Make Mine a Scotch and Soda! the Curious Case of Excessively High Brand Share: a Consumer Behaviour Conundrum

ABSTRACT - This paper explores the relatively unusual case of a beverage alcohol brand, a Scotch whisky in Thailand, which enjoys a 51 per cent total market share in volume and some 60 per cent in value. Moreover, 26 actively promoted Scotch whisky brands compete for this A$400 million market with the average drinker consuming 10.3 litres of Scotch per annum. Heavy users of the market-dominant brand drink 41.4 litres per annum of their favourite tiple. In a very liberal regulatory environment where all types of marketing and promotion are permitted, including beverage spirit television commercials, only three brands enjoy a market share greater than 5 per cent. With the assistance of 4 AC Nielsen Scotch Whisky Tracking Studies conducted in 1990, 1992, 1994 and 1996 the paper seeks an explanation of the dominant label’s excessive brand share. The answer would not seem to lie with the brand management school of marketing such as represented by Aaker and Keller. But the study does suggest that Ehrenerg was right when he wrote Athere are no strong brands and weak ones but only big brands and little ones@. However, the concept of Double Jeopardy does not hold the key to our curious case of excessively high brand share either. Rather, the conundrum seems to be solved by the simplest of all consumer behaviour models.

Citation:

Rujirutana Mandhachitara AA@ (2001) ,"Make Mine a Scotch and Soda! the Curious Case of Excessively High Brand Share: a Consumer Behaviour Conundrum", in AP - Asia Pacific Advances in Consumer Research Volume 4, eds. Paula M. Tidwell and Thomas E. Muller, Provo, UT : Association for Consumer Research, Pages: 230-238.

Asia Pacific Advances in Consumer Research Volume 4, 2001      Pages 230-238

MAKE MINE A SCOTCH AND SODA! THE CURIOUS CASE OF EXCESSIVELY HIGH BRAND SHARE: A CONSUMER BEHAVIOUR CONUNDRUM

Rujirutana Mandhachitara "A", Thammasat University, Thailand

[The following paper is based on data that were collected entirely by other people, very generous people who made them available to me on no conditions other that I do not sell it to a third party. Out of courtesy to my supporters, specifically but not exhaustively AC Nielsen, Thailand; The International Wine and Spirit Record; London and Media and Data Resources, Bangkok. I have referred to the dominant brand and its runner up in the Scotch whisky market in Thailand as just Brand A and B. The real identity of these brands, however, be immediately obvious to anyone who has visited the kingdom for more than a night or two.]

ABSTRACT -

This paper explores the relatively unusual case of a beverage alcohol brand, a Scotch whisky in Thailand, which enjoys a 51 per cent total market share in volume and some 60 per cent in value. Moreover, 26 actively promoted Scotch whisky brands compete for this A$400 million market with the average drinker consuming 10.3 litres of Scotch per annum. Heavy users of the market-dominant brand drink 41.4 litres per annum of their favourite tiple. In a very liberal regulatory environment where all types of marketing and promotion are permitted, including beverage spirit television commercials, only three brands enjoy a market share greater than 5 per cent. With the assistance of 4 AC Nielsen Scotch Whisky Tracking Studies conducted in 1990, 1992, 1994 and 1996 the paper seeks an explanation of the dominant label’s excessive brand share. The answer would not seem to lie with the brand management school of marketing such as represented by Aaker and Keller. But the study does suggest that Ehrenerg was right when he wrote "there are no strong brands and weak ones but only big brands and little ones". However, the concept of Double Jeopardy does not hold the key to our curious case of excessively high brand share either. Rather, the conundrum seems to be solved by the simplest of all consumer behaviour models.

THE CONUNDRUM

This paper is an exercise in armchair marketing detective work. It attempts to solve the case of how Brand A achieved its towering 51 per cent market share (see Figure 1 for full details). The investigation got off to a very uninspiring start and I partly blame this early disappointment on the very thin indeed literature available on excessively high- brand share labels in repeat purchase markets. When, as you will learn, the black art of marketing lore failed to explain Brand A’s success, and of course, Brand B’s failure, I turned to the white science of consumer behaviour and its bright new laws. After 4 of these were found to be in opposition to the data I have had assembled, I thought I had better stop and take up crocheting. Even alchemy started to seem like junior high chemistry compared to discovering how Brand A did what you will read it did. One positive thing that resulted from Brand A’s exploits was a goal kick for consumers. They were up and running without regard for the sensibilities of the academy.

Being a bit of an anthropologist and unable to find the successful alchemist’s prescription, I did not even have to move from my armchair, to reach for inspiration from the bookshelf in a dusty copy of S.F. Nadel (1957) "The Theory of Social Structure". Nadel as a field anthropologist found that his ethnographic studies of his African tribe of adoption became much more effective when a "critical incident" happened. Such an event might be an unexpected death or even an incestuous marriage, for example. When a critical incident occurs the social structure of the tribe opens up and becomes transparent to reveal fundamental relationship which would be missed by the normal complexity of daily life.

It came to me that Brand A’s excessively high-brand share was a critical incident of sorts for my work as a student of marketing and consumer behaviour. A transparency did occur and the most fundamental of relationships between the consumer, his landscaper the brand manager, the scholar and his theories became more apparent. I did find a methodological answer to the conundrum which so preoccupied me, and I found it in a most unexpected place. But still I think my investigation raised more questions than it answered, and I feel I have just about "Come out by the same Door I went".

I have added a postscript to the paper because in the opinion of a reader I respect but who has absolutely no tolerance of ambiguity. He maintains that I had no right to lead him by the nose through most of this paper and then leave him ignorant of the reasons for Brand A’s success. On the other hand, I thought the discovery of the importance of stated "brand liked best" and its corresponding brand share was the answer, not unlike Ehrenberg’s "gravity in physics being a weak force compared with electricity, yet holding the world together", p.51. My respected reader disagreed and said it was more likely giving him an Agatha Christie detective novel knowing that the two final pages were missing.

THE MARKET

Thailand is second only to the US in terms of Scotch whisky consumption. In 1996, the last year of analysis treated in this paper i.e. before the Thai currency devalued in mid-1997, the value of the Scotch whisky market in Thailand exceeded A$400,000,000 (15 million bottles) at retail prices. Thais, particularly Bangkokians (10 million population), and other urbn middle class and upwardly mobile males prefer Scotch to all other alcoholic beverages. Most affluent Thai Scotch whisky consumers usually starts with standard grade (aged 3-5 years) and quickly trade up to premium grade (aged 12 years) by their early twenties. Also large drinkers of self made men graduate to Scotch from locally distilled brown spirits (AC Nielsen, Media Index 1996). A 75 cl bottle of premium Scotch cost A$26 dollars at the end of 1996 and a similar size bottle of standard Scotch cost between A$10 and 15 (International Wine and Spirits Record 1997).

Figure 1 shows the evolutionary structure of the drinking spirits market over a nine year period from 1988-1996. In this respect it is not particularly surprising, showing as it does the trading up and new entries into a more sophisticated market. The growth of the Scotch whisky market is shown by the size of the circles which are expanding at a strong rate. The size of the circles represent the retail sales of some 26 labels of Scotch as reported by regulation on a monthly basis collected and collated in the International Wine and Spirit Record, Thailand Edition (1986-1996). What is remarkable about the document is Brand A’s excessively high brand share as shown by the stripped areas of the circles. All the 26 Scotch whisky brands are represented in Thailand by some 10 sole importers and distributors who very actively advertise and promote their products. Advertising production values are also high and alcoholic beverage advertising is allowed in all media, including television.

FIGURE 1

THE ANATOMY OF AN EXCESSIVELY HIGH-BRAND SHARE MARKET

PLATE 1

TYPICAL PREMIUM BRAND PRINT ADVERTISING

And it is here that we encounter the first curiosity of the Thai Scotch whisky market, almost all premium brand advertising is very similar in message and execution. Aspirational, boy meets girl, decides to drink the night-away with her rather than embrace the more obvious alternative. See Plate 1: Typical Premium Brand Print Advertising.

While the standard brands often make use of humour in their advertising, no creative director would dream of compromising the prestige of a premium Scotch brand with levity. Brand marketers in Thailand were happily unaware of Nelson’s (1999) research which empirically establishes that spirits advertising has no effect on demand, only on brand share movements

TABLE 1

CONSUMPTION OF SCOTCH WHISKY IN STANDARD 9 LITRE CASES BY CATEGORY AND SEGMENT BRANDS A AND B: 1988-1966

And the similarity in marketing mix elements for premium Scotch whisky brands is nearly ubiquitous. Channels of distribution are identical for all brands and every standard and premium product is present in virtually 100 per cent of liquor stores and other licensed retail outlets. Sales forces are organised in very much the same way and very effectively so, to best service the regulated distribution system. Salesmen are remunerated according to much the same formula of salary and incentives. During a salesman’s career it is not unusual for him to have worked for 2 or 3 of the major Scotch importers and distributors. Terms of trade are also very similar between the 10 marketing companies and all run variations of basically the same trade promotions (lucky draw for a Mercedes) and consumer promotions (lucky draw for 7 Honda Civics). Merchandising offers more scope for innovation but one brand’s breakthrough one month will tend to be followed by an improved imitation within two months (McLaughlin and Moore 1997).

As illustrated in Figure 1, Brand A has come from only a moderate leadership in market share of 22 per cent, to Brand B of 18 per cent in 1986 to a 51 per cent of the total Scotch whisky market in Thailand in 1996. A market share of 51 per cent of a A$400 million market made up of 26 keen competitors does seems somewhat excessive!

Perhaps one of the more curious things about Brand A’s market growth is that the label has increased 371 per cent in volume during the 9 years under review. Brand B’s category sales volume increased only 120 per cent over the same nine years and its share of the Scotch whisky category actually declined 3 percentage points from 18 to 15 per cent by 1996. Table 1 refers.

There are two more elements of the Scotchwhisky market which may help explain first its internal similarities and lack of differentiation, and second may help us to better understand its high brand share. First, when consumed long as Scotch universally is in Thailand, with slightly salty soda water and ice made from unfiltered tap water, most brands tend to taste alike and only very few drinkers can even identify their own brands in a blind test (Roa and Saifon 1983). Second, of the 26 actively competing standard and premium labels only 3 have been able to achieve a 5 per cent or more brand share. (International Wine and Spirit Record 1986-1996). There are two principal questions which have pricked the writer’s curiosity and represent the heart of this paper. Why did Brand A do so well, and why did Brand B do so relatively poorly over the nine years under review?

One approach and possibly the most influential is the marketing management "building strong/powerful brands" school represented by Aaker (1996), Keller (1998) and de Chernatony and McDonald (1998). These writers who emphasise the efficiency of well deployed marketing mix variables with which to create powerful brands do not shed much light on our quest either. Such prized brand management motherhoods as pricing, positioning, promotions, advertising and distribution which build competitive advantage in so many other markets, do not appear to do so in the case of Scotch whisky in Thailand. Figures 2 and 3 refer.

Still, as Ehrenberg would courteously remind us, we should already have known there are no strong brands and weak ones but only big brands and little ones (Ehrenberg et al 1997). But this case does not lend itself to the Ehrenbergian school of analysis either. Although it is axiomatic that high-share brands receive extra repeat purchase loyalty as opposed to small market share brands which tend to have fewer buyers who tend to buy less often. This disadvantageous position for a lower share brand is referred to as Double Jeopardy (Ehrenberg 1988). Fader and Schmittlein (1993) have researched the "repeat purchase premium" (Ehrenberg, Goodhardt and Barwise 1990) in their study of high-share brands using the Dirichlet Model on Japanese and US panel data. They demonstrate that high-share brands will receive measurably, but still incrementally low, more repeat purchase loyalty than the model would allow. However, the disproportionately high market share of our product under review in Thailand takes it far beyond the parametres established by Fader and Schmittlein.

FIGURE 2

ADVERTISING SUPPORT : 1989-1996

FIGURE 3

RECOMMEND RETAIL PRICE: 1989-1996

DOES IT PAY TO ADVERTISE?

Figure 2 should be considered in the light of the fact that advertising media spending for both Brand A and B only really deviates significantly in 1996. In 1995 they were closer than had been in several years. It is tempting to take one of the advertising threshold theories, such as Little’s (1979), which stipulate a threshold beyond which brand media weight produces more significant results to scale on spending. But the existence of such thresholds is quite doubtful and from our point of view even if Brand A crossed one it could only have been in 1996, too late to explain the unrelenting growth of an already very high market share. Indeed, Simon (1980) and Simon and Arndt (1980) seem to reflect our own possible conclusions when they report there are no discernible economies of scale in advertising that favour large brands.

Moreover, Figure 2 should be seen in the context of total the Scotch whisky category advertising (26 brands) which reached A$21 million in 1996. Of this total Brands A and B contributed jointly only 40 per cent. For 1995 the Brand A and B share of total above-the-line media spend was 45 per cent of the category. Table 2 below demonstrates how the level of media spend is unlikely to be a major factor in explaining Brand A’s excessively high brand share.

Indeed, the relatively insignificant role of dvertising in building Brand A’s leadership position and inhibiting the growth of Brand B would seem to support Raj (1982) that advertising exposure is more effective in increasing volume purchased than in promoting brand switching. This viewpoint is supported by Tellis (1988) who observes that advertising is an ineffective method in precipitating brand switching, "Overall, advertising is one of the less important determinants of purchase behaviour", p.142. We concur with Tellis in respect of the case under discussion here.

TABLE 2

SCOTCH WHISKY BRAND PURCHASES & MEDIA SPEND: 1988-1996

TABLE 3

ANNUAL ELASTICITIES AMONG BRANDS A AND B

THE COST OF PRICE

As Figure 3 shows, the period 1991-1993 saw very rapid price growth for both Brand A and Brand B, but particularly Brand B. During these 3 years Brand B maintained a 13.5 per cent premium over the market leader’s recommended retail price. However, most of the premium was dealt back to consumers as promotional gifts with purchase. Subsequently they pursued a virtual price parity policy. Table 3 below examines the price elasticities for the two brands during that period, the results of which are somewhat counter-intuitive (Manning, Blumberg and Moulton 1995).

There are many published studies which attempt to explain the consequences of a price premium or a price increase in terms of the product’s level and type (see Roa & Monroe 1996 and Shakar 1996 for a partial review of this extensive literature). Roa and Monroe are also instructive on maintaining super high price premiums such as we encounter with Brand A and B. Typical of the top end of the genre is Krishnamurthi and Raj’s (1991) study of relationship between consumer price eleasticity and a number of consumer behaviour dimensions, based on brand loyalty. Their data base is very robust comprising two panels; one from BURKE and the other from IRI. In a 16 row (elasticities) and a 9 column (brand dispositions) matrix, only 4 of the 144 cells contain a positive price elasticity. On the other hand Brand A and Brand B display all positive demand elasticities in Table 3. This is a fairly rare outcome.

PAUSE FOR REFLECTION

On balance the work of consumer behaviourists permits us a somewhat better understanding of the rather unusual phenomenon we have under review. But there are still numerous disparities in this data between what consumers say about their brands, particularly in the many ways that we are accustomed to asking them in our research studies, and their actual behaviour. Much of the theory of consumer behaviour and its resultant working models rely heavily on what consumers think, feel, aspire, remember, plan and mentally trade-off, together with a lost of other psychological and neural processes. As researchers we have to capture, analyse and interpret this data, however indirectly, through what consumers tell us, again however indirectly, about what is going on in their heads. This is all fine and well. But when consumers tell us one thing and do the bi-polar opposite, not on just one but several dimensions of measurement, we should have cause for concern. In many ways this is particularly worrying when the inconsistent explanation is unlikely to be methodological artifact, with data collected as it has been by one of the world’s most trustworthy research companies AC Nielsen (1990, 1992, 1994 and 1996).

Let me demonstrate the problematical natures of the point made in the previous paragraph in respect the apparent failure of advertising to effect changes in brand imagery. Table 4 refers.

The statement versus behaviour inconsistency can be easily shown by examining the changes in the percentage of respondents agreeing with some 17 brand image statements. These have not changed over the period 1990-1996 during the four Nielsen studies. What were sample "brand image" measures in 1990 as Miller and Berry (1998) point ou, their importance is now reflected in the terms "perceived advantage versus competition" and ultimately "consumer bonding" (Dyson Farr and Hollis 1996). However no matter what we call them brand imagery, remains one of the work horses of any marketing research omnibus questionnaire. Their changing scores are used by brand managers to indicate the impact of advertising and other marketing mix variables. This is why brand imagery is so important.

TABLE 4

COGNITIVE, AFFECTIVE AND CONATIVE IMAGE SCORE: 1990-1996

Although Table 4 contains quite a substantial amount of information, most of it is unremarkable for several reasons. First Brand A starts out much stronger in 1990 image terms than Brand B, which somewhat surprising given their 1986 competitive brand shares of 22 per cent versus 18 per cent. But as we are not concerned with etiology here we are obliged to take this as our point of reference. Second although Brand A did significantly better than Brand B on most of the 17 image dimensions, Brand B scored consistently better on 3 cognitive and 2 conative scales. Although the cell numbers change over time, sometimes substantially but mostly not, the variance row numbers (in Italics) from one period to the next are marginal. This is highlighted by a comparison of total brand image scores on all 17 dimensions between Brand A and Brand B at the beginning of our study and at the end. Table 5 refers.

TABLE 5

BRAND IMAGE CHANGES: 1990-1996

TABLE 6

BRAND DRUNK LAST AND BOUGHT LAST: 1990-1996

The above effect provides pause for reflection as to whether a similar picture may not have drawn for us by consumers when they answered the brand image questions if Brand A had not spent A$21.4 million and Brand B A$12.2 million on advertising at all over this 7 year period.

The very foundation on which the study of consumer behaviour is built is a belief in the reliable connection between a consumer’s stated (no matter how elicited) beliefs, feelings, predisposition to act, his experiences, his preferences, the images he has of things, his hierarchical ordering of alternative choices, among others in relation to and his actual and manifest behaviour. When these connections cannot be even vaguely established we have should perhaps recognise cause for concern about the power of our conceptual framework which underlies the established methodology. This is, of course, an unusual market situation we are analysing but still when critical leading-edge indicators of brand standing, and therefore ultimately patronage; to wit brand image, spontaneous brand awareness and television commercial recall for both Brands A and B do not seem to be associated at all with actual media spend as analysed earlier in Table 2, we may have a problem.

The above does not imply the fieldwork for these surveys has been poorly conducted for this is clearly not so for reasons of external and internal reasons. Further to the second, when asked in two very different parts of the questionnaire which brands they had drunk last and the brand they had bought last, respondents answered in precisely the way they would have been predicted them to do because the event involved the conscious act of brand choice.

CONCLUSION

The case of Brand A’s increasingly excessive high-brand share versus the failure of Brand B, or any other of the 26 competitive labels actively participating in the Scotch whisky market growth during the period 1988-1996, cannot be explained in any way by the usual group of suspects: marketing variables (advertising spend and price positioning) and the more established consumer behaviour tools (brand imagery, brand salience and spontaneous brand advertising recall).

The study of consumer behaviour has become so complex in recent years, particularly with the proliferation of models and constructs that the essential simplicity of the mechanism of consumer choice has perhaps been obscured. This can be eloquently demonstrated by the very close correspondence between "brand bought last" and "brand drunk last" of the periods Table 6 or example.

Could this rather exceptional case of the Scotch whisky market in Thailand provide us with an opportunity to sense something more fundamental? I believe it provides a portal for us to take note that perhaps the very complexity of explanatory models can under certain circumstances obscure a simpler and even more predictive way of understanding consumer choice.

The textbook "Consumer Behavior" by Engel, Blackwell and Miniard (1968) launched the discipline through being the first comprehensive compilation of theoretical and empirical materials in consumer behaviour. I do believe however that there were at least two earlier published American consumer behaviourists than Engel et al.. Jonathan Bareham (1995) describes how in an experiment in the early thirties of the last century, an anthropologist called La Piere (1934) traveled around the USA with a Chinese couple during a time when anti-Chinese racial discrimination was widespread. Of the 200 hotels and restaurants they visited, they were refused entry only once. He later wrote to all the places they had dined and stayed at asking them if they would accept Chinese; 90 per cent replied that they would not. La Piere demonstrated as we have seen in our study that the relationship between attitudes and behaviour can be very inexact. Although a word of caution should be added that both the La Piere experiment and our own case represent very out-of-the ordinary situations.

In our failed search for an explanation of Brand A’s excessively high-brand share we did fall upon the simple option of examining answers to the question "which brand do you like best?" As Table 7 shows the simple statement of preference for a brand i.e. "brand liked best" is the best predictor of purchase intentions and brand share we have encountered.

The key to the Thai Scotch whisky brand share conundrum seems to be that consumers simply drink what they like to drink best. No bells and no whistles! Or as perhaps as America’s first consumer behaviourist, Abraham Lincoln (1964), put it:

"People who like this sort of thing will find the sort of thing they like."

TABLE 7

ASSOCIATION BETWEEN BRAND LIKED BEST AND MARKET SHARE

POSTSCRIPT

Like nature, many scholar marketing and consumer behaviour abhor a vacuum, and I really do wish I could appease their discomfort over this vulnerability. For example, I could put before the reader such statistics as provided by Phongpaichit and Baker (1998), i.e.:

1) the Thai economy was the fastest growing in the world of 10 years during the 1980s and early 1990s,

2) middle class incomes increased three fold during this period, and

3) Brand A tended to be drunk disproportionately more often in high consumption group drinking sessions.

These factors and many identifiable others certainly did have a role to play in the Brand A excessive high-brand share conundrum but in themselves are insufficient to provide an answer to our question.

I have to emphasise the essence of the answer to the Brand A’s high-brand share conundrum is what factors underlie the simple affect of "liking a brand best". If we can avoid the temptation to decompose the concept of "liking" into less meaningful component parts but the sum of which is greater than "liking best", we will make a significant start in understanding our conundrum nd perhaps others too.

In our case an understanding of what marketing, social and even historical forces had contributed to the brand image profile and the 65 per cent and 25 per cent "brand liked best" score first recorded in 1990 would be very helpful but it is not available and so we must use it as our point of departure. Perhaps we have to accept the real possibility of a Double Jeopardy effect in operation in respect of Brand A which will influence present and future consumers as to what they "like best".

REFERENCE

Aaker, David A. (1996), Building Strong Brands. New York: The Free Press.

AC Nielsen Co. Ltd. (1997), Media Index 1986-1996. Thailand.

AC Nielsen Co., Ltd (1997), Scotch Whisky Tracking Studies: 1990, 1992, 1994, and 1996. Thailand

Anonymous (1998), "How Superbrands Score Over Rivals", Marketing, October, 21, 25.

Bareham, Jonathan (1995), Consumer Behaviour in the Food Industry: A European Perspective. Oxford: Butterworth-Heinemann.

De Chernatony, Leslie and Malcolm McDonald (1998), Creating Powerful Brands in Consumer, Service and Industrial Markets. Oxford: Butterworth-Heinemann. 2^nd Edition.

Dyson, Pual; Andy Farr and Nigel Hollis (1996), "Understanding, Measuring and Using Brand Equity", Journal of Advertising Research, 36 (6), 9-21.

Ehrenberg, Andrew (1988), Repeat-Buying: Facts, Theory and Applications. New York: OUP. 2^nd Edition.

Ehrenberg, Andrew; Patrick Barwise and Gerald Goodhardt (1990), "Double Jeopardy Revisited", Journal of Marketing, 54, 82-91.

Ehrenberg, Andrew; Neil Barnard and John Scriven (1997), "Differentiation or Salience", Journal of Advertising Research, November-December, 7-14.

Ehrenberg, Andrew and Mark Uncles (1997), "Dirichlet-type Markets: A Review (part I and II)", working paper, School of Marketing, The University of New South Wales

Engel, James F.; Roger D. Blackwell; Paul W. Miniard (1995), Consumer Behaviour. Texas: The Dryden Press, 8^th Edition.

Fader, Peter S. and Schmittlein, David C. (1993), "Excess Behavioural Loyalty for High-Share Brands: Deviations from the Dirichlet Model for Repeat Purchasing", Journal of Marketing Research, November, 30, 478-493.

International Wine and Spirit Record (1986-1996), Thailand Report. London: The International Wine and Spirit Record.

Keller, Kevin L. (1998), Strategic Brand Management. Building, Measuring, and Managing Brand Equity. New Jersey: Prentice Hall.

Krishnamurthi, Lakshman and S.P. Raj (1991), "An Empirical Analysis of the Relationship between Brand Loyalty and Consumer Price Elasticity", Marketing Science, Spring,10 (2).

Lincoln, Abraham, quoted by C.W.E. Russell in "Collection and Recollections", Chapter 30.

Little, John D.C. (1979), "Aggregate Advertising Models: The State of the Art", Operations Research, July-August, 27, 629-667.

Manning, Williard G.; Linda Blumberg, Linda and Lawrence H. Moulton (1995), "The Demand for Alcohol: The Diferential Response to Price", Journal of Health Economics, 14, 123-148.

McLaughlin, Ian and Andrew Moore. (1997), Riche Monde Bangkok Limited: The Story of Johnnie Walker in Thailand, in Strategies for Asia Pacific: INSEAD. France.

Miller, Stephen and Lisette Berry (1998), "Brand Salience versus Brand Image: Two Theories of Adertising Effectiveness", Journal of Advertising Research, September-October, 38 (5), 77-82.

Nadel, S. F. (1957), The Theory of Social Structure. Melbourne: Melbourne University Press.

Nelson, Jon P. (1999), "Broadcast Advertising and U.S. Demand for Alcoholic Beverages", Southern Economic Journal, April, 65 (4), 774-790.

Phongpaichit, Pasuk and Chris Baker (1998), Thailand’s Boom and Bust. Chiang Mai: Silkworms Books.

Raj, S.P. (1982), "The Effects of Advertising on High and Low Loyalty Consumer Segments", Journal of Consumer Research, June, 9, 77-89.

Rao, Akshay R. and Kent B. Monroe (1996), "Causes and Consequences of Price Premium", Journal of Business, 69 (4), p.511-535.

Rao, Kuchibhotla Mallikarjuna and Somjate Saifon (1983), The Impact of Adulteration and imitations on Trade Mark Owner’s Business in Thailand, dissertation for master’s degree, December, Asian Institute of Thailand, Bangkok.

Shankar, Venkatesh and Lakshman Krishnamurthi (1996), "Relating Price Sensitivity to Retailer Promotional Variables and Pricing Policy: An Empirical Analysis", Journal of Retailing, Fall, 72 (3), 294.

Simon, Julian L. (1980), "On Firm Size and Advertising Efficiency: A Comment", Journal of Economic Literature, September, 18, 1066-1074.

Simon, Julian L. and Johan Arndt (1980), "The Shape of the Advertising Response Function", Journal of Advertising Research, August, 20, 11-30.

Tellis, Gerard J. (1988), "Advertising Exposure, Loyalty, and Brand Purchase: A Two-Stage Model of Choice", Journal of Marketing Research, May, 25, 134-144.

The Rubaiyat of Omar Khayyam (1859), translated by Edward Fitzgerald. London.

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