Markov zincirleri ile pazar payı tahmini ve renkli televizyon pazarına ilişkin bir uygulama

Abstract

ÖZET Pazar payı; işletmeler arasında yarattığı belirgin farklar, stratejik yönetim açısından önemi kârlılık, yatırımlar ve pazar lama planlaması üzerindeki etkileri ile son derece önemli olduğu kadar pazarın yapısı ve işleyişi ile pazarlama bileşenlerinin op timum düzeylerinin belirlenmesinde de etkin rol oynamaktadır. İş letmeler açısından birçok stratejik değişkenin oluşumunda etkili ve kantitatif nitelikte bir karar verme kriteri olan pazar payı nın diğer yönünde ise sözkonusu payın oluşumunu sağlayan tüketici ler yeralmaktadır. Renkli TV pazarında, 1986 yılından sonra talepte görülen dü şüşler, birçok işletmenin pazar payı kaybına ve bir süre sonra pazardan çekilmesine neden olurken, bazı işletmeleri de dış pazar lara yöneltmiştir- 1983-89 yılları arasında pazar paylarında gö rülen hızlı 'değişmeler, büyük firmaların teknolojik ve mali açı dan güçlü duruma gelmeleri ile hız kazanan rekabet, küçük işlet melerin pazardan çekilmesine ve pazarın oligopolistik bir yapı ka zanmasına neden olmuştur - Bu denli dinamik bir yapıya sahip renkli TV pazarında yeralan işletmelerin 1990-95 yılları arasındaki kısa ve uzun dönem pazar payları Markov Zincirleri ve beraberindeki varsayımlar altında elde edilmiştir. Sonuçlara göre, 1990-95 yılları arasıda Telra, Vestel ve Philips pazar payları giderek artan işletmeler olurken, İES pazar dan çekilme, Bekoteknik ise pazar payı en hızlı düşme durumundaki işletmeler arasında yeralmaktadır. -vıı-

SUMMARY MARKET SHARE ESTIMATION OF COLORED TV WITH MARKOV CHAINS FOR THE PERIOD OF 1990-1995 Traditional and modern analysis accept that market share show the achivement level of a business in competitive market. - Market share identifies the fundamental reasons, differences and deviations among sales and - Market share is a comparison scale. This shows that the position of a product or a business in the market is not independent from the others. The reasons for increasing importance of market share are press effect, capital needs of businesses, costs, return on invest ments, economic growth, international relations and economics, profitability. Some basic strategies to increase market share are price, new product development, marketing quality, advertising and the other promotional techniques. Two different approaches to market share are from the tradi tional school which are (short-term) and long-term analysis. The traditional or short-term approach of market share 15 based upon : - The function of marketing expenditures - Sales-market share relationships. Long-term approaches to market share are `Profit Impact of Marketing Strategy` (PIMS) and `Boston Consulting Group Technique` (BCG). To investigate the relationship between profitability and market share, a major empirical investigation has been conducted called the Profit Impact of Marketing Strategy` (PIMS). The objectives of the PIMS study are to answer the basic questions of what factors influence the profitability and by how much. -vxxi-Two of the most important factors affecting profitability are market share and investment intensity. Businesses with a large market share have more experience than their competitors and also have lower total unit costs. Additional support for this had been observed with the PIMS data, businesses with high levels of market share have lower ratios of marketing expenses to sales. Another important factor of profitability defined by the PIMS study is investment intensity, the ratio of total intensity the lower the return on investment of the business. On the other hand, PIMS relies on economics of scale, market power and manage rial ability in the market. The main objective of a business to grow faster than the market to achieve a poweful share position and to increase current profit for market position and long-term profitability. As the market growth for these products slows, their requirement for cash is reduced. But with their high share position they are good cash earners and the excess cash can be applied to other growing markets. This approach has been developed by the Boston Consulting Group (BCG) who make use of the growth/ share matrix. Products in the business are classified according to their rela tive market share and the growth rate of the market they serve. The relationships between market share and marketing mix are most important points of the analysis to allocate optimal marketing strategies for a business. The product (both new product and existing product) /market scope of the business consists of the products it offers and the markets it decides to compete in. The product/market scope is an operational definition of the mission of the business and plays a key role in determining the profitability of the business. Price is one of the elements in the marketing mix. Which affects market share of a product. Pricing strategy is very important for a business because it has a direct effort on profitabilty both in the short term and long term. Decision rules for setting the advertising strategy and especially advertising budget are developed for the case in which market share is an additive function of previous market share and of current and previous advertising share. Planning in business or nunbusiness organizations requires an understanding of demand. In the marketing plan, analysis of market share is a prerequisite to the development of strategies and tactics that enables the organization to reach its goals. After the plan has been executed feedback hoops from control systems enable the revision of initial market share estimates. -ix-The demand and market share for economic goods and services are the results of decisions by individiuals who try to meet their physiological, social and individual needs within their economic limitations. Measures of potential demand are made by identifying the determinants of these decisions. The process of market share analysis lies on theories models and measures from demographical data, economics, sociology and phychology. The demographics of market share consist of locating unmet needs in time and space. Demographic dimensions include age, educational level, occupation and geographic location. Two demographic dimensions used in industrial marketing are Standard Industrial Classification (SIC) and number of employees. Demographic data are more available in secondary sources than other market share variables. Therefore demographic market share analysis have generally the lowest cost. Demographic segmentation provides promotional strategies because of media report circulations according to demographic classifications. Socio-economic market share analysis contributes to the marketing activities of product planning, the identification of new markets for existing products and the allocation of promoti onal effort. Product planning which is the beginning of a product's life cycle, requires an analysis of the economic conditions at three levels (general business conditions, industry conditions and the economic outlook for the consumer) to determine if the economic climate is suitable for the introduction of a new product. The Mustang meing a product-line extension for a consu mer good represents the required type of analysis. In the product planning for the mustang the demographic analysis revealed the potential for an automobile designed to meet the needs of the 18 to 24 year old market segment. It is necessary to analyze the general business and economic conditions within the industry and the purchasing power of the segment to support the introduction of such an automobile. In the early 1960's, whenFord was considering the Mustang, the general business outlook was uncertain. On the other Land, the demand for new cars for consumer use has come from three sources : new owners of cars, present owners buying additional cars and the replacement of cars presently owned. Moreover, the market segment has a potential for purchasing a new automobile. When a product reaches a stable level of sales, the best method is to locate new markets for the product. Industrial market share analysis is then developed using Censuses of Business which classify firms according to a standard Industrial Classifi cation (SIC). The identification of relevant variables (demographic, economic, social and socio-psychological) is simplified by the availability of multivariate statistical techniques. A technique -x-called the Automatic Interaction Detection (AID) algorithm is used for identifying demographic and economic variables for generic and brand share analysis. Economic models show a buyer's ability to buy a product or service and reflect.his buying motivations only indirectly. Early economic models of generic demand considered only the effect of price. Later models consider the effects of personal income, savings and credit terms. The relationships between sales and these independent variables are expressed as elasticities. Economic models may be used to estimate the market share for exis ting products. Because elasticities are sersitive to the product life cycle, and to trend, cycle, and seasonal effects, marketers should try to determine these effects when computing elasticities and when building marketing strategies upon them. The recent economic models of generic demand includes product attributes attitudes and purchasing intentions. Economic models of brand demand estimate the effects of the elements of the marketing mix on sales. These models introduce the problems of accounting for competitive effect and computing the cumulative effects of marketing. effort. The use of regression models specific measurement problems. These include : the identification of variables; the identification of the structure of the model; adjustment of the data to conform to common classifications such as geographic areas; the collection of competitive data for market share models; and the advertising- sales relationships. Society affects demand of market share in three ways. First, it meets the social needs of individuals as belonging new and love. Secondly, it describes goals and behavioral patterns for determining all individual needs such as physical, social, and psychological. Finally, it provides communication networks that perform important roles in controlling and changing behavioral patterns. Value differences among cultures become important to the multinational firm as it plans its marketing effort. Market segmentation and promotional strategics that use social variables require an understanding of the concepts of role and values. These concepts determine the effect that groups will have on an individual's demand for goods and services. Social class implies values that may affect demand but class variables have not proven to be good predictors of buying behavior. Reference groups are especially influential when a product is consumed publicly. Perceived social pressure to use a product may overcome perceived negative product attributes and result in the purchase of a product. An individual -may be influenced by a group in which he holds membership or to which he aspires. -xi-The concept of role in a group may be defined with the concept of role in a play. Large purchase decisions are difficult because of conflicts in group values and roles. In most societies, an individual can develop his own values, beliefs and personality. Measures of values are not used widely in marketing and personality variables lack the ability to predict buying behavior. On the other hand, beliefs are used widely in marketing such as awareness and unawareness. Social communication networks influence demand through perso nal influence and variations in media effectivenss during adoption process. Social patterns may influence strategies of distribution and pricing. The most important application of psychological concepts to demand analysis is the positioning of a brand in psychological and competitive space. The brand may be new or an old brand that is repositioned through product development and promotion. Positi oning requires first the identification of product attributes and consumer benefits that are important during the decision process and secondly the positioning of buyers' perceptions of competitive brands along these dimensions. In psychological terms, these brand positions are buyers' attitudes toward the brands. An attitude is a learned mental state that reflects a person's attraction toward or repulsion from an object, a person, or a concept. In marketing research, the current trend is to regard an attitude as containing two elements as beliefs and values. Attitudes are formed as the result of the processing of information which includes the acquisition, rationalization and forgetting of information. Perceived risk affects the search for and the pro cessing of information. Psychological theories of motivation need hierarchy, cognitive balance, communication and attitude change are central to an un derstanding of the dynamics of demand in the market place. Behavioral theories have their origins in hedonism. Motiva tion consists of two elements, a perception of needs and the evaluation of the means for meeting these needs. Needs may be arrayed as a hierarchy ranging from the physiological to the psychological. Behavioral models are in a constant state of refinement. The marketing planner must accept the fact that there is no general model of behavior and must develop his own models by using gene rally accepted variables and functions. The principal limitation of measures of attitudes topard objects is their inability to predict behavior toward objects. Most of the information to which a person is potentially exposed is filtered during the first stage of processing through selective -xii-exposure and selective perception. On the other hand awareness of a need reduces the filtering out of information. Theoretical models of information processing tend to stop with attitude forma tion. But Marketing models of the information process emphasize choice behavior and then ignore central processing completely by linking an awareness of an advertisement to final brand choice. The commonly used market share models are all based on the implicit assumption of a homogeneous population. However, studies of individual brand choice behavior tend to reject this basic premise. Individual consumers are the basic behavioral units and the market share is the composite process at the aggregate level. Panel data pertaining to a brand should be analyzed with the aid of a heteregeneous brand choice models. Then, consumers follow the Bernoulli process, the gravitational market share model should be used. When consumers follow the Linear Learning Model or divided between Bernouilli and Markov followers, the deviations in predictions made using the approximating homogeneous model in place of the heterogeneous one ought to be evaluated. If these deviations are small and our results are satisfactory when the market share is high, a homogeneous model can be used. If devia tions are significant, then while the homogeneous market share model will work well in the fitting of time-series data, its subsequent use as a basis of multiperiod forecasting and, as a structure for determining the optimal marketing strategy may be inappropriate. Use of homogeneous models may lead to errors in managerial decisions and in their anticipated outcomes. Markov process is a method of analyzing the current movement of some variable in an effort to predict future movement of those same variables. As a management tool, the Markov process has been used as a marketing aid forexamining and predicting the behavior of customers in terms of their brand loyalty and their switching from one brand to another. In this thesis, Color TV industry is the subject for analysis of the role of Markov process and how it applied in different mar ket problems such as competition between brand names and predict the future share of business. As markov chains are based upon the movement of customers, random sampling technique is used to reach sample size which will represent the whole customers having at least one colored TV. With using this technique, sample size (n) was obtained as 388 under 5 % degree of freedom. Firstly, the movement of customers from one brand to another and the number of customers and the actual exchanges of customers (gains losses) is illustrated as a result of data obtained from sample size (n=388). The whole matter is that simple analysis in terms of net gain or loss of customers is inadequate, it is needed more detailed analysis` concerning the rate of gains from and loss -xiii-to all competitiors. With such data, it is easy to make an effort to four problems which are concerned by that more than ten brands: - Predict the share of market a seller will have at some future time - Predict the rate at which a sell will gain or loss its share of market in the future time. - Predict whether or not some market equilibrium (constant or level market shares) will obtain in the future. - Analyze a business1 s efforts in terms of exactly what effect they are having on its gain and loss of market share. To solve for all problems, the rate of 8 lass, the number and the probability of customers unchanged (brand loyalty) were obtained. As the markov process offers us these tools for market analysis, by employing these tools, it is easy to draw more accu rate conclusion both marketing position, both present and future. Then transition probabilities which are nothing more than the probability that a certain business will retain its customers were computed. Because calculation of a complete set of these transi tion probabilities would require data on the flow of customers among all the brands, the rates at which all the brands gain new customers is needed. A matrix of transition probabilities includes for each brand the retention probabilities and probabilities of its loss of customers to its other all competitiors. The rows in this matrix show the retention of customers and the gain of customers.the columns represent the retention of customers and the loss of customers. The first-order Markov process is based on the assumption that the probability of 1990 depends upon the outcomes of the last year (1989) and not at all on any earlier buying behavior. A second-order markov process (illustrated in Appendix C) assumes that customer choices in 1991 may depend upon their choices during the immediate past two years. In turn, a third order process is based upon the assumption that customers behavior is best predic ted by observing and taking account of their behavior during the past three years. It is reasonable to assume in colored TV problems that a state of equilibrium might be reached in the future regarding market shares; that is, exchange of customers under equilibrium would be such as to continue -to freeze- the three market shares which obtained at the moment equilibrium was reached. As equilibrium is approached, the gains losses become smaller and smaller. In the case of our problem, the matrix of transition probabilities remains fixed, that the propensities of all brands to retain gain and lose customers do not change over time. That is to say, we will always -xiv-end with the same final proportion of customers no mater what the original share were. As a result of analysis, short and long term market shares for all the brands, which are in colored TV market having very dynamic structure, is estimated for the period of 1990-95 using with above processes, basic assumptions and some hypotheses. For the period of 1990-95,, It is shown that Telra, Vestel and Philips will have increasing market share continously although. Bekoteknik will have sharply decreasing market share and IES will be out market. -xv-