internal consistency. (Cooper&Emory 1995, p.153) The most common type of reliability measurement evaluates the internal consistency of items in a scale. Two types of internal consistency can be measured: 1) average inter-item correlation, and 2) Cronbach’s alpha, which measures the internal consistency of items in a scale (Garson 1982).

Internal consistency is the degree of homogeneity among the items that constitute a measure. Internal consistency is determined by statistical examination of the results obtained, typically equated with Cronbach’s (1951) coefficient alpha. Cronbach’s alpha measures true variance over total variance. According to Nunnally (1978) the alpha of a scale should be greater than 0.70 for the items to be used together as a scale. The alpha for the total scale is also computed on the assumption that the items under examination are deleted. Nunnally (1978) gives a common guideline for the alpha standards of reliability: a) early stagy of research, alpha = 0.5-0.6, b) basic research, alpha = 0.7-0.8, and c) applied settings, alpha = 0.8-0.9.

3.3.2. Validity

Historically the most common definition of validity is that it refers to the extent to which a test or a set of operations measures what it is supposed to measure (Ghiselli et al. 1981, p.266). It can be epitomized by the question: are we measuring what we think we are measuring? Thus emphasis in this question is on what is being measured. Although the most common definition of validity was given above, it must be emphasized that there are more than only one validity. The most important classification of types of validity is content, criterion-related, and construct. (Kerlinger 1981, p. 457)

The content validity of a measuring instrument is the extent to which it provides adequate coverage of the topic under study (Cooper&Emory 1995, p.149). Content validity refers to the degree to which a specific set of items are representative and appropriate sample of the content contained in the instructional objectives the attainment of which the test is intended to measure. The theory of content validity suggests that a measurement has a content validity when its items are a randomly chosen subset of appropriate items. (Cronbach 1951) Content validation is guided by the question: is the substance or content of this measure representative of the content or the universe of content of the property being measured.

Content validation is basically judgmental. The items of a test must be studied, each item being weighed for its presumed representativeness of the universe. This means that each item must be judged for its presumed relevance to the property being measured. The universe of content must be clearly defined; that is, the judges must be furnished with specific directions for making judgments, as well as with specifications of what they are judging. (Kerlinger 1983, p.458-459)

The two most commonly used methods of validation involve the use of logical reasoning and personal judgments of groups of experts in the field (Tull&Albaum 1973, p.92). But depending on the circumstances and the characteristics to be measured there are a number of ways that expert judgments about content validity can be augmented. These are in no way a replacement for subjective judgments. They are merely some possible strategies for enhancing it.

Item homogeneity is an approach to enhancing subjective judgment and giving a more precise statement of the degree of validity is to view the relative homogeneity of the components of the test as an indication of content validity. The argument is that if the component parts of the test are intended to measure the same trait, their scores should be positively correlated; and the higher their cross-correlations, the more content-valid is the test. Another procedure that could enhance our confidence in judgments of content validity is for two or more panels for experts to go through the content-validity procedure independently and to compare the final product. (Ghiselli et al. 1981, p.274-279) In the present study, content validity was increased by planning the questions accurately.

Scientifically speaking, construct validity is one of the most significant advances of modern measurement theory and practice. (Kerlinger 1973, p.461). Construct validity is concerned with knowing more than just that a measuring instrument works. It is involved with the factors that lie behind the measurement scores obtained; with what factors or characteristics account for, or explain the variance in measurement scores. (Tull&Albaum 1983, p.91)

In widest sense construct validity could be thought as pertinent to any theory of behaviour. That is, understanding the construct, or knowing the theory, permits a host of predictions about behaviour when the construct is used as the independent variable. In this sense, construct validation is a deductive process, and construct validity is established to the extent that specific hypotheses deduced from the theory are substantiated in empirical studies. By contrast, in inductive process what’s being validated is not a theory about some characteristic of individual differences, but the content of a particular test. Construct validity would be established to the extent that a large number of relationships between the test and other variables was determined, and the pattern of relationships that was found clearly indicated the meaning of the test score. (Ghiselli et al. 1981, p.282-283)

Construct validity is concerned with knowing more than just that a measuring instrument works. It is involved with the factors that lie behind the measurement scores obtained; whit what factors or characteristics account for, or explain, the variance in measurement scores. Construct validity refers to the degree to which a test measures the target construct, or psychological concept or variable, inferred from all of the logical arguments and empirical evidence available. Construct validity is thus directly concerned with the relationship of a variable to other variables. (Gage 1991) One significant contribution to testing construct validity is Campell and Fiske’s (1959) procedure, called the multitrait-multimethod matrix, that uses ideas of convergence and discriminability and correlation matrixes to bring evidence to bear on validity. In a mulitrait-multimethod analysis, more than one construct and more than one method is used in the validation process thereby obtaining a method-by-measure matrix. Another method of construct validation is factor analysis, which is a method for reducing a large number of measures to a smaller number called factors by discovering which measures measure the same thing (Kerlinger 1981, p. 468).

3.4. Model construct

In model construction there are observed factors that are expected to be most likely influential on perceived prices of mobile services and preference for acquiring mobile services in bundles. In this construct there was measured the price perception amongst the three customer segments who are using or expected to take into use new mobile services. There are four factors that are expected to be influential on price perception and interestedness in mobile service bundles: (i) customers’ innovativeness; (ii) price sensitivity (elasticity); (iii) satisfaction to operator’s services, and (iv) readiness to invest more on mobile services. This construct is partly adopted from the works of total quality management (TQM) and value analysis/value engineering (VA/VE) researchers (e.g. Kotler 1997&Miles 1989).

Deriving from these theories there is modified a concept of value mix e.g. by Ho & Cheng (1999). Value mix is conceptualised as the combination of function, quality and price (figure 7). The concept embraces the essence of the traditional approaches to value and quality. Value mix describes customers’ determination of the value of a product or service in terms of function, quality and price. Cost, as not regarded by customers, is excluded. It is these three elements of value mix that every organisation should consider when designing and delivering product and service. (Ho & Cheng 1999)

FIGURE 7. Components of value mix (Ho & Cheng 1999).

Also Zeithaml (1988) has found four broad expressions which are the sources of value:

(i) value is low price;

(ii) value is whatever I want in a product or service; and

(iii) value is the quality I get for the price I pay; and

(iv) value is what I get for what I give (Gouvêa et al. 2001).

And to go further, according to Lovelock (Lovelock 1991, p.237), value is based on this last expression meaning that value is the sum of all the perceived benefits minus the sum of all the perceived costs. As we want to modify more this conceptual model presented in figure 7, we will propose a theory of perceived price as an alternative for value expression. Chen et al. have defined perceived price as the customer’s judgment about a service’s average price in comparison to its competitors (Chen et al. 1994) and perceived price will include monetary as well as non-monetary prices proposed by Zeithaml (1988). E.g. Gerstner (1985) has also found concrete evidence of clear positive relationship with quality and price (Chen et al. 1994, p.26). On the grounds of these notions there can be stated that perceived value and perceived price are closely connected. Therefore we will replace the value-mix in figure 7 with perceived price in order to emphasize the importance of pricing as an inducer of adoption of new mobile services.

In this research there will be measured the relationship of the four factors discussed earlier towards perceived price of mobile services. In comparison to the model above, there is replaced function and quality factors with satisfaction to operator’s services as quality is a broad and multidimensional term and would therefore require larger work for defining its meaning in this research. Satisfaction to operator’s services is instead more straight-forward and is directly related to customers’ perceptions. This way there can be also produced more operator specific results which would be easier to apply to business practices. Price factor in the figure 7 is replaced by price sensitivity. Customers’ innovativeness in added into the figure as it is expected to explain part of the differences in price perceptions. And fourth factor included into the figure is readiness to invest more on mobile services. As it brings out the aspect of future predicting which customers are most likely to increase their mobile service usage and on what grounds. These are the factors observed in this study which addition, naturally, they is also examined demographic variables and their relationship with customers’ price perception.

In figure 8 there is constructed a model for this research. There are two phenomenons towards which the factors are compared – price sensitivity and preference for mobile service bundles. Demographic variables are anticipated to possess a negative interaction with price perceptions. It is also hypothesized that price elasticity/sensitivity is positively correlating with price perception. Also satisfaction to operator’s services is expected to have negative relationship with price perception. Innovativeness and perceived price are expected to correlate positively as the most innovative customers are more likely to possess more emphasize on other service related factors than price. Readiness to invest more on mobile services is naturally expected to have a positive relationship with price perception. A customer who regards mobile services very expensive is unlikely to increase investments of those services.

For creating hypothesis for the factors influencing on mobile service customers’ price perceptions there is first examined the price sensitivity and satisfaction, which are interrelated as through satisfaction there can be increased/decreased price sensitivity. (Fornell et al. 1996). This is based on the fact that companies with higher satisfaction values are able to receive higher prices from customers. Price acceptance is closely related to the economic concept of consumers surplus (and thus consumer price perception): …”the excess of the price which a man would be willing to pay rather than go without having a thing over what he actually does pay is the economic measure of his satisfaction surplus” (Marshall, 1980, p.260). Thus we expect customers who are satisfied with a product or a service to accept higher prices, or other words perceive prices inexpensive. According to this, the following two hypotheses can be formulated:

H1: A customer’s satisfaction influences significantly on a customer’s price perception.

H2: A customer’s price sensitivity has a significant influence on a customer’s price perception.

Innovativeness and investment readiness are related to innovation diffusion and consumer adoption processes. Roger defines a person’s innovativeness as “the degree to which an individual is relatively earlier in adopting new ideas than the other members of adopters. Roger also sees five adopter groups as differing in their value orientations. Innovators, who are willing to try new ideas at some risk. Early adopters are guided by respect. The early majority are deliberate. The late majority are sceptical. And laggards are tradition bound. (Rogers 1983, p.336-337) Rogers and Stanfield have also found that in general an innovator is likely to be educated and knowledgeable with a high income, to have a positive attitude towards change and high aspirations, and to be linked to external information sources of media and change agents. (Urban&Hauser 1993, p.503) Thus, on the grounds of aforementioned characteristics of different levels of innovativeness there has been drawn the third and forth hypothesis:

H3: A customer’s innovativeness is a significant factor influencing on a customer’s price perception and innovative customers are significantly higher educated and earn more than other customers.

H4: A customer’s intention to invest on mobile services indicates his/her price perception of mobile services.

Using bundle pricing sellers combine several of their products and offer the bundle at a reduced price or increased value. In value-added bundling a company offers to price sensitive customers an additional value of a kind that less price-sensitive buyers do not want. (Nagle&Holden 2002, p.246; Kotler 2001, p.375) By bundling services into larger blocs, consumers need to evaluate only one bloc of services rather than several items separately. Hence total costs are more transparent and total value is easier to perceive. From the basis of these observations, there can be come to a conclusion that through bundling of mobile services there can be modified service bundles which have higher cost transparency, greater perceived value, and lower price elasticity. And thus there is created fifth hypothesis:

H5: Through bundling mobile services there can be affected significantly on a customer’s price perception.

The formulated hypotheses are expected to be valid in all of the three segments. That is, the amount of usage of mobile services does not change a customer’s response to upper mentioned circumstances. It is thus assumed that customers, regardless the level of usage, are corresponding to hypothesized situations according to same logic.

4. results of study

In analysing the research results there is first described respondents by their demographical backgrounds. After having knowledge on basic characteristics of the respondents there is examined the research questions and hypotheses in detail. First by studying respondents’ price perception and factors influencing onto it; and then elaborating how respondents perceive acquisitions of mobile services through service bundles, and which factors influence on their preference for mobile service bundles.

The response rates and respondents’ basic characteristics by the customer segments are summarized in the table 2. Even though, in general, the response rate 25.9% (778) was left a bit lower than expected it was sufficiently high for our study. The response rates were also found to differ rather much between the segments. Response rate for prospective users’ segment was 31%, moderate users’ 25.7 %, and heavy users’ 21.1 %. The responsiveness was therefore clearly user segment related. In the segments that used less often mobile services had higher response rates. Answer to this notion should be obtained by examining customer segment more closely.

The segments differed also when examined respondents’ genders and ages. These variables seemed to be rather much different between the three segments. For example respondents’ gender distributed equally between male and female in prospective users’ segment, whereas in the two other segments the distributions were ¾ in favour of male respondents. One of the rather unexpected results occurred in educational backgrounds as in the all three user segments mean respondent possessed only a secondary level of education. Even though it was hypothesized, that customers in heavy users’ segment would be clearly higher educated than customers in other segments.

TABLE 2. Respondents characteristics by user segments.

	Heavy users	Moderate users	Prospective users
Response rate, % Tota 25.9 %, N 778	21.1 %, n 211	25.7 %, n 257	31 %, n 310
Male respondents	74.4 %	77.7 %	50 %
Female respondents	25.6 %	22.3 %	50 %
Respondents by age groups: - 18-24 - 25-34 - 35-49 - 50-64 Total	mean 18-24 29.8 % 38.9 % 20.7 % 8.2 % 100 %	mean 25-34 12.3 % 37.9 % 32.8 % 14.2 % 100 %	mean 25-34 12.6 % 20.1 % 32.4 % 24.6 % 100 %
Respondents’ mean annual income	20.001-30.000€	30.001-40.000€	20.001-30.000€

4.1. Background of heavy users

In addition to upper mentioned notions on respondents’ basic characteristics it was also found that respondents among heavy users’ segment were most often 18 to 34 years of age. This age distribution was rather much different if compared to other two segments as heavy users were significantly younger than users in other segments. But on the other hand this was expected before the study, which only confirmed that the most of the heavy users of mobile services are fairly young in comparison to prospective users. Nearly 68 percent of respondents in this segment were between 18 – 34 years of old. This further confirmed that “innovators” and “early adopters” are most often rather young customers, at least in mobile services business, who are also probably more willing to invest more on new technologies regardless of higher prices.

As the heavy users of mobile services were found to be respectively young compared to other groups, it reflects automatically also to respondents’ marital status. Thus, it was no surprise that users in this segment were most often unmarried (50.5) and only approximately 1/8 of respondents were married. This might also reflect the economical aspect of the respondents’ willingness to invest on new technologies. That is, as there is no family requiring continuous financial investments there is left more money on other interests.

TABLE 3. Distribution respondents’ marital status.

		N	%	Valid %
	Married	27	12.8	13.4
	Live-in	60	28.4	29.7
	Unmarried	102	48.3	50.5
	Widow	1	0.5	0.5
	Divorced	12	5.7	5.9
	No response	9	4.3
In total		211	100.0
Std. deviation		0.940

Contrary to what was expected before the questionnaire survey, educational level was unexpectedly low in heavy users’ segment (as mentioned earlier). Over 40 percent (40.5%) of the respondents in this segment possessed a secondary level education were as only 13 percent had university level degree. It can be thus stated that higher education does not mean higher interest in mobile services; rather it is other way round so that a person with higher education is more causes in taking in use new mobile services. However, it should be also kept in mind in this kind of surveys that there is always a possibility for systematic errors especially as the response rate in this segment was rather low, 21.1%. Meaning that users of higher education might have left the questionnaires unanswered.

TABLE 4. Respondents’ level of education

		N	%	Valid %
	Elementary school	24	11.4	11.5
	Trade school	16	7.6	7.7
	Vocational school	69	32.7	33.2
	Technical school	18	8.5	8.7
	Polytechnic	21	10.0	10.1
	University/collage	27	12.8	13.0
	High school graduate	31	14.7	14.9
	Other	2	0.9	1.0
	No response	3	1.4
	In total	211	100.0	100.0
Std. deviation		1.952

Income level was also mentioned to be one of the elements differentiating “innovators” from other consumers. Therefore it was expected that innovators or early adopters belong to higher income categories also in a case of mobile services business so that they are able to invest more on new technology services. But when examining the table below there can be recognized that heavy users are fairly average in income level. Mean annual income amongst heavy users’ segment was 20.001-30.000 euros and 72 percent earned 30.000 euros or less per annum.

TABLE 5. Respondents’ annual gross income

	N	%	Valid %
Below 10.000 €	33	15.6	16.5
10.001-20.000 €	54	25.6	27.0
20.001-30.000 €	59	28.0	29.5
30.001-40.000 €	25	11.8	12.5
40.001-50.000 €	14	6.6	7.0
50.001-60.000 €	4	1.9	2.0
60.001-70.000 €	4	1.9	2.0
70.001 € or more	7	3.3	3.5
No response	11	5.2
In total	211	100.0	100.0
Std. deviation	1.650

4.2. Background of moderate users

Amongst the moderate users’ segment gender was rather similarly biased towards male users as in the heavy users’ segment. More than ¾^th of the respondents in this segment were male, 78% versus 22%. In addition to gender, also distribution of education was rather similar in the both segments; the most of the respondents possessed a secondary level of education. Despite that for the gender or education parts there were no differences compared to heavy users, differences started to appear in other demographical variables.

Especially when observed age distribution in moderate users’ segment, there was noticed that respondents were apparently older than in heavy users’ case (see table 6). There appeared a clear change in age distribution from young adults to more middle aged ones and mean age was 25-34 years. Two largest groups were thus 25-34 and 35-49 years of age. These two groups received a proportion of over 70%. Young 18-24 years of age users were significantly diminished compared to heavy users. Their proportional representation was only 12 % as it was in heavy users’ case almost 1/3^rd.

TABLE 6. Distribution respondents’ age.

Years of age		N	%	Valid %
	Below 18	2	0.8	0.8
	18-24	31	12.1	12.3
	25-34	96	37.4	37.9
	35-49	83	32.3	32.8
	50-64	36	14.0	14.2
	65 or more	5	1.9	2.0
	No response	4	1.6
In total		257	100.0
Std. deviation		0.974

Along with sift to older users the respondents’ distribution of marital status was also changed. Users in this segment were more often in steady relationships as 40 percent of respondents were married (40.4%) and almost 30 percent were in live-in relationship (27.6%). The changes in age and marital status seemed to affect also to respondents’ income levels as mean income in this segment was 30.001-40.000 euros per annum (std. deviation 1.980).

TABLE 7. Distribution of respondents’ marital status.

		N	%	Valid %
	Married	101	39.3	40.4
	Live-in	69	26.8	27.6
	Unmarried	61	23.7	24.4
	Divorced	19	7.4	7.6
	No response	7	2.7
In total		257	100.0
Std. deviation		1.154

Speaking of income, the most visible change had occurred in lowest income levels which had diminished to half of the proportion that was in heavy users’ case (from 16.5% to 8.5%). The change had benefited the higher income levels which had increased their proportions with the equal proportion. Now income categories 10.001 – 40.000 € obtained 50.4% of all respondents when the same categories in heavy users’ segment was only 32%. In general, respondents’ were more often earning over 30.000 euros per year than heavy users. This however doesn’t mean that they’d had more money for investments on new mobile services as moderate users were often living in steady relationships, meaning that higher proportion of the income goes on family needs.

TABLE 8. Respondents annual gross income

		N	%	Valid %
	Below 10.000 €	21	8.2	8.5
	10.001-20.000 €	48	18.7	19.5
	20.001-30.000 €	87	33.9	35.4
	30.001-40.000 €	37	14.4	15.0
	40.001-50.000 €	23	8.9	9.3
	50.001-60.000 €	8	3.1	3.3
	60.001-70.000 €	7	2.7	2.8
	70.001-80.000 €	5	1.9	2.0
	80.001 € or more	10	3.9	4.1
	No response	11	4.3
	Total	257	100.0	100.0
Std deviation		1.875

4.3. Background of prospective users

In prospective users’ segment there was observed to be clearly higher proportion of female respondents than in other two segments. In this category respondents were equally often male and female (50%/50%). Users in this segment were also rather different in educational background. There was notably less vocational school graduates, 23.5% (vs. 33.5% in moderate segment), but at the same time the proportion of university level graduates had respectively increased with 10 percentage unit to 17.6% (vs. 7.9% in moderate segment). This kind of transition from lower education levels to higher ones was pretty much opposite to expectations; it was expected that this same alike transition would have happened in heavy users favour rather than prospective users’.

TABLE 9. Education level of respondents

		N	%	Valid %
	Elementary school	48	15.5	15.6
	Commercial school	29	9.4	9.4
	Vocational school	72	23.2	23.5
	Technical school	35	11.3	11.4
	Polytechnic	28	9.0	9.1
	University/collage	54	17.4	17.6
	High school graduate	31	10.0	10.1
	Other	10	3.2	3.3
	No response	3	1.0
	In total	310	100.0	100.0
Std. deviation		2.063

Even though education level was a surprise the age distribution and its tendency, from young heavy users to older prospective users had remained steady (table 10). Even though the mean respondent in this segment was between 25-34 years of age as in the previous segment there had occurred some notable changes. There was a clear movement from age group of 25-34 (37.9% à 20.1%) to age group of 50-64 (14.2% à 24.6%). In other parts proportional distribution of age was quite stable. Although there were differences in distribution of age, respondents’ marital status was rather similar to users in moderate segment. Nothing else had changed distinctly but the proportion of divorced which had risen from 7.6% to 14.2%.

TABLE 10. Age distribution of respondents.

Years of age		N	%	Valid %
	Below 18	4	1.3	1.3
	18-24	39	12.6	12.6
	25-34	62	20.0	20.1
	35-49	100	32.3	32.4
	50-64	76	24.5	24.6
	65 or more	28	9.0	9.1
	No response	1	0.3
Total		310	100.0
Std. deviation		1.196

Due to higher proportion of retired respondents average annual gross-income levels were naturally lower than in moderate users’ segment. The mean income was placed to income category of 20.001-30.000 euros per year; same as heavy users. This similarity between the incomes of heavy and prospective users might occur from the fact that in both segments there are consumers out of active working life (students and retirees). Therefore in this segment two lowest income categories received higher proportions than in other segments, totalling to 41.9%. These additional percentages were taken mainly from the income category of 20.001-20.000€ which had diminished from 35.4% to 24.5% compared to moderate segment.

TABLE 11. Annual cross-income of respondents

		N	%	Valid %
	Below 10.000 €	43	13.9	14.4
	10.001-20.000 €	82	26.5	27.5
	20.001-30.000 €	73	23.5	24.5
	30.001-40.000 €	40	12.9	13.4
	40.001-50.000 €	30	9.7	10.1
	50.001-60.000 €	12	3.9	4.0
	60.001-70.000 €	7	2.3	2.3
	70.001 € or more	11	3.5	3.7
	No response	12	3.9
	Total	310	100.0	100.0
Std deviation		1.741

4.4. Factor analyses

Kerlinger (1973, p.659) has described factor analyse as “the queen of analytic methods”. The general purpose of factor analysis is to summarize the information contained in a large number of variables into a smaller number of factors. Factor analysis is referred to a diverse number of techniques used to discern the underlying dimensions or regularity in phenomena. (Zikmund 1991, p.574) Factor analysis begins by constructing a new set of variables based on the relationships in the correlation matrix. While this can be done in a number of ways the most frequently used approach is principal components analysis. This method transforms a set of variables into a new set of composite variables or principals components that are not correlated with each other. These linear combinations of variables, called factors, account for the variance in the data as a whole. (Cooper&Emory 1995, p.538)

Factor analyse was used for structuring and analysing the variables which explain differences in price perceptions of mobile service users. Respondents were asked whether they perceive mobile services expensive or inexpensive. The factors were constructed according to factor loadings which suggest the number of factors obtained. If factor loadings are very high, it suggests that only one factor is needed. If they are near zero, it suggests that there are no common factors. If they are moderately high (e.g. around .50), it suggests that more factors may be needed. Nunnally also states that the factors with very small loadings (under .30) are not interpreted. (Nunnally 1967, p.292, 303) Usually approval limits of 0.3 and 0.5 are both regularly used. For this research the approval limit 0.3 was seen the most.

Another criterion for factors is called latent root or eigen-value criterion. The latent roots criterion holds that the amount of variation explained by each factor or latent root must be greater than one. That is, each factor should account for the variation in at least one variable if the factor is to be considered useful from a data summarization perspective. (Churchill 1995, p.972)

According to discussed criterions for creating factors there was obtained four factors for each segment, although factors were different between the segments. When the factors are finally created and qualified factors are found, these factors should be named on the bases of what the variables loading on a given factor seem to have in common (Churchill 1995, p.976).

The internal consistency of the factors was examined by Cronbach’s alpha. It was perceived suitable for this study as e.g. Carmines and Zeller have stated that for examining the internal consistency of statistical tests that include combined or multiple variables Cronbach’s alpha is most suitable. Alpha changes between 0 and 1, and the closer to 1 the more consistent is the meter (1976, p.46). And Nunnally (1967, p.226) has also noted that for basic research the sufficient level of reliability is .80.

4.4.1. Heavy users

In heavy users’ segment there were identified four factors according to the factor matrix below. These factors were approved by the both criterions discussed previously; they obtained factor loadings above 0.3, and eigen-values above one. All the four factors explained the differences in customers’ price perceptions of mobile services, but were measuring different aspects of price perception. For naming the factors, there was first scrutinized the variables arranged under the factors and then using the pre-assigned terms for the variables or assessing the common elements describing the variables.

Variables in Factor 1 were observed to measure a customer’s satisfaction to operator’s services. This factor obtained the highest explanation power by explaining 15 percent of total variance of price perception (rotated eigen-value 15.025%). Variables in factor 2 were measuring customers’ willingness or readiness to invest more on mobile services. It got the second highest explanation power with 10 percent (rotated eigen-value 10.176%) and was thus also explaining quite well the differences in price perceptions. Factor 3 can be stated to measure respondents’ innovativeness with explanation power of 8 percent (rotated eigen-value 8.606%). And factor 4 measured customers’ price sensitiveness which explanation power was lowest out of the four qualified factors, only 4 percent (4.277%).

TABLE 12. Factor Matrix(a)

	Factor 1	Factor 2	Factor 3	Factor 4
Satisfaction to operator’s services
My problems are taken care of delicately and confidentially.	.739
I get my problems solved at the time important to me.	.725
Support services offered by my teleoperator fulfil my service needs.	.672
Software and services needed in usage of mobile internet are well up-to-date.	.669
Speed of data transfer is equivalent to promises of my teleoperator.	.618
My teleoperator is never too busy for answering to my questions.	.614
I trust to confidentiality and security of data transfer under my teleoperator.	.614
I am remembered with personal benefits by my operator.	.515
I am satisfied with Internet sites of my teleoperator.	.514
New Internet connection is installed in a promised schedule.	.437
Readiness to invest more on mobile servces
I could obtain mobile services with higher data speed.		.848
There would be a possibility to get more versatile service packages than usual.		.804
There was a possibility to tailor service packages I prefer by myself.		.783
Services are delivered me without delay.		.625
Innovativeness
Usage of new technology products is natural and enjoyable.			.696
I trust in usage of electronic channels in dealing with daily affairs.			.669
I am trying actively to learn new things and methods.			.639
I am well aware of different service channels.			.554
There can be increased service quality with new technologies.			.421
I appreciate more comfort improving devices than personal service.			.351
Reforms and innovations bring comfort improving devices rather than better personal service.			.316
Price sensitiveness
I am not troubled even though I have to pay more to acquire new technology services or products.				.626
I am willing to pay more in order to get services and products of higher quality.		.432		.572
Price and costs are very important factors in considering to acquire new products and services.				-.422
Expensiveness of mobile services is the main barrier in increasing the usage of mobile services.				-.323
Cronbach’s alpha	0.8389

Extraction Method: Principal Axis Factoring. Rotation Method: Varimax with Kaiser Normalization.

a Rotation converged in 5 iterations.

With the factors presented above there can be explained 38 percent of the variance in price perception after rotation. Especially innovativeness (8.6%) and price sensitiveness (4.3%) factors received rather low factor loadings if compared to the other two factors. The factors were also correlating to some extent with each other, especially the factors 2 and 4 (r .305). The reliability and consistency of the factors was confirmed by Cronbach’s alpha which reached the satisfactory level of .84.

TABLE 13. Factor Correlation Matrix

Factor	1	2	3	4
1	1.000	.230	.262	.092
2	.230	1.000	.271	.305
3	.262	.271	1.000	.118
4	.092	.305	.118	1.000

Extraction Method: Principal Axis Factoring.

Rotation Method: Promax with Kaiser Normalization.

4.4.2. Moderate users

The same factor analyse was used also for structuring and analysing the variables explaining the differences in price perceptions of mobile services in moderate users’ segment. The factor analyse produced also four qualified factors. The factor loadings were in general quite sufficient in all factors. Only exception was the fourth factor in which the loadings were divided into two separate groups. But as those variables were observed to measure the same phenomenon there was no need to divide those variables into two separate factors or reject the variables with low factor loadings.

The factors consisted of the same alike variables than in the previous segment and were thus also named accordingly. Factor 1 measured customers’ satisfaction to operator’s services. Factor 2 measured respondents’ innovativeness. Factor 3 was measuring readiness to invest more on mobile services. And the last factor measured respondents’ price sensitiveness.

TABLE 14. Factor matrix for price perception of mobile services

	Factor
	1	2	3	4
Satisfaction to operator’s services
I get my problems solved at the time important to me.	.791
My problems are taken care of delicately and confidentially.	.721
Support services offered by my teleoperator fulfil my service needs.	.701
My teleoperator is never too busy for answering to my questions.	.645
I am satisfied with Internet sites of my teleoperator.	.614
Software and services needed in usage of mobile internet are well up-to-date.	.599
I trust to confidentiality and security of data transfer under my teleoperator.	.588
Speed of data transfer is equivalent to promises of my teleoperator.	.563
I am remembered with personal benefits by teleoperator.	.540
New Internet connection is installed in a promised schedule.	.454
Innovativeness
I am trying actively learn new things and methods.		.742
There can be increased service quality with new technologies.		.720
Usage of new technology products is natural and enjoyable.		.667
I trust in usage of electronic channels in dealing with daily affairs.		.663
I am well aware of different service channels.		.633
Reforms and innovations bring comfort improving devices rather than better personal service.		.523
I appreciate more comfort improving devices than personal service.		.489
Readiness to invest more on mobile services
I could obtain mobile services with higher data speed than usual.			.859
There was a possibility to tailor service packages I prefer by myself.			.854
There would be a possibility to get more versatile service packages than usual.			.802
Services are delivered me without delay.			.654
Price sensitiveness
I am willing to pay more to get more versatile service possibilities.				.875
I am willing to pay more in order to get services and products of higher quality.				.809
I am not troubled even though I have to pay more to acquire new technology services or products.				.677
Price and costs are very important factors in considering to acquire new products and services.				.372
Expensiveness of mobile services is the main barrier in increasing the usage of mobile services.				.337
Cronbach’s alpha	0.8585

Extraction Method: Principal Axis Factoring. Rotation Method: Varimax with Kaiser Normalization.

a Rotation converged in 6 iterations.

With the four factors presented there was explained 42 percent of the variance in price perception. The factor 1 gained the best explanation power of 14 percent which was quite much above the values of the other three factors. Factors were also correlating to some extent with each other. The highest correlations were between the factors 2 and 4 (0.305), and lest correlating factors were 1 and 4 (0.092). Factor analyse provided also a level of consistency among the variables that was considered sufficient and reliable as Cronbach’s alpha was approximately 0.86.

TABLE 15. Factor Correlation Matrix

Factor	1	2	3	4
1	1.000	.250	.314	.208
2	.250	1.000	.224	.067
3	.314	.224	1.000	.281
4	.208	.067	.281	1.000

Extraction Method: Principal Axis Factoring.

Rotation Method: Promax with Kaiser Normalization.

4.4.3. Prospective users

Factor analyses in this segment produced us also with four qualified factors which explained the differences in perceived prices of mobile services. The factor loadings in general were rather high compared to other two segments. There was also one difference in naming the factors as the factor 4 was named innovation resistance, replacing thus the readiness to invest on mobile services. The innovativeness was thus divided into two factors.

The factor 1 measured respondents’ satisfaction to operator’s services and obtained highest explanation power with rotated eigen-value of 16.1 percent out of total variance (16.119%). Factor 2 measured users’ innovativeness and received second highest explanation power with 12 percent (12.020%). Factor 3 was measuring respondents’ price sensitiveness with explanation power of 5.6 percent (5.630%). And factor 4 measured users’ innovation resistance which had equal explanation power with the third factor 5.6 percent (5.565%).

TABLE 16. Rotated Factor Matrix for price perception

	Factor
	1	2	3	4
Satisfaction to operator’s services
My teleoperator is never too busy for answering to my questions.	.741
My problems are taken care of delicately and confidentially.	.725
Support services offered by my teleoperator fulfil my service needs.	.719
I get my problems solved at the time important to me.	.678
New Internet connection is installed in a promised schedule.	.633
I trust to confidentiality and security of data transfer under my teleoperator.	.565
I am satisfied with Internet sites of my teleoperator.	.536
Speed of data transfer is equivalent to promises of my teleoperator.	.535
I am remembered with personal benefits.	.473
Software and services needed in usage of mobile internet are well up-to-date.	.334
Innovativeness
I am trying actively learn new things and methods.		.696
I trust in usage of electronic channels in dealing with daily affairs.		.680
I am well aware of different service channels.		.639
I appreciate more comfort improving devices than personal service.		.613
Usage of new technology products is natural and enjoyable.		.612
There can be increased service quality with new technologies.		.576
Reforms and innovations bring comfort improving devices rather than better personal service.		.467
Price sensitiveness
Price and costs are very the most important factors in considering to acquire a new service.			.576
High costs are the main barrier in starting the usage of mobile services.			.499
I would be willing to pay extra to get more versatile service possibilities.			.429
Innovation resistance
I prefer to deal my affairs with customer service person.				.637
I don’t like changes in my normal methods.				.623
I don’t like computer based or automated devices or services.				.478
I lose my interest to new services when I get to know that they are more expensive than old ones.				.306
Cronbach’s alpha	0.8359

Extraction Method: Principal Axis Factoring. Rotation Method: Varimax with Kaiser Normalization.

a Rotation converged in 5 iterations.

With these factors there can be explained 39 percent of the variance in price perception (after rotation). Satisfaction to operator’s services (16%) and users’ innovativeness (12%) factors obtained the best explanation powers. Whereas price sensitivity and innovation resistance factors were both explaining only 5.6 percent of the total variance in price perception. The factors were also correlating to some extent with each other. The highest correlations were between factors 1 and 3 (0.425), and lest correlating factors were 2 and 4 (-0.051). The reliability and consistency of the factors was confirmed by Cronbach’s alfa 0.84 as it exceeded the critical value 0.8.

TABLE 17. Factor Correlation Matrix

Factor	1	2	3	4
1	1.000	.280	.425	.173
2	.280	1.000	.197	-.051
3	.425	.197	1.000	.286
4	.173	-.051	.286	1.000

Extraction Method: Principal Axis Factoring.

Rotation Method: Promax with Kaiser Normalization.

In summary, it can be stated that satisfaction to operator’s services and a customer’s innovativeness were the most important factors explaining the variance in price perception of mobile services. Thus, a company can best influence to price perception through increasing the satisfaction to services and by differentiating the services according to customers’ innovativeness levels. Rather surprising was that users’ price sensitiveness had the second lowest explanation power in variance of perceived prices.

4.5. Price perception

From the factor analyses we obtained four factors for each segment which explained the differences in variances of respondents’ price perceptions. In this chapter we have further analyzed the classified factors and their relationships with price perceptions. We have also taken a brief glance at the levels of perceived mobile service prices in each segment.

To further elaborate and analyse the relationships of these factors and customers’ price perceptions, we have applied regression and correlation analysis. Regression analyses technique is used as it attempts to predict the values of a continuous, interval-scaled dependent variable from the specific values of the independent variable. The applied regression analyses in linear regression investigates a straight-line relationships of the type Y=a+βX. (Zikmund 1991, p.548) In other words, with regression analyses there is investigated relationships of one or more variables towards dependent variable. Regression model can be also perceived as causal model according to Suppes’ criteria: If x and y don’t co-vary, then β_x=0. If x is the predictor and y the dependent variable, then in the regression model x precedes y. Therefore it includes precedence relationships as Suppes requires. And finally, β_x provides a statistical version of control. (Cohen et al. 1994)

For testing the goodness of data and a model there are two basic tests: t-test and f-test. In bivariate regression, t and F tests produce the same results since t² is equal to F. In multiple regression analyses, the F-test has an overall role for the model, and each of the independent variables is evaluated with a separate t test. (Cooper&Emory 1995, p.498) From these choices we have applied F-test or analyses of variance summary table for comparing relative magnitudes of the mean square. There is also used squared multiple correlation coefficient, R², which reflects the proportion of variation explained by the regression line. (Zikmund 1991, p.548) R²varies between 0-1 and the closer the value is to 1 the higher is the explanation power of the variable.

Even though the segments differed rather much in usage frequencies of mobile services and demographic characteristics, perceived prices of mobile services were surprisingly uniform. Basically users in all segments perceived mobile services quite expensive and were stating that the expensiveness is an important barrier for increasing the usage frequency.

4.5.1. Heavy users

For acquiring accurate knowledge on how respondents perceived prices of mobile services there was enquired whether they perceived mobile services expensive or inexpensive to use (figure 9). The result was rather plain as mean response was 2 (std. deviation 2.016) meaning that a majority of respondents perceived mobile services to be rather expensive. Out of all responses 51.7 percent considered mobile services expensive (1) or very expensive (0). All alternatives were however chosen and received some popularity. The tendency was though rising towards the expensive end of the scale.

FIGURE 9. Price perception of mobile services in heavy users’ segment.

By conducting a regression analysis there is investigated the relationships of the factors found in the factor analyses towards respondents’ price perception. We are especially interested in acquiring knowledge on the factors that have a significant relationship with respondents’ price perception. We are also interested in knowing how well the factors explain customers’ price perceptions; and thus try to find out the most influential factors.

In heavy users’ segment it was found that out of the original four factors only three of them were had a significant influence on respondents’ price perceptions. These factors were price sensitivity, innovativeness, and readiness to invest more on mobile services. Surprising was that satisfaction to operator’s services was not found to have significant influence on respondents’ price perception.

Price sensitiveness factor appeared to have strongest influence on respondents’ price perception as the multiple correlation coefficient was highest amongst the three factors, R=.462. It was also noticed, from beta coefficients that between the independent variables there was only slight inter-correlations between each other; thus it contributed also a satisfactory level multiple correlation.

Price sensitivity did also explain respondents’ price perception quite prominently, by 21 percent (R².213). And when taken into account the number of independent variables used in the analyses the explanation power was decreased only slightly to 20 percent (adjusted R² .197). The significance was confirmed by ANOVA which resulted a significant F-value 13.298 as the critical value at 0.01 level was 3.32 (v₁ 4 and v₂ 196).

TABLE 18. Regression analyses for price sensitivity and price perception.

Model	R	R²	Adjusted R²	Std. Error of the Estimate
1	.462(a)	.213	.197	1.817

a Predictors: (Constant), Price sensitivity

ANOVA

Model		Sum of Squares	df	Mean Square	F	Sig.
1	Regression	175.677	4	43.919	13.298	.000(a)
	Residual	647.318	196	3.303
	Total	822.995	200

a Predictors: (Constant), Price sensitivity

b Dependent Variable: Price perception

The second highest influence on respondents’ price perception of mobile services was found to has on a person’s level of innovativeness, with multiple correlation value of .314. Even though the multiple correlation coefficient was quite sufficient the explanation power remained on a rather low level explaining only 10 percent (R² .099) of the variance in price perception. And if using adjusted multiple correlation then the explanation power was only 7 percent (adjusted R² .066). From beta coefficients it was however noticed that the lowness of the explanation power was partly resulted from the correlation between few independent variables. But, as the independent variables were measuring a respondent’s innovativeness (same phenomenon) it was not a surprise to notice the correlation. The relationship was significant at the 0.01 level as calculated F-distribution was 3.008 compared to critical F-value 264 (v₁ 7 and v₂ 192).

TABLE 19. Regression analyses for innovativeness and price perception.

Model	R	R²	Adjusted R²	Std. Error of the Estimate
1	.314(a)	.099	.066	1.936

a Predictors: (Constant), Innovativeness

ANOVA(b)

Model		Sum of Squares	df	Mean Square	F	Sig.
1	Regression	78.960	7	11.280	3.008	.005(a)
	Residual	719.995	192	3.750
	Total	798.955	199

a Predictors: (Constant), Innovativeness

b Dependent Variable: Price perception

There was also found that a person’s readiness to invest on mobile services was significantly correlating with price perception even though the correlation coefficient was lowest compared to other factors, R .252. According to the criterion that a correlation coefficient under 0.3 can be assessed to be moderate if amount of responses is over 50 (Tähtinen&Isoaho 2001, p.108), this factor was taken account in explaining a respondent’s price perception. Investment readiness can be seen to be rather close to satisfaction to operator’s services as –factor. It could be hypothesised that a person that is not happy to operator’s services would not be willing to invest more money on any operator’s services. Explanation power of investment factor was only 6 percent (R².063), and when adjusting the explanation power with other factors it fell to 4 percent (adjustedR².044). Despite of these low values a person’s willingness to invest on mobile services was found to be significant at 0.05 level as calculated F-value 3.334 was above the critical value 2.37 (v₁ 4 and v₂ 192).

TABLE 20. Regression analyses for readiness to invest on mobile services versus price perception.

Model	R	R²	Adjusted R²	Std. Error of the Estimate
1	.252(a)	.063	.044	1.979

a Predictors: (Constant), Readiness to invest more on mobile services.

ANOVA(b)

Model		Sum of Squares	df	Mean Square	F	Sig.
1	Regression	52.239	4	13.060	3.334	.011(a)
	Residual	771.741	197	3.917
	Total	823.980	201

a Predictors: (Constant), Readiness to invest on mobile services.

b Dependent Variable: Price perception.

In addition to presented three main factors influencing on respondents’ price perception of mobile services, there was distinguished also some demographic variables that had significant relationship with price perception. However, it was noticed that in general demographic variables predicted quite poorly a respondent’s price perception. Only marital status and number of adults in household received significant relationship with price perception. The number of adults in household correlated with price perception negatively with the rate of -.205 (r -.205**; p<.01). Meaning that, the more adults in the household the more expensive mobile services were perceived.

Marital status was also found to correlate even though the correlation coefficient was low even when the amount of responses was taken into account (r .142*; p<.05). Especially married respondents seemed to possess a view that usage of mobile services is very expensive with mean response 0.92 (0 expensive, 6 inexpensive) and standard deviation 0.940. Divorced respondents seemed to perceive mobile services only slightly expensive although the deviation of responses was higher than in other groups.

There were also measured correlations between price perception and channel choices; delay acceptance; and respondents’ number of customerships with more than one operator. It was found that frequency of usage of service machines had a moderate and positive correlation, r = .165* (p<.05). This result is understandable as nowadays service machines have second highest service charges after personal services in Finland. And even though the correlation coefficient was rather low it can be considered to be moderate as Tähtinen & Isoaho. There was also found a positive correlation between price perception and number of customership (r= .209**; p<.01). Thereby, the more mobile interfaces a customer had and the more often he/she used service machines, the cheaper was the perceived price of mobile services.

TABLE 21. Correlations for perceived price versus usage frequency of service machines; and amount of mobile phone connections.

		Price perception
Usage frequency of service machines	N	197
	Pearson Correlation	.165(*)
	Sig. (2-tailed)	.020
Number of different customerships with different operators	N	204
	Pearson Correlation	.209(**)
	Sig. (2-tailed)	.003

* Correlation is significant at the 0.05 level (2-tailed).

** Correlation is significant at the 0.01 level (2-tailed).

Somewhat surprising was that only marital status and size of the household were significantly explaining differences in variance of price perceptions out of all demographic variables. There could have been expected that also income level and education could have influence on perceived prices. But instead price perceptions seemed to be rather same in all income/education/occupation levels. Therefore, differences ought to be looked for from other usage and user related variables.

To sum up, the factors and variables significantly explaining and predicting respondents’ price perception of mobile services in heavy users’ segment are constructed into a figure below. There is depicted the multiple correlation coefficients between price perception and the factors found to have significant influence. Demographic variables are shown in Pearson’s correlation coefficients.

FIGURE 10. The correlation coefficients for the factors with significant influence on respondents’ price perception of mobile services

4.5.2. Moderate users

Respondents’ price perception in moderate users’ segment was measured by enquiring “how they perceived mobile services compared to other service channels”; expensive or inexpensive. Responses in total were quite clear as over 50 percent of respondents considered mobile services to be more expensive than services in other channels. The mean response was 3.08 with rather high standard deviation 2.780 representing that respondents were varying very much in perception of mobile service prices. Thus, as the figure below shows price perception of mobile services is not well established among the moderate users (almost 28 percent responded “don’t know” whether mobile service are expensive or inexpensive). And this gives opportunities for business actors to influence on customers’ price perceptions by pricing and other marketing activities.

FIGURE 11. Customers’ price perception on mobile services in moderate users’ segment.

In a case of moderate users’ segment it was found that all the four factors created in the factor analyses were explaining significantly the differences in variance of respondents’ price perceptions: price sensitivity, satisfaction to operator’s services, readiness to invest more on mobile services, and innovativeness. The three factors were the same than in the previous segment but the satisfaction factor made its first appearance in this segment.

Notable was also that in this segment the price sensitivity did appear to be the most important factor influencing to perceived prices of mobile service users. The correlation coefficient was higher than in heavy users segment R= .486 and the explanation power was also strong 24 percent (R².236). And when taken account the number of variables the adjusted explanation power was decreased to approximately 23 percent (adjusted R².226). The relationship was significant at 0.01 level as calculated F-value exceeded the critical value 3.78 (v₁ 3 and v₂ 235).

TABLE 22. Regression analyses for price sensitivity versus price perception.

Model	R	R²	Adjusted R²	Std. Error of the Estimate
1	.486(a)	.236	.226	2.434

a Predictors: (Constant), Price sensitivity

ANOVA(b)

Model		Sum of Squares	df	Mean Square	F	Sig.
1	Regression	429.749	3	143.250	24.178	.000(a)
	Residual	1392.310	235	5.925
	Total	1822.059	238

a Predictors: (Constant), Price sensitivity; b Dependent Variable: Price perception

The second most important factor was the satisfaction to operator’s services with the correlation coefficient R= .327. It was found to be significant at 0.01 level as the F-value 2.665 was over the critical 2.32 (v₁ 10 and v₂ 222); although the marginal was rather small. The factor explained approximately 11 percent (R² .107) of the variance in a customer’s price perception. Noteworthy was also the rather weak adjusted explanation power, 6.7 percent (adjusted R².067) implying the high amount of independent variables included into the analyses. The explanation power was quite weak compared to the price sensitivity factor. But if ignoring that, the correlation is rather high and satisfaction factor is therefore essentially influencing and predicting persons’ price perceptions in mobile services business.

TABLE 23. Regression analyses for price sensitivity versus price perception.

Model	R	R²	Adjusted R²	Std. Error of the Estimate
1	.327(a)	.107	.067	2.673

a Predictors: (Constant), Satisfaction to operator’s services

ANOVA(b)

Model		Sum of Squares	df	Mean Square	F	Sig.
1	Regression	190.394	10	19.039	2.665	.004(a)
	Residual	1585.889	222	7.144
	Total	1776.283	232

a Predictors: (Constant), Satisfaction to operator’s services; b Dependent Variable: Price perception.

The third factor having a significant relationship with price perception according to ANOVA-table was the readiness to invest on mobile services. The significance was confirmed at the 0.01 level as F-value 6.518 was rather clearly over the critical level 3.32 (v₁ 4 and v₂ 236). The correlation was quite alike with the satisfaction factor, R= .315; only slightly under. Also the explanation power was naturally approximately at the same level, explaining 10 percent of the variance in price perception. But when recognizing the amount of independent variables that the multiple correlation coefficient was constructed, the adjusted explanation power was over 8 percent (adjusted R² .084).

TABLE 24. Regression analyses for price sensitivity versus price perception.

Model	R	R²	Adjusted R²	Std. Error of the Estimate
1	.315(a)	.099	.084	2.659

a Predictors: (Constant), Readiness to invest more on mobile services.

ANOVA(b)

Model		Sum of Squares	df	Mean Square	F	Sig.
1	Regression	184.317	4	46.079	6.518	.000(a)
	Residual	1668.487	236	7.070
	Total	1852.805	240

a Predictors: (Constant), Readiness to invest more on mobile services.

b Dependent Variable: Price perception.

The forth and last significant factor influencing on mobile services customers’ price perceptions was the level of innovativeness. The multiple correlation coefficient was rather low, R= .280, which was the lowest among the four factors. But still, it was considered to be significant as the number of responses was so high. The significance was confirmed at 0.01 level with F-value of 2.786 as it was a bit over the critical 2.64 (v₁ 7 and v₂ 229). The explanation power was also left to a lower level than in other factors only 8 percent (R².078). The explanation power was further decreased to 5 percent as the independent variables were taken into account as the variable number was rather high.

TABLE 25. Regression analyses for price sensitivity versus price perception.

Model	R	R²	Adjusted R²	Std. Error of the Estimate
1	.280(a)	.078	.050	2.704

a Predictors: (Constant), Innovativeness

ANOVA(b)

Model		Sum of Squares	df	Mean Square	F	Sig.
1	Regression	142.622	7	20.375	2.786	.008(a)
	Residual	1674.551	229	7.312
	Total	1817.173	236

a Predictors: (Constant), Innovativeness.

b Dependent Variable: Price perception.

When respondents’ price perceptions were examined according to gender it was noticed that there was significant differences in price perception between male and female respondents (r=.273**, p<.01). Female respondents perceived mobile service prices to be less expensive than male respondents; which was rather unexpected. Females’ mean response was 4.77 while male users mean was only 2.75. The difference was also found to be significant at 0.01 level as calculated F-value was 27.231 (v₁ 1 and v₂ 230). Gender appears to explain approximately 11 percent of the variance in price perception (eta² .106).

Male Female

FIGURE 12. Perceived prices of mobile services by gender.

In addition to gender also income seemed to correlate with customers’ price perceptions. The correlation was negative and rather weak even at this amount of responses, r -.142 (p<.05). The correlation direction was rather unexpected as it meant that higher the income level the more expensive is the perceived prices of mobile services. Amongst other demographic variables there was found no significant correlations with price perception.

I’d like to present very briefly an rather interesting relationship found through spearman’s correlations, eve though it is a bit out of the intended nomological net. The finding was significant as there was found rather high correlations and explanation powers between pricing and price perception variables. Pricing was found to correlate strongly with price perception. Persons who were most satisfied with the charging method of their mobile services were also those who perceived prices most inexpensive, r_s.547 (p< .001). And if taking squared Spearman’s rho (r_s² .329) we noticed that satisfaction to charging method explained 33 percent of the variance in price perception. Thus a satisfaction to pricing method, by it-self, explains 1/3^rd of the variance in price perception among moderate users’ segment.

Also the preference for different charging methods did correlate quite well with respondents’ price perceptions. The table below shows that persons who perceived pricing the most inexpensive did also prefer pricing methods which were based most directly to actual usage. Therefore, users who perceive mobile services most expensive could be most willing to take in to use different kind of fixed or bundle pricing methods.

TABLE 26. Nonparametric correlations for charging methods and perceived prices.

			Price perception
Spearman's rho	I am satisfied with the practice mobile services are charged at the moment.	Correlation Coefficient	.574(**)
		Sig. (2-tailed)	.000
		N	241
	What do you think on following charging methods of mobile services? According to actual usage	Correlation Coefficient	.309(**)
		Sig. (2-tailed)	.000
		N	226
	Fixed price	Correlation Coefficient	.193(**)
		Sig. (2-tailed)	.003
		N	234
	Service based	Correlation Coefficient	.323(**)
		Sig. (2-tailed)	.000
		N	216
	Bundle price	Correlation Coefficient	.241(**)
		Sig. (2-tailed)	.000
		N	216
	Separately from services and data transfer	Correlation Coefficient	.317(**)
		Sig. (2-tailed)	.000
		N	217
	Time-based pricing	Correlation Coefficient	.291(**)
		Sig. (2-tailed)	.000
		N	220
	What is the method your mobile service usage is charged at the moment?	Correlation Coefficient	.129(*)
		Sig. (2-tailed)	.047
		N	236

** Correlation is significant at the 0.01 level (2-tailed); * Correlation is significant at the 0.05 level (2-tailed).

In the case of moderate users’ segment there was found the four discussed factors affecting significantly to respondents’ price perception of mobile services: price sensitivity, satisfaction to operator’s services, readiness to invest more on mobile services, and innovativeness. In addition to these factors there was found two demographic variables that correlated significantly with perceived price of the respondents’: gender and income. As the pricing related variables in themselves were out of the intended nomological net they are not presented in the figure 13. The figure depicts the multiple correlation coefficients between price perception and the significant factors affecting to it; demographic variables are shown with Pearson’s correlation coefficients.

FIGURE 13. The correlation coefficients of the factors with significant influence on respondents’ price perception of mobile services

4.5.3. Prospective users

Price perception of mobile service was examined in the prospective users’ segment with questioning whether they perceived mobile services expensive or inexpensive. In this segment proportion of those users that perceived mobile services very expensive was much higher than in the other two segments, 33.7%. But on other parts the scale, presented in the figure 14, was quite alike with the other segments; only don’t know option had received a bit higher proportion (10.1%). Mean of responses was 2.2 with standard deviation 2.39 showing high differences in price perception between respondents. Also the “don’t know” responses received rather high proportion as seen in the figure below. Price perception among prospective users was thus rather equal with heavy users but recognizably higher than in moderate users’ segment.

FIGURE 14. Distribution of price perception of mobile services.

There was examined the factors expected to have significant relationship with price perception through regression analysis. Similarly with the heavy users’ segment there was found only three significant factors influencing to respondents’ price perception: satisfaction to operator’s services, price sensitivity, and innovation resistance. In examining the factors there was noticed few differences compared to the other two segments. First of all the level of multiple correlation coefficient of the satisfaction factor was the highest amongst the factors in this segment. Its importance has been increasing along with the segments. And second, the innovation factor was replaced by innovation resistance factor. It is somewhat logical as the prospective segment is consisted of consumers that are expected to be least innovative; and of consumers that are expected to take new services in use after the consumers in other two segments.

In this segment the satisfaction to operator’s services did appear to be the most important factor influencing to respondents’ price perception (unlike the other segments). The significance at the 0.01 level was confirmed as F-value 2.606 was over the critical one (2.32; v₁ 10 and v₂ 268). The multiple correlation coefficient was rather low 0.298 if compared to the most significant factors in the other two segments. Thus also the explanation power was left to a rather low level, explaining only 9 percent of the variance in price perception (R².089). And when taken account the number of independent variables the explanation power decreased to 5.5 percent (indicating the high number of independent variables included into the factor).

TABLE 27. Regression analyses for satisfaction to operator’s services versus price perception.

Model	R	R²	Adjusted R²	Std. Error of the Estimate
1	.298(a)	.089	.055	2.282

a Predictors: (Constant), Satisfaction to operator’s services

ANOVA(b)

Model		Sum of Squares	df	Mean Square	F	Sig.
1	Regression	135.682	10	13.568	2.606	.005(a)
	Residual	1395.401	268	5.207
	Total	1531.082	278

a Predictors: (Constant), Satisfaction to operator’s services; b Dependent Variable: Price perception.

The second highest correlation scores received, not surprisingly, price sensitivity which also in this segment was found to correlate significantly with respondents’ price perceptions. The significance was confirmed at 0.01 level as F-value was 4.258 and exceeded the critical value 3.32 (v₁ 4 and v₂ 228). The correlation coefficient was however left to a low level, only .241, and thus caused also the low level of explanation power. The price sensitivity explained only 6 percent of the variance in price perception. And when converting it to a comparable form the explanation power was decreased to a bit over 4 percent (adjusted R² .044). The multiple correlation coefficient was still moderate but the explanation power was very weak and its usefulness is thus under question.

TABLE 28. Regression analyses for satisfaction to operator’s services versus price perception.

Model	R	R²	Adjusted R²	Std. Error of the Estimate
1	.241(a)	.058	.044	2.345

a Predictors: (Constant), Price sensitivity

ANOVA(b)

Model		Sum of Squares	df	Mean Square	F	Sig.
1	Regression	93.636	4	23.409	4.258	.002(a)
	Residual	1523.020	277	5.498
	Total	1616.656	281

a Predictors: (Constant), Price sensitivity; b Dependent Variable: Price perception

The third and last formally significant factor was innovation resistance (compared to innovativeness in other segments). The significance was confirmed at 0.05 level as the calculated F-value 3.246 exceeded the critical value 2.37 (v₁ 4 and v₂ 228). The correlation coefficient R=.208 was even lower than in other two factors which also affected directly to the explanation power. The explanation power 4 percent is rather meaningless and of little use.

Thus as all of the factors received rather low explanation powers it should be questioned whether these factors are useful in predicting customers’ price perception in this segment. More likely there could be found some other factors that are more accurate for describing how prospective customers price perception are formed. We are therefore hoping to find some significant correlations amongst the demographic variables.

TABLE 29. Regression analyses for satisfaction to operator’s services versus price perception.

Model	R	R²	Adjusted R²	Std. Error of the Estimate
1	.208(a)	.043	.030	2.348

a Predictors: (Constant), Innovation resistance.

ANOVA(b)

Model		Sum of Squares	df	Mean Square	F	Sig.
1	Regression	71.579	4	17.895	3.246	.013(a)
	Residual	1587.834	288	5.513
	Total	1659.413	292

a Predictors: (Constant), Innovation resistance; b Dependent Variable: Price perception.

Despite of the expectations amongst the demographic variables there was found only one significantly correlating variable. Respondents’ income level was found to have a positive relationship with a person’s price perception. But the correlation was rather minimal and thus is not good predictor of a person’s price perception either, r= .142* (p<.05). The higher the income level the less expensive a person perceives mobile services.

TABLE 30. Correlation for price perception and usage frequency of computer while mobile phone as modem.

		Price perception
Annual gross income	Pearson Correlation	.142(*)
	Sig. (2-tailed)	.017
	N	286

* Correlation is significant at the 0.05 level (2-tailed).

To sum up, in this prospective users’ segment there was found only three significant factors correlating with price perception. The factors in this segment were partly different compared to other segments. The highest correlation scores obtained the satisfaction factor with 0.298. This was followed by price sensitivity and innovation resistance factors which both received rather low correlation score. In addition to the factors there was found one demographic variable, income level, which correlated significantly but weakly with price perception.

There was inspected also variables outside the nomological net but, unlike to moderate segment, there was found no significantly correlating variables. Therefore, the formation of mobile service price perception among prospective users was left somewhat open for future studies as this study did not provide any clear answer to it. Thus, there should be studied other elements relating to prospective mobile service users and their characteristics to be able to explain which factors have a significant influence on a formation of their perception on mobile service prices.

FIGURE 15. The correlation coefficients of the factors with significant influence on respondents’ price perception of mobile services

Next there is examined the factors individually to define the variables that have a significant influence on them. It is also important to obtain a knowledge how there could be affected to customers’ price perceptions indirectly. Therefore we have next provided a detailed description on the variables having significant relationship with the factors identified in this chapter.

4.6. Innovativeness of mobile service users

Consumers’ innovativeness has been recognized to be one of the most important factors determining consumers’ phase of adoption of new products or services. Therefore we have in this study expected that a person’s innovativeness probably influences also on one’s price perceptions. We have measured innovativeness with the variables shown in the factor analyses. And from that basis we have conducted correlation analysis and analysis of variance to find out whether innovativeness is influenced by some of the demographic and other user or usage related variables.

When observing innovativeness between the three segments it was noticed that differences between the groups were rather small. Mean responses were quite similar; only minor differences existed. The most innovative and comfortable with new technologies seemed to be the users in moderate segment with a mean of 4.94 (std. deviation 1.293). The least innovative segment was prospective users with a mean of 4.35 (std. deviation 1.611). Notable though was that the higher was the mean innovativeness the lower was the standard deviation as seen from the figure 16.

FIGURE 16. Respondents’ mean innovativeness by user segments

4.6.1. Heavy users

When inspected the demographic variables and their ability to explain differences in the variance of respondents’ innovativeness, there was expected that most of the demographic variables might be significant. However, through correlation tests there was found no demographic variables significantly correlating with a customer’s innovativeness. But instead, analyses of variance brought out that income level was significantly explaining those differences, although the explanation power was rather low. Respondents’ annual gross income explained differences in innovativeness by 7 percent (eta² .069). Mean innovativeness among the respondents answered to this question was 5.70 (std. deviation 1.316). Innovativeness seemed to rise towards both ends of income scale: users in lowest (mean 6.00) and highest (mean 6.50) income classes perceive to be most innovative. The mean 5.70 was also clearly higher than the average mean of the all respondents in this segment which was 4.78 (std. deviation 1.367). Indicating that the customers responded to this question was more innovative in average in this segment.

TABLE 31. ANOVA for respondents’ innovativeness versus annual gross income.

			Sum of Squares	df	Mean Square	F	Sig.
Innovativeness versus income	Between Groups	(Combined)	23.137	5	4.627	2.729	.021
	Within Groups		310.271	183	1.695
	Total		333.407	188

There was also found that usage of fixed internet connection was significantly explaining differences in variation of respondents’ innovativeness. Usage frequency and innovativeness was correlating positively, and the correlation coefficient level (Spearman’s rho) can be considered to be moderate, r_s.229, at significance 0.01. This can be also noticed from the graph below as there is a rather clear linear relationship with usage frequency and innovativeness. As the ANOVA is more robust test than correlation analyses there was further confirmed the relationship of internet usage and innovativeness. The relationship was confirmed at 0.01 as the F-value 3.159 exceeded the critical value 3.02 (v₁5; v₂184). The more often a person used a fixed internet connection the more innovative person was. Despite the clear relationship the explanation power was left to a rather low level, only 8 percent (eta² .079).

FIGURE 17. Respondents’ innovativeness by usage frequency of fixed internet connection.

To obtain more specific information on users’ characteristics concerning innovativeness or non-innovativeness there was further examined the correlations through Spearman’s rho. There was found a negative relationship between service channel through which money and economy services were used and users’ innovativeness. Meaning that the more up-to-date service channels respondents were using for money and economy services the more innovative they were (r_s = -.232; p<.01). With other service categories there was found no significant relationship with users’ innovativeness.

TABLE 32. Correlation for innovativeness versus service channel for money and economy services.

	Through which service channel do you use following services?		Innovativeness
Spearman's rho	Money and economy	Correlation Coefficient	-.232(**)
		Sig. (2-tailed)	.001
		N	188

** Correlation is significant at the 0.01 level (2-tailed).

It was also notable that those persons who were charged according to fixed price practice seemed to be less innovative than those who had decided to employ some other pricing methods. The correlation was rather weak though can be taken into account as the number of responses was rather high, r_s =.159 (p<.05). The direction of the correlation coefficient was also quite understandable as this method is the easiest and the clearest if a person is not well acquainted with mobile services.

TABLE 33. Correlation for respondents’ innovativeness versus pricing method in use for mobile services.

	What is the method through which usage of mobile services is priced?		Innovativeness
Spearman's rho	Fixed price	Correlation Coefficient	.154(*)
		Sig. (2-tailed)	.030
		N	197

* Correlation is significant at the 0.05 level (2-tailed).

Out of the variables examined and noted significant there was a clear emphasizes on pricing and service channel related variables. Especially pricing method seemed to have rather strong influence on respondents’ level of innovativeness. And also usage of fixed internet connection had some prediction power on user’s innovativeness. Unexpected however was that demographic variables didn’t correlate significantly with innovativeness (though analyses of variance showed that income level explains differences in innovativeness with an explanation power of 8 percent).

4.6.2. Moderate users

In moderate users’ segment there was discovered only one demographic variable with significant relationship with a respondent’s innovativeness. There was found a positive correlation with a respondent’s level of education and innovativeness, r = .168 (p<.01). Describing that the higher the level of education the higher the level of innovativeness. Also in this case the correlation coefficient was rather low; but as the response amount was quite high it has been taken into account. As the other demographic variables did not show any significant relationship with respondents’ innovativeness, there cannot be predicted a customer’s level of innovativeness with these variables.

TABLE 34. Correlation for respondents’ innovativeness versus level of education.

		Innovativeness
Education	Pearson Correlation	.168(**)
	Sig. (2-tailed)	.008
	N	245

** Correlation is significant at the 0.01 level (2-tailed).

Respondents’ innovativeness seemed to be more influenced by usage related variables. Usage frequency of mobile services appeared to correlate significantly (though rather weakly) with respondents’ innovativeness, r_s.149 (p<.05). The most often respondents brought out that they are using mobile services only occasionally (52%). And the mean of innovativeness among the respondents answered to this question was rather high 5.78 (std. deviation 1.252) compared to the heavy users’ segment’s 4.94. The more often a person uses mobile services the more innovative he/she is. In a case of respondents using mobile services 1-2 times per month (see figure 18), the innovativeness level was “questionable” as there were only three responses in this category. In other parts the upward trend was rather clear till 3-5 usage frequency. Daily users seem to be less innovative than other groups which rather surprising.

FIGURE 18. Respondents’ innovativeness versus usage frequency of mobile services.

The significance of this relationship was further confirmed through analyses of variance. The relationship was confirmed at 0.01 level even though the F-value 3.448 was quite barely above the critical value 3.02 (v₁ 5; v₂ 226). Also the explanation power was left to a rather low level, 7 percent (eta² .071).

Relating closely to usage habits there was also found some correlations between service channel choices and innovativeness. In a case of four service channel choices there was found significant and negative correlations with respondents’ innovativeness. Meaning that the more traditional service channels a customer used the less innovative a person was. This notion was quite in line with expectations but was now confirmed. Especially entertainment services were quite well correlating with respondents’ innovativeness (r_s= -.209; p<.01).

TABLE 35. Correlations for respondents’ innovativeness versus service channel choices service by service.

	Through which channel do you use following services?		Innovativeness
Spearman's rho	Money and economy	Correlation Coefficient	-.131(*)
		Sig. (2-tailed)	.046
		N	232
	News services	Correlation Coefficient	-.141(*)
		Sig. (2-tailed)	.039
		N	216
	Entertainment	Correlation Coefficient	-.209(**)
		Sig. (2-tailed)	.002
		N	209
	Reservations	Correlation Coefficient	-.146(*)
		Sig. (2-tailed)	.031
		N	219

* Correlation is significant at the 0.05 level (2-tailed); ** Correlation is significant at the 0.01 level (2-tailed)

As we are in this research dealing with pricing issues we were naturally interested in whether respondents’ innovativeness and wished pricing models were correlating. As a result there was found some rather slight but significant correlation with time-based pricing and a customer’s innovativeness. Respondents who wished time-based charging were less innovative than those wishing some other charging models. However, it must be noted that the correlation, r_s -.142, was rather weak even taken account of the response amounts (p<.05).

TABLE 36. Correlations for innovativeness and preferred pricing model

	What is the pricing model that you most prefer?		Innovativeness
Spearman's rho	Time-based pricing	Correlation Coefficient	-.142(*)
		Sig. (2-tailed)	.036
		N	220

* Correlation is significant at the 0.05 level (2-tailed).

In the moderate segment the best predictors for respondents’ innovativeness was the frequency they were using mobile services. Also the frequency of using fixed internet services did show some significant relationship with innovativeness. Rather good predictors for users’ innovativeness were also channel choices. It was found that the more up-to-date service channel a respondent chose for using one of the four services mentioned the higher was the level of the person’s innovativeness.

4.6.3. Prospective users

Unlike in the other two segments in this case there was found differences in innovativeness between the genders. Even though there was not found a significant correlation between the variables there was how ever found significant differences in means of innovativeness between the genders. Male respondents seemed to be a bit more innovative than female respondents. The difference was quite small despite that it was found to be significant at 0.05 level as F-value 3.891 exceeded the critical value 3.84 (v₁ 1; v₂ 293).

FIGURE 19. Mean of innovativeness by gender

There was found some significant correlation between a customer’s innovativeness and usage frequencies of the two service channels presented in the table 37. But again, the correlation coefficients were left to a very low level, and thereof meaningfulness or usefulness of these variables in predicting a customer’s innovativeness is questionable. Through Spearman’s rho there was obtained correlation scores for the two service channels .114 for fixed internet (p<.05) and .122 for computer with mobile phone as modem (p<.05).

TABLE 37. Correlations for customer’s innovativeness versus usage of service channels

	How often do you use following service channels?		Innovativeness
Spearman's rho	Fixed internet services	Correlation Coefficient	.114(*)
		Sig. (2-tailed)	.048
		N	298
	Computer mobile phone as modem	Correlation Coefficient	.122(*)
		Sig. (2-tailed)	.041
		N	282

* Correlation is significant at the 0.05 level (2-tailed).

When innovativeness was compared to usage frequencies of different service channels there was found significant relationships with usage frequencies of fixed internet services and respondents’ innovativeness. The significance was confirmed at 0.01 level by calculated F-value 4.137 which was above the tabled value 3.02 (v₁ 5; v₂ 279). By observing the figure 20 there can be seen a rather clear relationship with the usage frequency of fixed internet connection and a customer’s innovativeness. The more often a person uses fixed internet services the more innovative he/she is. Only exception to that trend was the daily users who did appear to be a bit less innovative than the previous group. Surprising was also the notion that users that were never using fixed internet connection were measured to be most innovative with the mean answer 5.41 (std. deviation 1.497). Even though there was found some differences in analyses of variance the correlation analyses did not found significant correlation with the usage frequency and a customer’s innovativeness.

FIGURE 20. Respondents’ innovativeness versus usage frequency of fixed internet connection.

As was the case in examining the price perception in the prospective users’ segment, the similar lack of significantly correlating or explaining variables occurred. There was found only few variables having a significant relationship with respondents’ innovativeness. And though these variables were formally significant they correlation coefficients were so low that those can be hardly used for predicting one’s innovativeness. Therefore it is more than likely that there are some underlying factors that are much more accurate for predicting a one’s innovativeness.

4.7. Price sensitivity

Customers’ price sensitivity was found to be one of the most important factors influencing on respondents’ price perception of mobile services. As in a case of other factors, also price sensitivity is inspected more detailed: how price sensitivity is differentiating between user segments. We want especially find out what are the variables explaining differences in respondents’ price sensitivity. The examining is done through demographic and usage related variables.

4.7.1. Heavy users

When respondents were asked how they stand with paying more for new services the results were rather clear. Almost 30 percent of respondents (28.3%) perceived the high prices of new services did hinder they willingness to acquire new services. And 61.5 percent of respondents were disturbed by high prices more than moderate. Mean response was 2.19 which is quite clearly expressed that users also in this segment did consider prices to be important factor in a case of new services. There was however quite a bit deviation in responses, 2.055.

FIGURE 21. Distribution of responses for variable of price sensitivity.

Through demographic variables there was found no significant variables which would have explained the differences in variance of responses. Therefore we went on inspecting usage and other user related variables in order to be able to explain the differences in variance of responses. There was found some positive correlation between users’ prices sensitivity and channel choices. Those respondents who used relationship services mainly through more traditional channels did appear to be less price sensitive than others as the correlation was positive (r=.247; p<.01). Nonetheless the explanation power remained on quite low level, 6 percent (r²=.061). The appeared correlation also with the service category “money and economy” although the correlation was in this case negative, r=-.163 (p<.05). Thus, the more up-to-date service channels were used in this case the less price sensitive users are.

TABLE 38. Correlations and response means of price sensitivity by channel choices.

	Mean	Std. Deviation	N
Through which channel do you use following services? Relationship	1.24	1.474	143
Search	1.05	.703	198
Hobbies and leasure	1.36	1.386	193
Communication	1.04	.839	191
Home and family	1.95	1.846	170
Shopping	3.15	2.022	163
Games	1.60	1.809	162
Money and economy	1.82	1.648	191
Carrier and education	1.69	1.564	176
News	1.61	1.703	180
Entertainment	2.02	1.856	169
Reservations	2.15	1.936	176

Through which channel do you use following services?		Price sensitivity
Relationship services	Pearson Correlation	.247(**)
	Sig. (2-tailed)	.003
	N	141
Money and economy	Pearson Correlation	-.163(*)
	Sig. (2-tailed)	.026
	N	187

** Correlation is significant at the 0.01 level (2-tailed); * Correlation is significant at the 0.05 level (2-tailed).

This service channel relationship can be interpreted so that persons who use relationship services through traditional channels are willing to pay even some extra to be able to do so. Whereas persons who use money and economy services through conventional channels are more price sensitive even though they are not willing to take mobile service in use. The reason for this could be that mobile services are perceived rather expensive and are thus not used. There was also found that the more often a person uses a fixed internet connection the less price sensitive one is, although relationship was rather weak r=.153 (p<.05).

TABLE 39. Correlations and mean responses for price sensitivity and usage frequency of fixed internet connection.

	Mean	Std. Deviation	N
Price sensitivity	2.19	2.055	205
Fixed internet services	3.48	1.668	201

		Innovativeness
Fixed internet services	Pearson Correlation	.153(*)
	Sig. (2-tailed)	.031
	N	199

* Correlation is significant at the 0.05 level (2-tailed).

Interesting finding was also the clear positive correlation between satisfaction to pricing model in use and price sensitivity. Meaning that those person that were most satisfied with the pricing model in use were surprisingly less price sensitive (r_s=.456; p<.001). And this variable was also quite good predictor on a person’s price sensitivity (together with the perceived expensiveness of mobile services) with the explaining almost 21 percent of the variance in price sensitivity (r_s²=.208).

Also in a case of this factor it was found that two most important variable categories influencing on respondents’ price sensitivity were price related and service channel related variables. Especially satisfaction to charging model was explaining well users’ price sensitivity. The more satisfied person was on charging method in use the less price sensitive he/she was.

4.7.2. Moderate users

When price sensitivity was measured in moderate users’ segment the respondents were asked “on what grounds did they chose mobile services; how relevant was price?”. Amongst the respondents, service prices were perceived to be the most important variable influencing on mobile service choices. Approximately 27 percent of the respondents stated that a product price possessed a very important role in the decision (“totally agree”). And if the 3 last bars (in figure 22) are considered to mean those who considered a service price to be more important factor than other factors, then 57.5 percent of respondents were influenced by service prices more than any other variables together.

FIGURE 22. Respondents’ price sensitivity.

Amongst the respondents the mean price sensitivity was 3.99, indicating rather high price sensitivity. As standard deviation was also rather high, 2.064, there was obtained rather much differentiation amongst respondents in moderate users’ segment. Therefore it was interesting to examine the factors that would explain those differences. For this intension we first explored the demographic variables, but found no significantly explaining variables. Instead there was found some slight but significant correlations between charging methods and price sensitivity.

TABLE 40. Correlations for hoped charging method versus price sensitivity

	Mean	Std. Deviation	N
Price sensitivity	3.99	2.064	243
What is the charging method that your mobile services usage is charged at the moment?	2.13	2.423	240
What is the charging method that you wished to be used in mobile services? According to actual usage	3.74	2.529	229
Fixed price	4.72	2.201	236
Service based	3.11	2.547	217
Bundle pricing	3.32	2.558	217
Separately from a service and data transfer	2.89	2.701	218
Time-based pricing	2.81	2.798	221

		Price sensitivity
Fixed price	Pearson Correlation	0.168(*)
	Sig. (2-tailed)	0.010
	N	235
Bundle pricing	Pearson Correlation	0.153(*)
	Sig. (2-tailed)	0.024
	N	217

* Correlation is significant at the 0.05 level (2-tailed); ** Correlation is significant at the 0.01 level (2-tailed).

Price sensitivity in this segment was rather peculiar compared to other factors and other segments as there was found only two pricing related variables correlating with price sensitivity. There was found that price sensitivity correlated positively with the fixed pricing practices. Both fixed charging (r=.168; p<.05) and bundle charging (r=.153; p<.05) were correlating positively with price sensitivity. Respondents who preferred fixed or bundle charging methods were less price sensitive than others. But as the correlations were rather weak the explanation powers were quite meaningless.

4.7.3. Prospective users

When there was examined the price sensitivity of mobile services amongst the prospective users’ segment, there was first found the high standard deviation, 2.258. It indicated the high differences amongst respondents’ price sensitivities. Nonetheless, there was still found some clear uniformity inside the segment. As depicted in the figure 23, there was a noticed apparent bias towards price sensitive end of the scale. Even though the figure is clearly biased the percentual differences were rather small (which was also indicated by the standard deviation). In the figure only slight proportion responded that price isn’t the chief variable deciding which services will be chosen. The highest three categories were chosen in 39.4 percent of responses, while the lowest three gathered only 27 percent of responses. There were also 1/5^th of respondents (20.3%) that couldn’t say.

FIGURE 23. Respondents’ price sensitivity in prospective users’ segment.

For examining the reasons for the variance of price perceptions there was first studied the demographic variables. It was found that only one variable had a significant correlation with a respondent’s price sensitivity – level of education. Respondents’ educational level were correlating significantly but rather weakly with price sensitivity (r= -.133; p<.05). As in the other segments, the demographic variables were possessing very weak relationships with the studied factors. Thus also the prediction and explanation powers were left to low (or even meaningless) levels. In this case the lower a respondent’s level of education was the lower was also a person’s price sensitivity.

As a surprise, according to responses, it was found that users that possessed either a comprehensive or secondary level of education were least price sensitive. Also F-test confirmed that explanation power of education was significant at 0.05 level as the F-value was 2.498 and thus exceeding the critical value 2.01 (v₁ 7; v₂ 229). As mentioned the explanation power was left to a very low level, only 7 percent (eta² .071).

FIGURE 24. Respondents’ price sensitivity by level of education

TABLE 41. ANOVA for price sensitivity versus level of education.

			Sum of Squares	df	Mean Square	F	Sig.
Price sensitivity vs. education	Between Groups	(Combined)	60.625	7	8.661	2.498	.017
	Within Groups		794.041	229	3.467
	Total		854.667	236

Even though the correlation test did not provide other demographic variables correlating significantly with price sensitivity, ANOVA did provide one. Analyses of variance found significant differences in respondents’ price sensitivities measured according to their occupations. The significance was confirmed at 0.01 level as the calculated F-value 2.687 exceeded the critical value 2.51 (v₁ 8; v₂ 229). Especially employees, students, and officials seemed be least price sensitive. But, despite of the significance, the explanation power was to only 9 percent (eta² .086).

There was also found moderate correlation between two charging related variables and price sensitivity. The both of the two variables were correlating positively with correlation coefficients of r_s .317 (p<.01) for knowledge on different charging methods and r_s.233 (p<.01) for satisfaction to charging method in use. To interpret the correlations, respondents who possessed a good knowledge of different charging methods and who were satisfied with charging methods in use, were most likely less price sensitive.

TABLE 42. Correlations for price sensitivity versus pricing variables.

			Knowledge on charging method of mobile services in use	Satisfaction to charging method in use
Spearman's rho	Price sensitivity	Correlation Coefficient	.317(**)	.233(**)
		Sig. (2-tailed)	.000	.000
		N	299	298

** Correlation is significant at the 0.01 level (2-tailed).

There was also found a significant correlation with usage frequency of fixed internet connection and price sensitivity. The correlation was negative but rather weak, r= -.144, explaining thus rather poorly the variances in price sensitivity. According to the result the more often a person uses a fixed Internet connection the less price sensitive he/she is.

TABLE 43. Usage frequency versus price sensitivity.

		Price sensitivity
Usage of fixed internet	Pearson Correlation	-.144(*)
	Sig. (2-tailed)	.013
	N	297

* Correlation is significant at the 0.05 level (2-tailed).

4.8. Satisfaction to service quality of teleoperator

Satisfaction to operator’s services was found to be important factor predicting and influencing on customers’ price perceptions. Only exception was the heavy users who did not perceive the satisfaction to influence significantly on their price perceptions of mobile services. But as the satisfaction was found to have significant influences on the two other segments it was considered to be necessary to explore in more detail: how satisfaction is formed among customers in these segments.

4.8.1. Moderate users

In a case of moderate users there must be first noted that mean satisfaction to a operator’s services was apparently lower than in heavy users group, only 3.23 (std. deviation 2.278). As the standard deviation also indicated there were rather high irregularities in responses through the whole scale which why it was hard to find any consistency. But these differences gave us also interesting challenges for examining the reasons causing the variances in responses. There was found three clear peaks in the figure 25; totally disagree, bar 3, and don’t know.

FIGURE 25. Satisfaction to operator’s services.

Amongst the demographic variables there was found one variable correlating significantly with the satisfaction variable – annual gross income. Income level was found to have a negative (rather weak) correlation coefficient, r= -.154 (p<.05). Meaning that the higher was a person’s income level the less satisfied the person was on operator’s services. Even though the relationship was significant the low correlation coefficient gives only weak explanation power for a customer’s satisfaction level. But it however is a sign that income level should be taken into account when predicting a customer’s level of satisfaction (if any other better predictors are not at hand). The significance was also confirmed at 0.05 level with F-test as calculated F-value was 1.896 and thus exceeding the critical 1.83 (v₁ 10; v₂ 226).

FIGURE 26. Satisfaction to operator’s services by level of income.

It should be also noted that credibility of response means in income groups over 50.001€ was rather weak as the amount of responses was very low in these high-end income classes (n<9). Even though income level was found to have significant correlation with the satisfaction variable, explanation power was left to a rather low level, 8 percent (eta² .077).

When there was inspected other variables that might indicate persons’ satisfaction to operator’s services there was observed usage and pricing related variables. Pricing did appear to be rather significant for influencing on persons’ satisfaction to operator’s services. Especially a “price transparency” was significantly correlating with satisfaction. Awareness of pricing method in use (r_s .151; p< .05) and a person’s knowledge on different charging methods (r_s .215; p<.01) correlated both positively but only moderately. Meaning that the better a person was aware of charging related practices the more satisfied he/she was. Therefore it is important for operators to create charging methods that are as transparent as possible.

TABLE 44. Correlation for satisfaction to operators services versus price transparency

			Knowledge on charging method in use	I’ve been well informed on selectable different charging methods
Spearman's rho	Satisfaction to operator’s services	Correlation Coefficient	.151(*)	.215(**)
		Sig. (2-tailed)	.020	.001
		N	239	239

* Correlation is significant at the 0.05 level (2-tailed); ** Correlation is significant at the 0.01 level (2-tailed).

Satisfaction to a charging method was also found to be significant in predicting a person’s level of satisfaction. It was found to correlate positively and moderately with satisfaction to operator’s services, r_s= .240, at 0.01 level. Thus, for an operator whose customers are satisfied with services are also likely to be satisfied with charging related issues and other way round.

TABLE 45. Correlations for satisfaction to operator’s services versus satisfaction with pricing of mobile services

			Satisfaction to charging method of mobile services in use
Spearman's rho	Satisfaction to operator’s services	Correlation Coefficient	.250(**)
		Sig. (2-tailed)	.000
		N	239

** Correlation is significant at the 0.01 level (2-tailed).

4.8.2. Prospective users

In prospective users’ segment the satisfaction to operator’s services was measured by enquiring correctness to a following statement: “my teleoperator is never too busy to answer my questions”. This variable obtained the best factor loadings in the factor analyses in describing a respondent’s satisfaction to operator’s services. The mean answer was 3.65 (0=disagree; 6=agree) and standard deviation 2.327, which both were higher than in the moderate segment. As seen from the figure 27, all options obtained responses quite equally (which also high std. deviation indicated). In addition, “don’t know” responses obtained very high proportion which naturally affected to reliability of the measurement.

FIGURE 27. Respondents’ satisfaction to operator’s services.

In order to find out the variables predicting or explaining the differences in variance of the satisfaction factor there was first explored the demographic variables. There was found two demographic variables that received significant explanation powers: age and profession. In a case of age there appeared to be a slight tendency towards older respondents to be more satisfied to operator’s services. The correlation between age and satisfaction was thus positive, r=.221 (p<.01). The significance was also confirmed in ANOVA at 0.01 level as calculated F-value 3.607 was above the tabled F-value 3.02 (v₁ 5; v₂ 298).

FIGURE 28. Satisfaction to services by age groups.

The second demographic variable that was found to be significant in explaining differences in variance of the responses was a respondent’s profession. From the figure 29 there can be noticed four profession groups that stood out from others: farmers, out of working life, retirees, and officers. All these professions received response means over 4 and were notably more satisfied with operator’s services than the average responder. Profession and satisfaction were also correlating together even though the correlation coefficient was left to a low level. The correlation was positive, r=.134, at significance level 0.05. The correlations shows that the more independently a person is working the more satisfied a person is to operator’s services.

FIGURE 29. Satisfaction to operator’s services versus respondent’s profession.

The significance was confirmed by ANOVA below. The calculated F distribution for the profession mean square was 2.365 whereas the critical F for 0.05 significance was 1.88 (v₁ 9; v₂ 293). The calculated F-value exceeded the critical F, and the null hypothesis of equal means was rejected. There is a difference between professions in the level of satisfaction to operator’s services. Squared eta-coefficient revealed that by a respondent’s profession there can be explained 7 percent of the variance in satisfaction to teleoperator’s services (eta² .068).

TABLE 46. ANOVA for respondents’ satisfaction to teleopertor’s services versus profession

			Sum of Squares	df	Mean Square	F	Sig.
Satisfaction to operator’s services vs. profession	Between Groups	(Combined)	110.367	9	12.263	2.365	.014
	Within Groups		1519.138	293	5.185
	Total		1629.505	302

4.9. Investment readiness on mobile services

The fourth factor with significant influence on price perception was investment readiness on mobile services. The importance of this factor was rather unexpected even though there was predicted that this factor indicates quite well a customer’s price perception. Meaning that, a person who is ready to invest more on mobile services is likely to perceive the prices also affordable. But to be able to bring more detailed insight into this factor and to variables influencing on it, there is in next chapter examined investment readiness segment-by-segment.

4.9.1. Heavy users

In heavy users’ segment investment readiness was found to be third important factor affecting to price perceptions. When trying to identify user related variables explaining the differences in investment readiness the demographic variables were first inspected. But amongst these variables there was found no significantly correlating variables. When observing pricing related elements there was found two variables that had significant influence on investment readiness. These variables were both correlating significantly but weakly with investment readiness: hoped charging methods and assessment method for mobile service prices.

TABLE 47. Correlations for investment readiness versus charging variables.

			Preference for fixed-fee charging	Assessment method for prices of mobile services
Spearman's rho	Investment readiness	Correlation Coefficient	.150(*)	-.149(*)
		Sig. (2-tailed)	.035	.035
		N	197	201

* Correlation is significant at the 0.05 level (2-tailed). ** Correlation is significant at the 0.01 level (2-tailed).

It was found that the more a person was hoping a fixed-fee charging methods the more willing a person was to invest more on mobile services. The correlation r_s .150 (p<.05) was quite weak but was still showing the positive aspects of fixed fee charging methods. There was also a significant relationship between assessment method of mobile services prices and investment readiness (r_s -.149; p<.05). This shows that a person who is assessing mobile service prices on the grounds of traditional service channels are more likely to be ready to invest more on mobile services.

Investment readiness was however left rather open for future studies as there was found only two weakly correlating variables. Therefore it is probable that there are other variables having significant and meaningful influences which were not found in this research. This was observed already in the factor analyses were we measured the factors cross-correlations. Investment readiness was noticed to correlate quite strongly with price sensitivity (r = .305).

4.9.2. Moderate users

As was the case in heavy users’ segment also moderate users did not possess any demographic variables that would significantly correlate investment readiness on mobile services. Therefore it is expected that variables relating to investment readiness will be more likely to be found under usage related topics.

Even though charging was noticed to have quite strong influence on many factors discussed earlier, it was still a surprise that all the significant variables affecting to investment readiness in this segment were charging related. Moreover, the correlation coefficients were all very close to each other as can be seen from the table 48. Charging model in use and hoped charging model were both correlating positively with r_s .147 (p<.05) and r_s .150 (p<.05). The more a person wants a fixed charging method the more willing he/she is to invest more on mobile services. Assessment method of mobile service prices possessed a negative relationship with investment readiness. Meaning that the more traditional was the assessment basis the more willing a person was to invest more on mobile services. This notion makes sense as the traditional services are rather expensive to use and does not possess the benefits of mobility.

TABLE 48. Correlations for investment readiness on mobile services versus charging related variables.

			Charging model of mobile services in use	Fixed fee as hoped charging method	Assessment method for mobile service prices
Spearman's rho	Investment readiness	Correlation Coefficient	.147(*)	.150(*)	-.149(*)
		Sig. (2-tailed)	.035	.035	.035
		N	205	197	201

* Correlation is significant at the 0.05 level (2-tailed).

This factor was explained rather weakly in both segments even taken into account rather high response amounts. Additionally there was found only few variables having significant relationships with the satisfaction factor. Therefore it should be noted that it is very likely that there are some other variables, which were not included into the study, that possess significant relationships with customers’ investment readiness. One explanation was found from the factor analyses where was also measured the factors cross-correlations. There was found in a case of both segments that satisfaction factor correlated moderately with the other factors.

4.10. Preference for mobile service bundles

In order to bring concrete and practical solutions for creating more effective pricing methods there is proposed a bundling strategy for mobile services. This strategy is seen to be one of the most interesting strategies for mobile service business. As surveys conducted amongst consumers in other businesses have shown, in general, consumers prefer service bundles. The bundling strategy also gives tools for creating service bundles which decrease users’ price sensitivity and increase the perceived value of mobile services. It is however necessary to examine its adoptability and usefulness from the customer point of view. Thus we have brought a closer look on users’ preference towards bundling mobile services, and factors affecting to users’ interestedness in acquiring mobile services in bundles.

4.10.1. Heavy users

First there was examined the heavy users’ segment and they preference for acquiring mobile services in bundles. Out of all respondents in this segment interestedness in service bundles was rather high as the response mean was 4.12 (0= not interested; 6= very interested in) with a standard deviation of 1.778. Male respondents did perceive service bundles to be more interesting option (mean 5.01, std. 1.642) than female respondents (mean 4.55, std. deviation 1.621) even though the difference wasn’t significant (p>.05). Also the other demographic variables were found to be insignificant in explaining the differences in the variance of this factor.

As there was also investigated the optimal size of the service bundles respondents’ were enquired “what would be the number of mobile services they would need?”. In this segment users tended to need greater number of mobile services than users in the other two segments. Mean response was 4.5 services although the standard deviation was rather high 2.235. Standard deviation was increased as 20 percent of respondents had responded that they would need more than 7 services.

FIGURE 30. Perceived need for amount of mobile services.

To further examine customers’ preferences for mobile service bundles there was investigated factors explaining the differences in the variance. For this purpose there has been employed a regression analyses which has produced us with insight into factors having a significant influence on customers’ preference for mobile service bundles. As a result there was obtained four factors that had a significant influence on the dependent variable (interestedness in acquiring mobile services in service bundles): price sensitivity (R²=.211), willingness to invest more money on mobile services (R²=.180), perceived price (R²=0.159), and price transparency (R²=.058).

The price sensitivity was found to have a significant influence on preference for mobile service bundles at 0.01 level as the F-value 7.735 exceeded the critical value 2.51 (v₁ 8; v₂ 194). This factor was found to have strongest relationship with the dependent factor with the multiple correlation coefficient 0.492. Price sensitivity was explaining approximately 24 percent of the variation in the dependent variable (R² .242). And taken into account the number of independent variables there was obtained an explanation power of 21 percent (adjusted R² 0.211). When compared the dependent variables under this category, there was found two variables that best predicted the users’ interestedness to acquire services in bundles: “high prices of mobile services as the main barrier to usage” and “need for tailored bundles” (beta .175 & .195; p<.05).

TABLE 49. Regression for respondents’ preference for acquiring mobile services in bundles versus price sensitivity

Model	R	R²	Adjusted R²	Std. Error of the Estimate
1	.492(a)	.242	.211	1.584

a Predictors: (Constant), Price sensitivity; b Dependent Variable: Interestedness in mobile service bundles.

ANOVA(b)

Model		Sum of Squares	df	Mean Square	F	Sig.
1	Regression	155.260	8	19.407	7.735	.000(a)
	Residual	486.770	194	2.509
	Total	642.030	202

a Predictors: (Constant), Price sensitivity; b Dependent Variable: Interestedness in mobile service bundles.

The investment readiness on mobile services was also found to correlate significantly with preference for mobile service bundles. The significance was confirmed by F-test at 0.01 level as F-value 5.935 was above the critical value 2.41 (v₁=9; v₂=193). This factor explained 22 percent of the variation in the dependent variable (R² .217). The comparable explanation power was 18 percent (R².180). Thus also this factor was found to have quite strong relationship with the dependent variable as the multiple correlation coefficient was almost equal with the price sensitivity factor.

TABLE 50. Regression for respondents’ preference for acquiring mobile services in bundles versus investment readiness

Model	R	R²	Adjusted R²	Std. Error of the Estimate
1	.466(a)	.217	.180	1.605

a Predictors: (Constant), Price sensitivity;

b Dependent Variable: Interestness in acquiring mobile services in service bundles.

ANOVA(b)

Model		Sum of Squares	df	Mean Square	F	Sig.
1	Regression	137.517	9	15.280	5.935	.000(a)
	Residual	496.887	193	2.575
	Total	634.404	202

a Predictors: (Constant), Price sensitivity

b Dependent Variable: Interestness in acquiring mobile services in service bundles.

The third strongest relationship with preference for mobile service bundles was respondents’ price perceptions on mobile services as the multiple correlation coefficient was 0.424. The relationship was also found to be significant at 0.01 level as the F-value 8.612 exceeded the critical value 2.21 (v₁=5; v₂=196). Therefore, as there was hypothesized, the service bundle and price perceptions factors where found to have a significant relationship. The price perception was also found to explain 18 percent of the variation in the dependent variable (R² .180). Whereas the adjusted explanation power was 16 percent (adjusted R² 0.159). There should how ever be noted that if variables of perceived price are separated three variables would be significant – “perceived usage of mobile services”, “match of mobile service prices and quality”, and “high prices of mobile services are the main barrier in increasing the usage of mobile services” (p<.05). These variables have the highest explanation power (Beta .142, .165, and .203).

TABLE 51. Regression for respondents’ preference for acquiring mobile services in bundles versus perceived price

Model	R	R²	Adjusted R²	Std. Error of the Estimate
1	.424(a)	.180	.159	1.619

a Predictors: (Constant), Price sensitivity

b Dependent Variable: Interestness in acquiring mobile services in service bundles.

ANOVA(b)

Model		Sum of Squares	df	Mean Square	F	Sig.
1	Regression	112.828	5	22.566	8.612	.000(a)
	Residual	513.553	196	2.620
	Total	626.381	201

a Predictors: (Constant), Price sensitivity

b Dependent Variable: Interestness in acquiring mobile services in service bundles.

The fourth and last significant variable influencing on respondents’ preference for mobile service bundles was price transparency. The significance at the 0.01 level was confirmed as the F-value 13.683 exceeded the tabled F-value 2.21 (v₁=5; v₂=196). The relationship was however significantly lower, R .250, than in case of other three factors. This also reflected to explanation power. This factor was explaining only 6 percent of the variation (R² .063). And when taking into account the number of independent variables the explanation power was decreased to 5.8 percent (adjusted R² 0.058).

TABLE 52. Regression for respondents’ preference for acquiring mobile services in bundles versus price transparency

Model	R	R²	Adjusted R²	Std. Error of the Estimate
1	.250(a)	.063	.058	1.726

a Predictors: (Constant), Price sensitivity

b Dependent Variable: Interestness in acquiring mobile services in service bundles.

ANOVA(b)

Model		Sum of Squares	df	Mean Square	F	Sig.
1	Regression	40.748	1	40.748	13.683	.000(a)
	Residual	610.470	205	2.978
	Total	651.217	206

a Predictors: (Constant), Price sensitivity

b Dependent Variable: Interestness in acquiring mobile services in service bundles.

The regression analysis showed that there can be found at least the four factors that influenced on respondents’ preference for acquiring mobile service in bundles. By affecting to these factors there can be influenced also on preference for service bundles. But in order to obtain more specific knowledge on how bundles should be constructed there is needed information about preferred service combinations.

FIGURE 31. Factors explaining the differences in variance of respondents’ interestedness in mobile service bundles.

For building service combinations customers were enquired the most preferred mobile services which usage they are most likely to increase. From these selected services there was chosen combinations of three services that were most often selected together. In the segment of heavy users there was found at five service combinations that were most often chosen (in 22.2% of responses).

Most often selected service combinations were Relationship (e.g. dating, relationship counselling, etc.) – Search (e.g. Google, Altavista, maps, etc.) – and Home and housing related services with a proportion of 7.6 percent. The second often chosen bundle construct was Relationship – Search – Remote control (e.g. activating burglar alarm or turn on a sauna from distance) services with 4.7 percent of responses. These after there were three equally often selected bundles with percentages of 3.3 and should also be considered as good basis in bundling services.

TABLE 53. Mobile service bundles most often preferred by respondents

	N	%	Valid %	Cumulative %
Relationship – Search – Home and housing	16	7.6	7.6	15.6
Relationship – Search – Remote control	10	4.7	4.7	8.0
Entertainment – Chat – Relationship	7	3.3	3.3	3.3
Relationship – Home and housing – Child and family	7	3.3	3.3	18.9
Relationship – Child and family – Fashion and beauty	7	3.3	3.3	22.2
Total	47	22.2	22.2

4.10.2. Moderate users

Out of all respondents in the moderate users’ segment preference for acquiring mobile services in bundles was prominently lower than in the heavy users’ segment, mean 3.78. Standard deviation was also rather high 2.124 indicating high differences in responses. Customers in this segment did also possess a view that the number of services they would get sufficiently along with wasn’t more than four services in average (mean 4.06 services; std. deviation 2.368) which was apparently lower than in heavy users’ segment (4.5 services). Also the standard deviation of responses was quite high in this case as the respondents perceived in 15 percent of cases that they would need more than seven services. The most preferred mobile service bundle sizes selected were 1 (17.1%), four (16.2%), and more than seven (15.3%).

FIGURE 32. Perceived need for amount of mobile services

In the case of moderate users’ segment there was also investigated the factors affecting to preferences for acquiring mobile services in service bundles. And result was rather surprising as there was found only one independent variable having a significant influence on the dependent variable. The variable was price sensitivity which received a moderate relationship as the multiple correlation coefficient was .359. The price sensitivity explained 13 percent of the variance in respondents’ preference for mobile service bundles (R² .129). And when taken into account the amount of independent variables included into the analyses the comparable explanation power decreased to 9 percent (adjusted R²=.09). Other three examined variables did not receive significant explanation powers. In other words, preference for acquiring mobile services in bundles was not influenced by users’ price perception, price transparency, or satisfaction to operator’s services. The significance of price sensitivity was confirmed by F-test at 0.05 level as F-value 3.291 exceeded the critical value 1.83 (v₁=10; v₂=222).

TABLE 54. Regression analyses for preference for acquiring mobile service in bundles versus price sensitivity

Model	R	R²	Adjusted R²	Std. Error of the Estimate
1	.359(a)	.129	.090	2.237

a Predictors: (Constant), price sensitivity; b Dependent Variable: Interestedness in mobile service bundles.

ANOVA(b)

Model		Sum of Squares	df	Mean Square	F	Sig.
1	Regression	164.740	10	16.474	3.291	.001(a)
	Residual	1111.398	222	5.006
	Total	1276.137	232

a Predictors: (Constant), price sensitivity; b Dependent Variable: Interestedness in mobile service bundles.

When examining the figure 33 there can be visibly noticed the difference between heavy users’ segment. In the moderate users’ segment there was found only one factor significantly influencing on respondents’ preference for mobile service bundles. As there wasn’t either obtained a significant relationship with the dependent variable and the price perception factor or significant relationship between price perception and preference for mobile service bundles (despite of the hypothesis).

FIGURE 33. Factors influencing on consumers’ preference for acquiring mobile service bundles.

Next there were modified mobile service bundles by enquiring the most used mobile services amongst the respondents. Service bundles in this segment were clearly more gender based than in heavy users’ case - home, family and gender related services have received more weigh than others. Two most often selected service bundles were “men and women – Search – Home and housing services” and “men and women – Search – Child and family services” which both received a proportion of 4.3 percent. Notable was also the appearance of shopping and gambling services which did not get this high popularity in the two other segments. In total, there is presented 5 service bundles that received some common popularity with a total proportion of 18 percent.

TABLE 55. Mobile service bundles most often preferred by respondents

	N	%	Valid %	Cumulative %
Services to men and women – Search – Home and housing	11	4.3	4.3	7.4
Services to men and women – Search – Child and family	11	4.3	4.3	11.7
Services to men and women – Home and housing – Child and family	9	3.5	3.5	15.2
Services to men and women – Search – Remote control	8	3.1	3.1	3.1
Shopping	8	3.1	3.1	18.3
Total	60	21.2	18.3

4.10.3. Prospective users

In prospective users’ segment the preference for mobile service bundles was approximately equal with the moderate segment as the mean answer was 3.80 with standard deviation of 2.497. Only difference was the slightly higher deviation which was caused by higher differences between the respondents preferences. Especially the proportion of “don’t know” answers was high 22.6 percent. Thus, because of the high deviation the reliability of these results is decreased.

When examining the respondents’ perceptions on the amount of mobile services they perceive to need, the tendency towards lesser needs had continued. There was found that customers in this segment perceive that they would get well along with 3.5 mobile services in average (std. deviation 2.2327). In heavy and moderate segments the mean value for needed amount of services was 4.5 (heavy) and 4.1 (moderate). When observing the figure 34 there can be noticed that a prominent proportion of respondents perceived that 1 or 2 services would be well enough. This is clearly indicating the lower mobile service needs and lower level of usage of mobile services.

FIGURE 34. Perceived need for amount of mobile services

In the case of prospective users’ segment there was, as well as in the case of the other two segments, investigated the factors affecting to preference for acquiring mobile services in service bundles. As a result there was obtained four factors having a significant influence on the dependent variable: price sensitivity, price transparency, price perception, and satisfaction to operator’s services. In total, with these variables there could be obtained an explanation power of 34.9 percent. When examining the demographic variables and their ability to predict a respondent’s preference for mobile service bundles there was found no significant results.

But out of the factors created in the factor analyses all of them were revealed to have a significant influence on customers’ preference for mobile service bundles. The strongest relationship was found to have with the price sensitivity with the multiple correlation coefficient of 0.404. The significance of the relationship was confirmed by F-test at 0.01 level as the F-value 14.018 was above the critical value 3.32 (v₁= 4;v₂= 288). Price sensitivity was found to explain 16 percent of the variance amongst the responses (R² .163). When the number of independent variables was taken into account the explanation power decreased to 15 percent (adjusted R² .151).

The explanation power of this factor was moderate when compared to the other segments. By observing more closely the variables under the price sensitivity there was found two variables that correlated better with the dependent variable than others. These two variables were “willingness to pay more to get more versatile service choices” and “acquiring mobile services based on price”. These variables explanation powers were Beta .262 (p<0.001) and later one Beta .188 (p<0.05).

TABLE 56. Regression for preference for acquiring mobile services in bundles versus price sensitivity

Model	R	R²	Adjusted R²	Std. Error of the Estimate
1	.404(a)	.163	.151	2.276

a Predictors: (Constant), Price sensitivity.

b Dependent Variable: Interested in acquiring mobile services in service bundles.

ANOVA(b)

Model		Sum of Squares	df	Mean Square	F	Sig.
1	Regression	290.355	4	72.589	14.018	.000(a)
	Residual	1491.358	288	5.178
	Total	1781.713	292

a Predictors: (Constant), Price sensitivity.

b Dependent Variable: Interested in acquiring mobile services in service bundles.

Price transparency had the second highest relationship with customers’ preference for mobile service bundles, R=.359. Also the explanation power was naturally the second highest; explaining 13 percent of the variance in the dependent variable (R² .129). And when this was adjusted by taking into account the number of independent variables, the explanation power decreased to 12.6 percent. The relationship was also confirmed by F-test at 0.01 level as the F-value 44.402 was clearly over the critical value 6.63 (v₁= 1;v₂= 301). The explanation power was considered to be quite sufficient when compared with the other factors in this segment. There can be further noticed that “high prices of mobile services” –variable seemed to be best predictor for users’ preference for service bundles (beta=0.162, p<.01).

TABLE 57. Regression analyses for customers’ preference for acquiring mobile services in bundles versus price transparency

Model	R	R²	Adjusted R²	Std. Error of the Estimate
1	.359(a)	.129	.126	2.328

a Predictors: (Constant), Price transparency

b Dependent Variable: I am interested in acquiring mobile services in bundles?

ANOVA(b)

Model		Sum of Squares	df	Mean Square	F	Sig.
1	Regression	240.588	1	240.588	44.402	.000(a)
	Residual	1630.924	301	5.418
	Total	1871.512	302

a Predictors: (Constant), Price transparency

b Dependent Variable: I am interested in acquiring mobile services in bundles?

Satisfaction to operator’s services had the third highest influence on a person’s preference for mobile service bundles. The relationship was however left to rather low level as the multiple correlation coefficient was only 0.266. Therefore, also the explanation power was left to very low level; explaining only 7 percent of the variance in the dependent variables (R² .071). The explanation power was especially weak if taken into account the number of independent variables, only 4 percent (adjusted R².037). The relationship between these two was confirmed by F-distribution at 0.05 level as the F-value 2.084 exceeded the critical tabled value 1.83 (v₁ 10; v₂ 273).

TABLE 58. Regression for customers’ preference for acquiring mobile services in service bundles versus satisfaction to operator’s services

Model	R	R²	Adjusted R²	Std. Error of the Estimate
1	.266(a)	.071	.037	2.431

a Predictors: (Constant), Satisfaction to teleoperator

b Dependent Variable: Interested in acquiring mobile services in service bundles.

ANOVA(b)

Model		Sum of Squares	df	Mean Square	F	Sig.
1	Regression	123.165	10	12.316	2.084	.026(a)
	Residual	1613.300	273	5.910
	Total	1736.465	283

a Predictors: (Constant), Satisfaction to teleoperator

b Dependent Variable: Interested in acquiring mobile services in service bundles.

The fourth significant factor affecting to customers’ preference for mobile service bundles was the price perception. The relationship was significant at 0.01 level as F-value 3.512 was above the critical value 3.32 (v₁ 4; v₂ 276). As a cause of the low correlation coefficient 0.220, the explanation power was also left to very weak level, R=.048. The factor’s ability to explain variance in the dependent variable became even weaker when adjusting it by taking into account the number of independent variables, adjusted R² .035.

TABLE 59. Regression for customers’ preference for acquiring mobile services in service bundles versus price perception

Model	R	R²	Adjusted R²	Std. Error of the Estimate
1	.220(a)	.048	.035	2.427

a Predictors: (Constant), Price perception.

b Dependent Variable: Interested in acquiring mobile services in service bundles.

ANOVA(b)

Model		Sum of Squares	df	Mean Square	F	Sig.
1	Regression	82.759	4	20.690	3.512	.008(a)
	Residual	1625.739	276	5.890
	Total	1708.498	280

a Predictors: (Constant), Price perception.

b Dependent Variable: Interested in acquiring mobile services in service bundles.

To conclude, amongst the prospective users there was found four factors influencing significantly on customers’ preference for mobile service bundles (figure 35). The factors can be seen to be divided into two groups according to correlation coefficients: the price sensitivity and price transparency factors; and the satisfaction to operator’s services and price perception factors. Especially price sensitivity and price transparency factors obtained explanation powers that were meaningful. But the other two factors had explanation power which usefulness was left to rather a low level.

TABLE 35. Factors influencing on respondents’ preference for mobile service bundles.

As in the other two segments, also in this segment there was studied preferred bundle constructs. There was obtained 5 service combinations which were perceived to obtain meaningful popularity. The total proportion of these bundle constructs was 21.5 percent. The most often selected service combinations was “men and women (e.g. Ellit-service for women) – Search (e.g. Google, Altavista, maps, etc.) – Remote control (e.g. activating burglar alarm or turn on a sauna from distance)” related services with a proportion of 5.8 percent. The second often chosen bundle combination was “men and women – Search – Remote control” related services with proportion of 4.8 percent. The third most often chosen service bundle was “men and women – Search – Remote control” related services with a proportion of 4.5 percent.

TABLE 60. Mobile service bundles most often preferred by respondents

	N	%	Valid %	Cumulative %
Services for men and women – Search – Remote control	18	5.8	5.8	5.8
Services for men and women – Search – Home and housing	15	4.8	4.8	10.6
Services for men and women – Search – Child and family	14	4.5	4.5	15.1
Services for men and women – Search – Hobbies and leisure	10	3.2	3.2	18.3
Services for men and women – Home and housing – Child and family	10	3.2	3.2	21.5

5. Discussion

The purpose of this research was, first of all, to examine the formation of price perception of mobile services amongst the mobile service customers. Special interest was on inspecting the differences in price perception between the three customer segments: heavy users, moderate users, and prospective users. The segmentation was conducted on the grounds of customers’ usage amount of mobile services. The customer segments were found to have distinctive differences in the constructions of price perceptions.

Customers’ price perceptions were examined through factors that were expected to have a significant influence on the formation of a customer’s price perception. On the grounds of theoretical modelling there was conducted factor analysis for constructing consistent factors which would best measure the formation of price perception. After the factor analysis there was measured factors’ ability to influence on and explain customers’ price perception through regression analyses.

As a result of factor analyses there was found, in total, five factors - price sensitivity, satisfaction to operator’s services, innovativeness, readiness to invest more on mobile services, and innovation resistance. Especially price sensitivity, satisfaction and innovativeness had the strongest effects on price perception through all the three segments. Notable was rather remarkable differences between user segments when measuring explanation powers and significances of the factors. For example in heavy users’ and prospective users’ segments there was found only three significant factors explaining respondents’ price perceptions. Whereas, in moderate users’ segment there was obtained four factors with significant explanation powers.

As there was in the beginning noted, pricing of mobile services is causing problems for mobile operators. And as traditional pricing or charging models are not viable for mobile services, new charging models need to be created. For this purpose there is, however, needed detailed knowledge on how mobile service price are perceived and what are the dynamics behind the pricing in mobile service industry. Therefore we have set a special attention towards charging issue. The second emphasised area in this research was examining, how pricing/charging method affects to price perception of mobile services.

Price and pricing methods were found to be very important factors in mobile services business according to responses. Pricing method was found to have significant relationship with almost all the measured factors, and in all of the three customer segments. Especially fixed pricing methods influenced frequently on most of the examined factors. Also price transparency was found to be important variable affecting to customers’ price perception. A customer’s awareness of the charging methods of mobile services had throughout the segments and factors a positive influence on price perception. The more aware a customer was of the charging methods the lower was the perceived price of mobile services. Therefore it was found that charging method related issues are highly important in formation of price perception of mobile services.

For answering to the third research question there was examined bundle pricing strategy; and how it was influenced (and influencing) by different factors. There was especially observed how bundle pricing was affected by price perception and investment readiness factors. The relationships were measured through regression analyses which resulted prominent differences between the customer segments. The results were divided into two: heavy users and prospective users were witnessing rather strong relationships between the dependent and the independent factors; where as the moderate segment was not significantly influenced by bundle pricing. Especially heavy users had rather high correlation with price perception and preference for bundle pricing. This segment received an explanation power of 18 percent; meaning that 18 percent of a person’s price perception was explained by interestedness in bundled services.

Despite of the differences between the segments there was also found some similarities in explanation powers and correlations of the factors. For example the price perception factor was influenced significantly in all the three segments by price sensitivity and innovativeness factors. Thus, it can be stated that operator’s should segment customers according to these customer characteristics and offer them tailored services to their special needs. And in a case of preference for service bundles there was found much more variability especially between the moderate segment and the two other segments.

5.1. Heavy users

Price perception in heavy users’ segment was perceived quite expensive as the mean response was 2 (0=expensive; 6=inexpensive). The price perception was influenced and explained significantly by three factors presented in figure 36. In addition, there was found two demographic variables correlating significantly with a customer’s price perception. The relationships between the dependent variable and the factors are presented by multiple correlation coefficients, R. Meaning that the higher the R-value the stronger the relationship. By squaring the R-values there is obtained an explanation power. Explanation powers were presented in the previous chapters.

In the heavy users’ segment there was found three factors through which the price perception was formed: price sensitivity (R=.462), innovativeness (R=.314), and readiness to invest on mobile services (R=.252). It was a surprise that the satisfaction to operator’s services factor did not receive a significant relationship with price perception. Which could indicate that customers in this segment are either not using operator’s other services or they perceive operator’s other services to be rather insignificant to them.

Multiple correlation coefficients R .462 (R² .213) of price sensitivity factor was the highest amongst the three factors showing rather strong correlation between a customer’s price sensitivity and price perception. Price sensitivity factor explained over 21 percent of the variance in the price perception variable. Also a customer’s innovativeness influenced moderately to his/her price perception even though the explanation power was left to 10 percent. Even though the explanation power was prominently lower than in price sensitivity case, it can be still considered to be important in formation of a customer’s price perception. The third factor, investment readiness, had weakest influence on a customer’s price perception. And the low multiple correlation also resulted very low explanation power, only 6 percent.

There was found, in fact, only two meaningful factors explaining a customer’s price perception in heavy users’ segment. Therefore it should be pondered whether there were some other factors significantly influencing on price perception which were excluded from this study. Meanwhile, these two or three factors gives a rather good tools for predicting and explaining customers’ price perceptions and, especially, a good ground for further studies.

When examining variables affecting indirectly to customers’ price perceptions, there was especially interested in charging related elements. There was found that satisfaction to charging method had a clearly best correlation with price sensitivity, r=456. Meaning that the less price sensitive a person was the more inexpensive was his/her perceived price of mobile services. And charging related variables had rather strong and multiple relationships with the all of the three factors (see figure 36).

Charging method related variables were thus clearly indicating the high importance of choosing a right method for pricing and charging mobile services. Satisfaction to charging methods had normally a positive influence on the factors and therefore, also a positive relationship with price perception. Also fixed fee pricing was revealed to have an important indirect relationship on a customer’s price perception.

There was also found that service channel choices were strongly influencing to all factors. Channel choices were rather well indicating a customer’s price sensitivity and level of innovativeness. In average the more traditional service channels a customer was using the more price sensitive he/she was. Channel choices should be therefore considered as on important element indirectly predicting and influencing on a customer’s price perception.

Third variable influencing indirectly but widely on a customer’s price perception was usage frequency of fixed Internet connection. Usage frequency was correlating positively with the price sensitivity and innovativeness. Meaning that the more a person used fixed Internet connection the more innovative and less price sensitive he/she was. Demographic variables showed only some rather weak or moderate predictability as there was found only two significant variables – adults in the household and marital status.

As there was discovered that charging method related elements were strongly influencing on formation of price perceptions of mobile services, there was also measured customers preference for service bundles. Acquiring mobile services in bundles was seen positively amongst the heavy users as the mean answer was 4.12 (0=not interested; 6=very interested) with standard deviation of 1.778.

There was further examined this element through regression analyses which was hoped to bring us better insight to the factors explaining a customer’s preference for mobile service bundles (especially, as none of the demographic variables showed any significant explanation power). Therefore the factors presented in the figure 37 were expected to explain which elements are significantly influencing on a customer’s preference on mobile service bundles and especially, what is the relationship between preference for mobile services and price perception.

As seen from the figure 37 there were found four factors significantly explaining respondents’ preference for mobile service bundles: price sensitivity (R=.492), readiness to invest more on mobile services (R=.466), price perception (R=.424), and price transparency (R=.250). It was notable that the three of the factors were directly dealing with pricing issues, whereas the fourth one was predicting a customer’s willingness to invest more on mobile services. With these four factors there can be explained rather well a customer’s preference for mobile service bundles. As the cumulative explanation power would have been 61.8.

It is notable that multiple correlation coefficients, R, were rather high amongst the three factors (showing strong relationship with the dependent factor), but the fourth factor had prominently weaker relationship. It should be also noted that in a case of the three factors there appeared the same variables indicating indirect relationships than in the previous chapter. Variables predicting and explaining price transparency were not examined in this study and should be therefore taken care of in future studies.

FIGURE 37. Multiple correlation coefficients for the factors possessing significant relationships with customers’ preference for mobile service bundles.

To sum up, there was found three, out of four, factors influencing significantly on a customer’s price perception. Which addition there was noticed some indirect relationship with variables influencing price perception through the factors. Especially charging method related variables were seen to have significant and multiple important indirect influences on a customer’s price perception. Thus it was expected that there would be significant relationship with a customer’s preference for mobile service bundles and a customer’s price perception.

This relationship was confirmed by regression analyses which resulted four significant factors having a significant relationship with the dependent variable. There was found a significant and meaningful relationship between the preference for bundles and price perception. Therefore by launching bundle or fixed fee related charging methods there can be significantly influenced on customers’ price perceptions.

5.2. Moderate users

Regression analyses were used also in the moderate users’ segment for examining and identifying factors significantly explaining and influencing on customers’ price perceptions. In comparison to heavy users’ segment there were identified four factors with significant influence and explanation power on the price perception of mobile services: price sensitivity (R=.486), satisfaction to operator’s services (R=.327), investment readiness (R=.315), and innovativeness (R=.280). There was also found higher number of variables indirectly influencing on a customer’s price perception than in the heavy users’ case.

As in the heavy users’ segment, the price sensitivity factor obtained the highest explanation power on a customer’s price perception by explaining 24 percent of the variance in the price perception. And while satisfaction to operator’s services did not receive a significant relationship in the heavy users’ segment, in this segment it received the second highest multiple correlation coefficient and explanation power of 12 percent. And when accounting together the explanation powers of the factors, there was obtained a cumulative explanation value of 42 percent. This level of explanation power can be stated to be rather high especially when comparing with the heavy users’ segment.

Even though there were some notable differences between heavy and moderate users’ segments when examining the factors’ explanation powers, the most prominent differences appeared amongst the variables with indirect relationships. First of all, there was found rather high amount of different variables significantly correlating with the factors and thus having indirect influences on a customer’s price perception. Especially when observing charging method related variables, there was found two changes: first, the correlations were stronger; and second, charging method variables influenced significantly through the three factors which had strongest explanation powers on a customer’s price perception.

Therefore, charging method was noticed to be very important element indirectly affecting to customers’ price perceptions. Especially satisfaction to charging model and preference for fixed fee based charging models were most frequently affecting to the factors. Service channel related variables were also in this segment found to be importantly influencing and predicting indirectly on a customer’s price perception. Therefore, these are the elements which should be first considered when examining and improving customer based charging practices.

FIGURE 38. Formation of price perception in moderate users’ segment.

As there was found, figure 38, the importance of charging methods influencing on a customer’s price perception, there should be considered practical solutions for business practitioners. Especially when there was found multiple and significant relationships of fixed fee based charging methods with the examined factors, we have suggested a service bundling method fulfilling the customer needs. Fixed fee based pricing methods were noticed to affect positively to customer’s price perception of mobile services. It was expected therefore that by creating mobile service bundles and pricing accordingly there could be obtained the benefits of fixed fee pricing in mobile environment.

There was examined customers’ preference for mobile service bundles through regression analyses which was expected to bring us similar results than in the heavy users’ segment. But instead, there was found only one factor correlating significantly with customers’ preference for service bundles – price sensitivity. The multiple correlation coefficient for price sensitivity was quite sufficient, R=.359, and thus possessing rather strong influence on a customer’s preference for service bundles. Explanation power R² was moderate 13 percent.

This kind of result was highly unexpected as there was found multiple indications that fixed fee based charging methods might be very useful and demanded for this segment. And, furthermore, there was found no significant correlation between preference for service bundles and price perception. Meaning that, even though, there were indirect relationships between these two variables there was found no direct relationships.

As there was obtained only this one factor having a significant relationship with the dependent variable, it leaves open the question about the formation of a customer’s preference for mobile service bundles. But on the other hand, we though obtained a result that in this segment service bundling strategy would with high probability be insufficient. Also the one explanation for this result can be found from the fact that the even though the mean preference for mobile service bundles wasn’t prominently lower than the other segments, there was high standard deviation, 2.124.

FIGURE 39 Multiple correlation coefficients for the factors possessing significant relationships with customers’ preference for mobile service bundles.

In conclusion it can be stated that there was obtained good picture of the formation of the price perceptions in the moderate users’ segment. There was found four factors significantly influencing and explaining a customer’s price perception. Furthermore, there was also found clear and multiple charging method related variables indirectly influencing on a customer’s price perception. This was implying the importance of charging model used in mobile services. The more satisfied and the more aware customers’ were of charging methods in use the less expensive they perceived mobile services.

Even though customers in this segment seemed to prefer fixed fee based charging practices in closer examination there was witnessed that bundle pricing method was not influencing on neither price perception nor investment readiness. But before making any final conclusions there is needed further studies for examining whether there are some other factors influencing on a customer’s preference for service bundles which were excluded from this study.

5.3. Prospective users’ segment

In the prospective users’ segment there was obtained results which raised multiple further questions. The results in this segment were not stating as clear conclusions as in the two previous segments, rather there was brought issues that will need further studies. In the figure 40 there is presented a summary of the conducted regression and correlation analyses. There was obtained three factors significantly explaining a customer’s price perception in this segment: satisfaction to operator’s services (R=.298), price sensitivity (R=.241), and innovation resistance (R=.208). In addition, there was also obtained high number of variables influencing indirectly to a customer’s price perception.

As all of the significant factors received correlation coefficients under 0.3 also the explanation powers were left to low levels. The satisfaction factor obtained explanation power of 9 percent before adjustment. And if taken into account the number of independent variables the explanation power further decreases. Therefore, with these three factors there cannot be done any accurate prediction rather they are indicating some potential directions of customers’ price perceptions. This kind of result clearly indicates that there are most likely other factors (which were not included into this study) that have significant influences on customers’ perceived prices.

There was also found that, for the first time, the investment readiness was not found to have a significant relationship with a customer’s price perception. The satisfaction factor had taken its place. A third characteristic in this segment was the appearance of innovation resistance factor. In a case of other two factors respondents’ innovativeness were better described by innovativeness which had reversed in this segment. Though, a customers’ innovation resistance was however quite expected as also theory suggests that the late adopters are most likely to be least innovative.

When observing variables having indirect influence on a customer’s price perception there was found four variable categories: charging, channel choice, usage frequency of fixed internet, and demographic variables. Especially charging method related variables were again having moderate and multiple relationships with the satisfaction and price sensitivity factors. Also channel choices were rather well influencing indirectly on a customer’s price perception through the same factors. Thus, these two variables were (as in the other segments) best influencing indirectly on a customer’s price perception. Especially important variables seemed to be satisfaction to charging method and price transparency. It was therefore expected that by modifying a charging method to be as transparent and easily assessed as possible, would best satisfy customers in this segment and influence best (positively) on their price perception.

FIGURE 40. Formation of price perception in prospective users’ segment.

Amongst the prospective users there was obtained a mean answer for preference for mobile service bundles of 3.80 (scale 0-6) with standard deviation of 2.497. The preference levels have thus been rather constant through all segments. The most significant differences have appeared in standard deviations. In this segment there was received the highest standard deviation. As the standard deviation is this high it his difficult generalize as there might be high differences between different customer segments.

In figure 41 there is depicted correlations between the dependent variable and the factors. If comparing the correlation coefficients to heavy users’ segment there is first notion the strength of relationships which was prominently lower in this segment. Only price sensitivity and price transparency were noticed to have sufficient correlations, R=.404 and R=.359. There was also found that satisfaction to operator’s services did influence and explain significantly respondents’ preference for acquiring mobile services in bundles. And the most importantly there was obtained a significant relationship between price perception and preference for service bundles. And even though the correlation between these two variables was rather low the dependent variable has indirect influence on a customer’s price perception through the other three variables.

FIGURE 41. Multiple correlation coefficients for the factors possessing significant relationships with customers’ preference for mobile service bundles.

Prospective users’ segment was found to be rather much different compared to the other two segments. There were multiple differences which were most clearly witnessed by the factors significantly influencing on a customer’s price perception – different factors and low multiple correlation coefficients. But the variables indirectly influencing on a customer’s price perception was rather similar with the other segments.

When there was observed charging method related variables indirectly influencing on a customer’s price perception, there were emphasized two categories: price transparency and satisfaction to charging method. But still, the charging method was observed to be very important element affecting indirectly on customer’s price perception. Therefore, bundle pricing strategy was expected to be very useful and needed for this segment.

In examining customers’ preferences for mobile service bundles there was found significant relationship by the four factors presented in the figure 41. A customer’s price perception and preference for bundles had also direct influence on each others. But it should be remembered from the factor analyses that there was rather prominent cross-correlations between these factors which why preference for service bundles and price perception have multiplied relationships through these other three factors.

Finally, to present more vividly the results of the study and the differences between the customer segments, there has been constructed a comparison table. There is presented customers’ price perception on mobile services, significant factors influencing on price perception, and factors’ multiple correlation coefficients. In bundle pricing section there is presented the same information, and additionally the three most preferred mobile service constructs.

TABLE 61. Comparison table for customer segment based results.

Heavy users

Moderate users

Prospective users

Price perception

Mean price perception

(0=expesive; 6=inexpensive)

mean 2.00

std. 2.016

mean 3.08

std. 2.780

mean 2.20

std. 2.39

Factors with significant influence on customers’ price perception

I Price sensitivity, R 0.462

II Innovativeness, R 0.314

III Investment readiness

R 0.252

I Price sensitivity, R 0.486

II Satisfaction to operator’s

services, R 0.327

III Investment readiness,

R 0.315

IV Innovativeness, R 0.280

I Satisfaction to operator’s

services, R 0.298

II Price sensitivity, R 0.241

III Innovation resistance, R 0.208

Bundle pricing

Mean preference for mobile service bundles

(scale 0=not preferred;

6= preferred)

mean 4.12

std. 1.778

mean 3.78.

std. 2.124

mean 3.80

std. 2.497

Factors with significant influence on customers’ preference for mobile service bundles

I Prise sensitivity, R 0.492

II Investment readiness,

R 0.466

III Price perception,

R 0.424

IV Price transparency,

R 0.250

I Prise sensitivity, R 0.359

I Price sensitivity, R 0.404

II Price transparency, R 0.359

III Satisfaction to operator’s

services, R 0.266

IV Price perception, R 0.220

Most preferred service bundle combinations

I Relationship–Search–

Home and housing

services

II Relationship–Search–

Remote control

services

III Entertainment – Chat –

Relationship services

I Gender based–Search–

Home and housing

services

II Gender based–Search–

Child and family services

III Gender based – Home

and housing – Child and

family services

I Gender based–Search–

Remote control services

II Gender based–Search–Home

and housing services

III Gender based–Search–Child

and family services

5.4. Limitations of the study

As Kerlinger (1980) stated, “reliability is the accuracy or precision of a measuring instrument”, instruments used and applied in studies should be robust and free from time or condition based errors. The reliability of this research was confirmed by Cronbach’s alpha, which measured the internal consistency of items in a scale. Alphas were measured in factor analysis which confirmed the internal consistency by providing alpha coefficients well over 0.8 which was above the critical value 0.7 suggested by Nunnally (1978).

The reliability of this study suffered also to some extent from the fact that sample group was obtained from one data base. This caused some problems for the generalisibility of the results. Though it was improved by taking stratified random samples to avoid biased sample groups or emphasize of mobile service customers of certain characteristics. Also the rather low response rate caused a danger of systematic error, i.e. only certain type of mobile service customers have responded.

As research validity was defined by Ghiselli (1981) by stating that it refers to the extent to which a test or a set of operations measures what it is supposed to measure. It can be stated that in measuring validity we are concerned that we are measuring what we think we are measuring. For improving content validity the questioner was carefully constructed and revised by experts. But still there was left some problems which would have been able to correct by testing the questioner more carefully before the actual study.

The construct validity was improved by factor analyses. But as there were few factors that would have been obtained only one or two variables on communality level 0.3 the critical communality level was decreased to 0.3. To improve the validity of this study the critical value for factor communalities should be set on 0.5 level.

Another limitation of this study was its scope. This thesis examines various areas of consumer behaviour, pricing behaviour as well as the field of mobile service pricing methods. The area of mobile service behaviour and mobile service pricing is absent from large-scale studies and this research cannot provide any complete model or theory in this respect. A better study design than the present on would include a clearer statement of the research problem and focus on either pricing or behavioural issues.

Moreover, this study was concerned the statistical methods used. It has been argued that correlation and regression techniques are not deficient, but more advanced tools (e.g. neural networks, comparative analysis) have been tested and used in e.g. Internet research. Additionally, the design of the research variable and their levels of measurement were based on previous studies. Also some of the survey questions seemed to be hard to answer.

5.5. Contributions of the study

As there was rather few studies concerned pricing issues in mobile service business this study has made an attempt to provide better understanding on the dynamics of pricing of mobile services. From theoretical point of view there was produced a significant insight into formation of price perceptions amongst mobile service customers. There were identified factors that were significantly influencing on customers’ price perception. As there was examined customers’ price perceptions in the three different segments, there was obtained also segment specific knowledge. Moreover, there were identified significant differences in constructs of formation of perceived prices between the three segments. These differences confirmed the findings of Kollmann to be true also in mobile services business.

In addition, there was provided evidences about the importance of charging models in customers’ price perception. There was noticed indirect but positive relationships between price perception and charging method related variables. It was found that there could be influenced positively on a customer’s price perception by fixed fee based and transparent charging models. Charging method related variables possessed thus quite similar influences on customers’ price perceptions in the all segments.

The information provided in this research can help theoreticians to identify the special characteristics related to pricing of mobile services and thus, redirect research areas for future pricing studies. Therefore, through this research we are few steps closer to ability to construct conceptual pricing models for mobile services business.

For business actors this study resulted a better understanding on mobile service customers and how they perceive mobile service prices in comparison to their demographic and usage related characteristics. When there is knowledge on how price perceptions are formed and which factors are significant, it gives practitioners tools for more accurate price differentiation. Important was also foundation of the importance of charging methods for customers’ price perceptions. Especially the information of being able to affect positively on customers’ price perceptions by fixed fee based and transparent charging methods is essential.

The service bundle chapter was especially directed for business practitioners as it provided some concrete segment specific evidence on customers’ preference for mobile service bundles. And most importantly, in this chapter there were constructed mobile service bundles (of three services) that were most preferred by respondents. By this information there is offered for operators a starting point for segment specific mobile service bundling activities.

Thus, operators can make use of the information provided in this study by more customer based business models by offering different service bundles to different customer segments with different charging methods and prices. Business practitioners are also better of as they can better predict how different marketing activities affect to customers’ price perceptions.

5.6. Future research

Even though there was studied multiple areas relating to price perceptions of mobile services many issues were left unexplored and unanswered. Future studies are thus needed as this study only described basic relationship between perceived prices of mobile services and the factors significantly influencing on the formation of price perceptions. In the future there should be obtained more accurate knowledge on the relationships between the price perception and the factors presented in this study. Especially interesting would be to obtain knowledge on causal relationships. It would be very useful for theoreticians and practitioners to get information on how these factors influence and what is the cause and effect relationship between the factors.

Second future’s research area would be investigating, in more detail, on which conditions mobile service customers’ would be ready to invest more on mobile services. Especially interesting would be information on the factors influencing on customers’ readiness to adopt mobile services. As this research would relate closely to diffusion research it would provide highly useful knowledge on the diffusion dynamics in mobile services business. There would be also obtained information on differences between customer segments in terms of adoption readiness. This would help business practitioners to induce adoption processes of new mobile services through more accurate marketing activities.

Even though we were in this research referring few times to price sensitivity issues and its influence on customers’ price perceptions, there was not obtained detailed knowledge on this issue and dynamics behind customers’ price sensitivity/elasticity. Even price sensitivity is essential to take into account in pricing and charging activities there is little, if any, research on mobile service customers’ price sensitivity. Therefore it would be important to examine more closely how price sensitivity differs between the mobile service customers, which factors influence on the formation of the price sensitivity, and how it affects to actual price perceptions of mobile services.

The fourth area discussed in this research but requiring more detailed research was bundle strategy in mobile services business. In this study there was provided some insight to bundling of mobile services and how it would affect to customers’ price perception. But much was still left uncovered. Especially moderate users’ segment left many open questions on how customers’ preference for mobile service bundles was formed, which factors affect to adoption of service bundles, and how it affects to price perception? Therefore it is necessary to further examine the viability of service bundle strategy in mobile services business.

REFERENCES

Aaker, D. & Kumar, V. & Day, G. 1995. Marketing Research, 5^th edition. New York: John Wiley.

Adams W. & Yellen J. 1976. Commodity Bundling and the burden of monopoly. Quarterly Journal of Economics, 90 (3), 475-498.

Arnold, S.E. 2000. The joy of six Internet content revenue models. Econtent, 23(4), 14-24.

Bagozzi, R. 1980. Causal Models in Marketing. New York: John Wiley & Sons.

Bontis, N. & Honsan, C. 2000. Case study - The evolution of software pricing: from box licenses to application service provider models. Internet Research: Electronic Networking Applications and Policy, 10(3), 246-255.

Campell, G. & Fiske, D.W., 1959. Convergent and discriminant validation by the multitrait-multimehtod matrix. Psychological Bulletin, 56, 81-105

Chase, R. & Tansik, D. 1983. The Customer Contact Model for Organization Design, Management Science, 29(9), p. 1037-1050.

Chen, I.J., Gupta, A. & Rom, W. 1994. A study of price and quality in service operations. International Journal of Service Industry Management, 5(2), 23-33. MCB University Press.

Christensen L. 1997. Experimental methodology, 7^th edition.

Churchill G. 1995. Marketing Research: methodological foundations. 6^th edition, Orlando: Dryden press.

Churchill G. & Iacobucci, D. 2002. Marketing Reserarch: Methodological Foundations, 8^th edition.

Cohen, P., Ballersteros, D., Gregory, D., Amant, R. 1994. Regression can build predictive causal models. Computer science technical resport 94-15.

Cohen, L. and Manion, L. 1980. Research Methods in Education. London: Croom Helm.

Cooper, D. & Emory W. 1995. Business Research Methods, 5^th edition. Richard D. Irwin, Inc.

Cronbach, L., 1951. Coefficient alpha and the internal consistency of tests. Psychometrika, September, 297-334.

Cushnie, J. & Hutchison, D. 2000. Charging & Billing models for future mobile internet services. QofIS’2000 DMRG, Lancaster University, Computing Department, September.

Dean, J. 1950. Pricing Policies for New Products. Harvard Business Review, 28, 45-55.

Engel, J., Blackwell, R. & Miniard, P. 1986. Consumer behaviour, 5th edition. CBS College publishing.

Finch, H.J., Becherer, R.C. & Casavant, R. 1998. An option-based approach for pricing perishable services assets. The Journal of Services Marketing, 12(6), 473-483, MCB University Press.

Fishburn, P.C. & Odlyzko, A.M. 1997. Fixed fee versus unit pricing for information goods: competition, equilibrium, and price wars. AT&T Labs – Research, 6.12.1997 (revised version). Ryan C.S., Princeton University.

Fornell, C., Johnston, M. D. Anderson, E., Cha, J., & Bryant, B. E. 1996. The American Customer Satisfaction Index: Nature, Purpose, and Findings. Journal of Marketing, 60 (October), 7-18.

Foxall, G. 1995. Cognitive styles of consumer initiators. Technovation, 15 (5), 269-288.

Garson, G.D., 1982. Political Science Methods. Boston: Holbrook, 288.

Ghiselli, E., Campbell, J., Zedeck, S., 1981. Measurement theory for the behavioural sciences.

Gouvêa, M., Toledo, G. & Filho, L. 2001. The prices of mailing services evaluated by companies. Marketing Intelligence & Planning, 19 (4), 282-294.

Groth, J. 1995. Exclusive value and the pricing of services. Management Decision, 33 (8), 22-29.

Hayek, F. 1954. The use of knowledge in society”, The American Economic Review, Vol. 15(4), 21-30.

Ho, D. & Cheng, E. 1999. Techniques: quest for value mix. Managing Service Quality, 9 (3), 204-208.

ITViikko 2002. An article of Telia’s mms services, 12^th of December, p.15.

Jeitschko, T.D. 2001. Issues in price discrimination: A comment on and addendum to “teaching price discrimination,” by Carrol and Coates. Southern Economic Journal, 68(1), 178-186.

Jonason, A. & Eliasson, G. 2001. Mobile Internet revenues: an empirical study of the I-mode portal. Internet Research: Electronic Networking Applications and Policy, 11(4), s. 341-348. MCB University Press.

Jonason, A. 2002. Innovative pricing effects: theory and practice in mobile Internet networks. European Journal of Innovation Management, 5(4), 185-193. MCB UP Limited.

Kerlinger, F.N (1981). Foundations of behavioural research. Second edition. Holt-saunders, Japan.

Kollmann, T. 2000. The price/acceptance function: perspectives of a pricing policy in European telecommunication markets. European Journal of Innovation Management, 3(1), 7-14. MCB University Press.

Kotler, P. 1997. Marketing management: analysis, planning, implementation, and control. International (9^th)edition.

Laffont, J. & Tirole, J. 2000. Competition in Telecommunications. The MIT Press, London, England.

Levin, A.M. 2002. Contrast and Assimilation Processes in Consumers’ Evaluations of Dual Brands. Journal of Business and Psychology, 17(1), 145-154.

Lichtenstein, D.; Blonch, D.; Black, W. 1988. Correlates of Price Acceptability. Journal of Consumer Research 15 (2), 243-252.

Link, F. 1997. Diffusion Dynamics and the Pricing of Innovations. Doctoral dissertation, Lund University Press.

Lovelock, C. 1991. Services Marketing, 2^nd edition. Prentice-Hall International, Englewood Cliffs.

Miles, L. 1989. Techniques of value analysis and engineering, 3^rd edition. Eleanor Miles Walker.

Monroe, K. 1990. Pricing: Making Profitable Decisions. New York: McGraw-Hill Book Company.

Mylonopoulos N., A.; Sideris, I.; Fouskas, K.; & Pateli, A. 2002. Emerging Market Dynamics in the Mobile Services Industry – white paper. The Mobicom project, www.mobiforum.org.

Nagle, T.T. & Holden R.K. 2002. The strategy and tactics of pricing: A guide to profitable decision making, 3^rd edition.

Nordström, B. 2001. Value, Content, Partnerships and Revenues in the Mobile-Internet Era. Advanced Internet Applications, Mobile commerce. Available online from: www.wmrc.com/businessbriefing/pdf/mcommerce2001/ reference/MobileInternet5.pdf, last accessed 15May 2003.

Nunnally, J. 1967. Psychometric theory. Mcgraw-hill book company.

Jacoby, J. & Olson, J. 1977. Consumer Response to Price: An Attitudinal Information Processing Perspective. Chicago, IL: American Marketing Association.

Poropudas, T. (2000). Venäjän hurja mobiilikasvu jatkuu. DigiToday|tele, 03.02.2003, www.digitoday.fi.

Reichl, P. & Haidegger, W. 2002. How to Combine Edge Pricing and End-to-End Performance Management Efficiently. Telecommunications Research Center (FTW). Available online from: http://userver.ftw.at/~reichl/publications/ATS02.pdf, last accessed 5 March 2003.

Rogers, E.M. 1995. Diffusion of Innovations, 4^th edition. The free press, New York.

Ross, E.B. 1984. Making money with proactive pricing. Harvard Business Review, 84(11/12), 145-155.

Runyon, K. E. & David W. S., 1987. Consumer Behavior and the Practice of Marketing, Merrill Publishing Co., Columbus OH.

Salinger, M.A. 1995. A graphical analyses of bundling. Journal of Business, 68, 85-98.

Strouse, K.,G. 1999. Marketing Telecommunication Services: New approaches for a changing environment. London.

Tse , A. 2001. How much more are consumers willing to pay for a higher level of service? A preliminary survey. Journal of services marketing, 15 (1), 11-17.

Tucker, S. 1966. Pricing for higher profit: Criteria, method, applications. McGraw-Hill Book company.

Tull, D.S. & Albaum, G.S. 1973. Survey Research: a decisional approach. New York: Internet books.

Urban, G. & Hauser, J. 1993. Design and Marketing of New Products, 2^nd edition. Englewood Cliffs: Prentice-Hall.

Uusitalo, H. (1991) Tiede, tutkimus ja tutkielma, johdatus tutkielman pariin; WSOY, Juva.

Venkatesh, R. & Vijay, M. 1993. A Probabilistic Approach to Pricing a Bundle of Products or Services. Journal of Marketing Research, 30 (November), 494-508.

Zeithaml, V.; Parasuraman, A.; & Berry, L. 1990. Delivering quality service: Balancing Customer Perception and Expectations. The Free Press, New York.

Zikmund, W. 1991. Business research methods, 3^rd edition.

[1] The paradigm is to sell for two different prices, at least one of which is above marginal cost, two identical units that can be efficiently produced separately. The separability of production costs makes marginal cost easily observable and the inefficiency of deviating from it obvious. Because it is hard to see how a firm could succeed with such a strategy without the power to affect total industry output. Thus pre-empting the possibility of someone else’s making the sale to the “high-price” customer at a price closer to marginal cost, it is easy to attribute the deviation of price from marginal cost to the exercise of market power. (Levin 20002)

[2] This is the practice by which a purchaser who buys at the lower price is able to resell or divert to those who are willing to pay a higher price. In economic terms, the need to prevent arbitrage arises because the competitive solution does not naturally result in separating equilibria in which consumers pay different prices for different product configurations. Such a result happens where there is little, if any, cross elasticity between the different product configurations. (Denis C. 2001)

[3] It is claimed that there must exist differences in consumers’ price elasticities that price discrimination could be profitably applied. Though, according to Jeitschko (2001) “there are many exceptions to this condition and differing price elasticities of demand is not a necessary condition for price discrimination”. But generally this can be observed as a basis for price discrimination and if it is not taken as a basic condition the other characteristics must support the discrimination strategy enough strongly that it can be omitted.