Categorization and Segmentation in Russian Consumer Behaviour

ABSTRACT - Russian consumer behaviour was analyzed basing on interviews taken in non-queue situations. Price differences for brands of shirts are not expected to be large. Price elasticity of the demand is rather high. The main demand forecast indicators are duration of the product search process and buyers' pre-shopping intentions. Brands are clearly eligible for psychographical categorization. Data presented suggest that advertising is efficient mostly for newly introduced models staying in fashion emulation or mass-fashion stage.



Citation:

Anna B. Kolbasova (1993) ,"Categorization and Segmentation in Russian Consumer Behaviour", in E - European Advances in Consumer Research Volume 1, eds. W. Fred Van Raaij and Gary J. Bamossy, Provo, UT : Association for Consumer Research, Pages: 244-248.

European Advances in Consumer Research Volume 1, 1993      Pages 244-248

CATEGORIZATION AND SEGMENTATION IN RUSSIAN CONSUMER BEHAVIOUR

Anna B. Kolbasova, Academy of National Economy, Moscow, Russia

ABSTRACT -

Russian consumer behaviour was analyzed basing on interviews taken in non-queue situations. Price differences for brands of shirts are not expected to be large. Price elasticity of the demand is rather high. The main demand forecast indicators are duration of the product search process and buyers' pre-shopping intentions. Brands are clearly eligible for psychographical categorization. Data presented suggest that advertising is efficient mostly for newly introduced models staying in fashion emulation or mass-fashion stage.

CATEGORIZATION AND SEGMENTATION IN RUSSIAN CONSUMER BEHAVIOUR

Considerable changes are now underway in Russia. One of the first results of liberalization process is that demand and supply are gradually coming to balance. However, enterprises demonstrate old patterns of behaviour; in spite of the fact that demand is shrinking in conditions of the overall non-payment crisis they are reluctant to cut down prices.

Similarly to entrepreneurs, psychology of consumers is strongly affected by traditional behavior patterns.

A few years ago the Russian consumer market was characterized by a poor and unstable assortment of commodities, steady prices, small differentials of prices among product types and time-consuming shopping. Because of structural shift in price levels in Russia buyers overestimated non-food goods and underestimated foods.

The research programme included:

B definition of product categories using psychographical and other product characteristics;

One could assert that due to the poor choice of goods, consumers are not sophisticated enough in distinguishing between product types: according to a survey made by Ovsyannikov 1989 et al, the indifferent-to-fashion buyers accounted for 29.2% of those polled.

B study of purchase choice criteria and predicting the new brand demand. Apart from purchase process characteristics, price, brand attributes and consumer demographical characteristics were considered.

It was assumed that owing to the product types scarcity, forecasts of consumer demand depended on that scarcity. On the contrary, probability of purchase by western consumers is relatively constant. According to a research on car models sales (Urban 1980), the real probability of purchase for those consumers who reported they would "certainly buy" a product is 0.6. In Russia answers are more uncertain due to irregular deliveries and high "expectation" time for better choice.

B analysis of advertising efficiency;

B investigation of buyers' responses to price. Because of generally low incomes the expected price difference for various product types is relatively small. Conversely, price elasticity is large.

The subject of our research conducted in 1988-1990 was the market of men shirts. The best firm shop of a well-known textile factory granted a unique chance to conduct field research of consumer behaviour. The general conditions of this market three years ago were most advanced in comparison with other products and are now to great extent the same. As before, the shirt market is characterized by mild competition, a rather diverse line of brands available without queues due to relatively high and free prices, and limited consumer budgets. However, the new factor of todayBthe high inflation ratesBprovides significant corrections into buyer behaviour. This factor should be held in mind while reading this paper.

METHODS

Samples

With respect to different objectives of the survey, as a total, about 30 samples of 100 to 500 respondents were examined.

Three series of in-shop intercept interviews were as follows:

1) With respect to a particular shirt model a questionnaire was worked out to study buyer response to price variations and to forecast future sales. Data were collected for 27 models with samples of 100 to 500 respondents.

2) With respect to the choice process among 15 models on sale, a sample of 211 respondents was derived.

3) The advertising efficiency was examined with a sample of 112 respondents.

In the first two series of interviews, visitors have been drawn uniformly during the whole week. To randomize the sample, the respondents were stopped at the shop entrance. Refusals accounted for less than 10%.

Data processing

The product categorization study included:

1) Categorization based on preference ratings. It consisted of a correlation matrix and its Thurstone scaling as well as factor and cluster analysis. For nearness estimates of the matrix demand cross-elasticity coefficients ("competitiveness coefficients") were compiled. A coefficient for a pair of brands indicated a percentage change in the sale of a brand if a new brand appears on the market. It is equal to a number of cases when a new brand has a greater preference rank than a former one divided by the number of consumers willing to buy the former brand regardless of its rank.

2) Products were categorized on the basis of ratings of product attributes with regard to the available products.

3) Product categories were formed on the basis of 27 new products surveys; each survey was done for a specific product introduction.

These three types of categorization provided categories that were interpreted in a similar manner. Spearman rank order correlation between the competitiveness characteristics of the second and the third types of categorization was 0.87. Nearness estimates: 17% within the clusters, 6% across the whole sample. F-ratios between segment characteristics is 7.03 with tabled Ftab =4.88 at a significance level a =0.05. That means that category differences are significant.

To position new models in categories of the first type of categorization, discriminant analysis was used. Sales forecasting was implemented with regression analysis. Advertising effect was studied with factor analysis. Price responses were approximated with Beta distribution.

The results were tested with the correlation significance measures and analysis of variances. The forecasts were validated with a separate sample.

PRODUCT CATEGORIZATION RESULTS

The main questionnaire was based first of all on behaviour and attitude characteristics. Here is a description of its blocks:

1 ) Buyers received a list of model attributes and were asked to rank them on importance. These characteristics are fashionability, originality, price level, utilization on special occasions, utilization for work, usage with a suit. In addition a respondent was asked to tell his/her requirements on shirt's colour, accessories, price, fabric.

2) Buyers were asked to range brand preferences.

3) Every buyer selected brands difference criteria from a given list of criteria and compared a first-rank-preference brand with "basic" ones using 5-point criterion scales. "Basic" brands covered main shirt styles and were specified by the researcher.

Clusters of brands

7 clusters of shirts were derived and interpreted as follows:

1) classic shirts; 2) one-Coloured, mannered shirts; 3) gaudy, two-coloured shirts; 4) fashionable shirts with small trimmings, for business; 5) decorated shirts; 6) white shirts; 7) shirts for a week-day.

Brand competitiveness (that is cross-elasticity coefficient) does not significantly correlate with the number of brands inside a cluster. Spearman's rank correlation coefficient is 0.21.

Preferences are determined by four main factors:

I ) Does a brand accentuate the user's manliness, willpower?

2) Is a brand decorated by accessories (in the framework of a given fashion)?

3) Should a brand be used in work or leisure time?

4) If a brand is gaudy, is it styled in an old-fashioned or in a classical manner?

Consumers' satisfaction

For satisfaction level analysis, every respondent was asked to rate the first- and the second-rank-preference brands using choice criteria.

The list of criteria is given below:

cheap-expensive; practical- non-practical; beautiful-plain; fashionable-obsolete; modest-mannered; original-flat; elegant-vulgar; for a businessman-for a professional man; for ceremonial occasions-for festivals-for week-days-for home; universal-for rare occasions; for the youth-for the old people; of urgent need-a similar brand is already acquired.

The share of buyers dissatisfied with non-price attributes was 42.6%.They accounted for 47.1 % of Moscow residents and 25.8% of non-residents, 25.8% of elderly people over 50.

Dissatisfaction is identified by two main factors:

1) Unpracticalness and insufficient fitness for week-days;

2) Lack of originality, brightness, smartness.

Spearman's rank correlation between dissatisfaction and competitiveness is 0.3. Though it is low, one can suppose that more decorated brands meet higher demand because of the manufacturer's unwillingness to produce these and correspond with more sophisticated tastes of consumers.

Consumer profiles

Groups of product consumers were identified and interpreted as follows:

1 . Hesitating buyers. These seldom visit the shop, cannot develop their style and aim at a single brand. If they miss the particular product they can't find an equivalent. Not risky, modest but pretend to be original.

2. Imitators. Conservative, old-fashioned. A little older than the first group. Prefer brightness.

3. More modest than the former group. They also do not plan the purchase but determine their style.

4. Consumers that make purchases for a certain occasion: a season, ceremony, etc. They look for products that would fit utilization situations.

5. Initiators. Innovative in choice. Visit the shop frequently, young, plan to buy an original brand, mainly Moscow residents.

6. Buyers of classic, high-quality original brands. Moscow residents, advanced in years, knowing their style.

The buyers segmentation did not identify clusters.

The factor of fashion appeared to be insignificant because consumers interpreted this attribute too subjectively.

SALES FORECASTING

Sales forecasts based on interview data were 1.5-6 times larger than in actual sales.

Three versions of forecast corrections were tested:

1 . The preliminary forecast was adjusted to a similar old product sales. Unfortunately, the determining of similarity is too complicated for clothes brands.

2. Equations connecting preliminary forecasts with actual sales with the help of various demand variables (share of buyers aimed at a particular purchase, purchasing experience, loyalty to a shop, attitude to brand attributes) were built.

3. Forecast accuracy was found to be dependent on product scarcity, in particular on product category competitiveness. This connection was used to correct preliminary forecasts.

Assessing demand factors

The following factors affecting the confidence level of respondents' answers were considered:

- share of people intending to buy a specific product before coming to the shop [aim];

- share of impulsive buyers that were formerly unaware of product but decided to buy it [impulsiveness];

- average frequency of shop visits per year [frequency];

- originality, defined through the price surplus that an average buyer would pay for the unique specimen of a brand [originality];

- share of those who had bought a similar product in the foregoing year [similar purchase];

- share of non-residents [non-residents];

- average buyer age [age];

- share of the dissatisfied buyers that had reproved a brand [dissatisfaction];

TABLE 1

CORRELATION OF THE FACTORS AND THE BRAND "FORECAST ACCURACY"

Designations of the criteria are put into squared brackets.

Regression analysis shows that the error of the sales forecast is first determined by the aimed people share.

Actual sales divided by respondents purchase expectations ("forecast accuracy") is equal to

Using category characteristics

"Forecast accuracy" correlates with the brand/segment competitiveness level.

Five brand clusters were ranged on "forecast accuracy":

1) Category of the highest "forecast accuracy" includes fashionable scarce brands of utilization on evident occasions.

2) Scarce original products requiring long search and series of futile visits.

3) Old-fashioned, decorated brands (two-coloured fabric).

4) Brands that do not fit any cluster.

5) Old-fashioned brands, bought mainly by non-residents.

This method of forecast correction provides better results: chi-square variable is 12.15 with the tabled chi-square 23.7 at significance level a = 0.05. Chi-square statistics for the method based on the factor of [aim] is 61.3.

In this paragraph our primary task is to explain the divergence between the predicted and the actual sales.

We have assumed that there exists a product space of potential demand combining products rated above a certain utility level. When a respondent considers a product during an interview he/she applies to this space. Furthermore, a space of real demand embraces only brands of preferences additionally filtered by buyers. Leaving alone non-contact respondents who refused to give answers (some of them accompanied real buyers or hurried because they had already made a purchase decision) we can set reasons for respondents' exaggerating the readiness for a purchase:

1) Some buyers imply that the product type is of the second preference rank.

2) Some of uncertain buyers dare not to buy new risky products.

3) Respondents misunderstand aquestion since they consider to buy a not-on-sale new product in the long future and are not serious.

4) When a product is available, it loses its image of a "perspective product", "a product of special interest".

ADVERTISING EFFECT FORECASTING

Study results obtained for advertising of a new product by radio and TV show that visitors drawn by advertisements are interested in brands similar to the advertised ones (they bought similar brands twice more often).

29% of the visitors-advertisement readers planned to come to the shop in the nearest future. The share of visitors who bought a new product among those who learned about it: through advertising-52%; through word-of-mouth-30%; who were not informed-15%. The total advertising effect on sales in that case was 30096'.

Nevertheless advertising can't provide a satisfactory explanation for the first-day-sale growth. The first factor of the growth specifies whether a new product style is adopted on the market. This factor explains mostly increase rate of shop visitors.

The second factor shows advertising intensity and affects mostly purchase frequency per visitor. The third factor is grade at which new product attributes meet the dominant fashion.

PRICE

Russian companies have not learned how to base price on the products' perceived value. They see product cost and their monopolistic position as the key to pricing.

In general, psychology of prices for Russian consumers seem to be similar to the western ones (see Scitovsky 1944, Helson 1964, Gabor & Granger 1966, Kalwany 1990, Monroe 1973, Dickson 1990). They use free price as an indicator of rarity and quality. According to our data, just 13% of consumers explained price variations for different shirt patterns by the producers cost. The others paid attention to product quality and attractiveness.

As for Russian peculiarities, I expect more negative consumer attitude towards price difference for similar products. Furthermore, "just" price is perceived mainly on the lowest price level among similar products.

Buyers' response to price was studied with the questionnaire based on recommendations by Lakhman et al (1987). Here is the text.

FIGURE 1

A BRAND: PRICE VERSUS DEMAND

Questionnaire

1 . Do you like the pattern presented here?

2. Which product type did you intend to buy before coming to the shop?

1) 1 didn't intend to buy anything;

2) 1 had no definite object and was apt to buy any thing I'd like;

3) 1 need another brand;

4) 1 need a similar brand;

5) 1 was looking for this very brand.

3. Is here in the shop any other brand that you prefer? Indicate the probability of your buying this brand: (no; most likely no; perhaps; most likely yes; yes).

4. At what maximum price would you agree to acquire the new brand today?

5. Would you buy two or more brand items? At what maximum price?...

6. Did you buy a similar product in the foregoing year? (yes, no)

7. How many times a year do you visit this shop?

The main purchase factors do not show significant correlation with the price. The correlation coefficient with the price of a similar past-purchase is r ~ 0.33 (significance level < 0.001).

Dissatisfaction does not affect price perception: average -prices for the "dissatisfied" and "fully satisfied" buyers are similar.

Actual sales correlate with price at r = -.25 (significance level 0.20). In consumers' purchase criteria ranking, price plays the secondary role among purchase factors. Price is not the main barrier to the purchase; it rather plays role when estimated and interpreted. However, within-segment price variation was 2 times lower than for the whole sample (6.5% against 12.4%). For respondents that preferred a new shirt, price elasticity was -15.3% (well-known Tellis' average elasticity is -1.76%!).

To study price elasticities, polygons for every product brand were constructed and approximated with respect to Beta-distribution.

3 function coefficients were adjusted with the criterion of chi-square minimization. Only one parameter explained for the shifts of different brands curves by average prices indicated by buyers. It validates that price perception is not correlated with other brand characteristics across the sample of consumers.

A polygon for one of the brands is presented on FIG. 1. One can clearly identify Gabor and Granger's price thresholds.

In 20 of 30 cases price was actually stated near a lower threshold. The lower threshold accounts for the highest daily sales volume. 'Me sections of high elasticity show that 64% of buyers are very sensitive to the price shift. The means, standard deviations and variations are summarized in table 2.

The last row gives estimates of sale shares per visitor obtained on the inter-view data.

CONCLUSIONS

Speaking about marketing methods worked out over the past 30 years for glutted (or heavy) markets it is essential to take into consideration the specifics of the Russian businessmen way of thinking.

The results we have obtained with our studies may be interpreted in the following way.

1. The Russian market is dominated by monopolistic state enterprises where mechanisms underlying deficiency of goods are different from those existing in developed market economics and are not necessarily governed by monetary factors. At the time of survey functional characteristics of deficiency (time of purchase, frequency of shop visits) were most essential in purchase process.

TABLE 2

PRICE LEVELS, DEMAND AND SALES UNDER DIFFERENT PRICE LEVELS

With good reason a buyer unconsciously settled the norm of the search time and later on made a purchase of a preferred product. The survey of Ovsyarmikov et al has shown that the share of impulsive purchases had steadily declined and that 74% of buyers made lists of planned purchases. In general, Russian consumers are more buyers than consumers because they pay too much attention to process of buying. Great share of purchases is made in advance so that consumption process drags out. Different marketing channels correspond to various social categories (black market people, currency holders, residents of cities of the first-order supply, rural residents, people using state administration privileged supply).

2. Price differences for various brands are expected to be relatively small. The average price elasticity is -15.3 (estimated basing on poll results), rather than - 1.76. Psychologically, Russian consumers are dependent on state subsidies and aid and not used to price hikes.

3. The study of intention to buy and other segmentation criteria gives no support to the hypothesis about inability of Russian buyers to select a purchase. We would rather accentuate the uncertainty of the search process, that manifests itself in the divergence of the predicted and actual data.

Similarly with the results of the western market research (Bass), segmentation on socio-demographic criteria provided no significant correlation with preferences. Fashion and originality characteristics were interpreted by respondents too subjectively. We failed to identify segments of shirts consumers. Obviously, a more thorough psychographical survey is required.

Study of dissatisfaction levels showed that quality of manufacturing caused the most of buyers negative responses.

4. Of relatively greater importance in Russia is the role of advertising and prices among marketing methods. According to the latest research (Jones 1990), 50% of advertising expenditures growth gave 5-15% of the sales growth. In our case an advertised item on TV, radio or newspaper provided up to 500% sales increase. Of the most importance is the advertising for newly introduced brands in the frameworks of well-known fashion trends adopted by the market.

REFERENCES

Brent Ritchie, J.R. (1974), "An Exploratory Analysis of the Nature and Extent of Individual Differences in Perception", Journal of Marketing Research, X1, Febr, 41-49.

Dickson, P.R., Sawyer, A.G. (1990), "Me Price Knowledge and Search of Supermarket Shoppers", Journal of Marketing, N4.

Gabor, A., Granger, C. (1966), "Price as an Indicator of Quality, Report on an Enquiry", Economica, 46, Febr, 43-70.

Helson, H. (1964). Adaptation Level Theory. NY: Harper & Row.

Jones, J.P. (1990), "The Double Jeopardy of Sales Promotions", Harvard Business Review, Sept-Oct., 148-150

Kalwani, M.U., Chi Kim Yim, Rinne, H.J., Sugita, Y. (1990), "A Price Expectations Model of Customer Brand Choice", Journal of Marketing Research, XXV 11, Aug.

Kaynak, E. (ed.) (1986). Global Perspectives in Marketing. NY.

Lakhman, I.L., Levin, M.I., Polterovich, V.M. (1987), "Mechanism of Contractual Prices", Izvestia AN SSSR, Series of economics, N3

Monroe, K.B. (1973), "Buyers' Subjective Perceptions of Price", Journal of Marketing Research, X, Febr, 70-80.

Ovsyannikov, A.A., Pettai, I.I., Rimashevskaya, N.M. (1989). Typology of Consumer Behaviour. Moscow: Nauka.

Scitovsky, T. (1944-1945), "Some Consequences of the Habit of Judging Quality by Price", Review of Economic Studies, X 11 (2), N32, 100-105.

Sewall, M.A. (1978), "Market Segmentation Based on Consumer Ratings of Proposed Product Designs", Journal of Marketing Research, XV, 557-567.

Sirgy, M.J. (1982), "Me Critical Survey of Theories of Self-perception Consumer Behaviour", Journal of Consumer Research, 9, N3.

Stanton, J.L., Lowenhar, J.A. (1974), "A Congruence Model of Brand Preference: a Theoretical and Empirical Study", Journal of Marketing Research, X1, May, 427-433.

Tellis, G.J. (1988), "Me Price Elasticity of Selective Demand: a Meta-analysis of Econometric Models of Sales", Journal of Marketing Research, 25, 331-341

Urban, G.L., Hauser, J.R. (1980). Design and Marketing of New Products.-Englewood Cliffs, N.J.

Wind, Y. (1978), "Issues and Advances in Segmentation Research", Journal of Marketing Research, 15, 317-328

----------------------------------------

Authors

Anna B. Kolbasova, Academy of National Economy, Moscow, Russia



Volume

E - European Advances in Consumer Research Volume 1 | 1993



Share Proceeding

Featured papers

See More

Featured

Growing Up Rich and Insecure Makes Objects Seem Human: Childhood Material and Social Environments Predict Anthropomorphism

Jodie Whelan, York University, Canada
Sean T. Hingston, York University, Canada
Matthew Thomson, Western University, Canada
Allison R. Johnson, Western University, Canada

Read More

Featured

Feature A Benefactor or A Victim? How Charity Appeals with Different Protagonist Foci Affect Donation Behavior

Bingqing (Miranda) Yin, University of Kansas, USA
Jin Seok Pyone, University of Kansas, USA

Read More

Featured

Associative versus Relational Processing: The Role of Elaboration in Evaluative Conditioning

Xiaomeng Fan, Northwestern University, USA
Galen V. Bodenhausen, Northwestern University, USA

Read More

Engage with Us

Becoming an Association for Consumer Research member is simple. Membership in ACR is relatively inexpensive, but brings significant benefits to its members.