Brand Preferenceh
Brand Preferenceh
Brand Preferenceh
net/publication/313797929
CITATIONS READS
0 2,524
2 authors:
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by Leykun Birhanu on 30 March 2019.
Abstract: In every product category, consumers have more choices, more information and higher expectations
than ever before. To move consumer from trial to preference, brands need to deliver on their value preposition,
as well as dislodge someone else from the consumer’s existing preference set. The study was on Factors
Determining Consumer Beer Brand Preference in Addis Ababa, Ethiopia. Therefore; the objective of the study
was to assess factors determining consumer beer brand preference in Addis Ababa, Ethiopia. The finding from
the multinomial logistic regression revealed what factors determine the probability attached to respondents beer
brand preference. Accordingly; age, perceived beer quality, perceived social benefit, situational influence and
peer influence had positive sign and significantly affect the probability of preferring St. George beer. Whereas;
family size had sign and significantly affect the probability of preferring St. George. Moreover; advertisement,
situational and peer influence had positive sign and significantly affect the probability of preferring Habesha
beer. Whereas family size had negative sign and significantly affect the probability of preferring Habesha beer.
Furthermore; advertisement and situational influence had positive sign and significantly affect the probability of
preferring Walia beer. Whereas sex and marital status had negative sign and significantly affect the probability
of preferring Walia beer. Still there is untapped market potential that producers should take such as a market
segmentation strategy and design their products in a manner that make the products appeal to different
categories of individuals that can influence of personal factors on customer satisfaction. A potentially successful
strategy can be that which provides products that correspond to and appreciate customers’ social status and age.
It is also advised that any advertisement for beer brands should convey information about the advantages which
the brand being advertised would offer over other brands. Moreover; situational influence was found to be
significantly relevant to brand preference of beer, producers should in their advertisement emphasis social
groups. They should exploit this further through segmenting the market into distinctive social classes.
1. INTRODUCTION
Beer consumption in developing countries is often seen as one revealing (if crude) proxy for gauging
the strength of economic activity. If beer sales are high and rising, incomes and economic activity are
presumably growing strongly while the reverse should be true if beer sales are flat or falling (Access
capital research, 2010). Ethiopia‟s beer industry is currently comprised of eight companies are
When J=1, 1 is the Kx1 vector of unknown parameters, and we get the binary logit model.
The outcome or response probabilities of this study was categorized based the number of
respondents' preference to beer brand. Accordingly the top three highly preferred beer brands and the
remaining as others was possible outcomes or response probabilities.
Unbiased and consistent parameter estimates of the MNL model in equation (1) require the
assumption of independence of irrelevant alternatives (IIA) to hold. More specifically, the IIA assumption
requires that the probability of preferring a beer brand in one category by a given respondent needs to be
independent from the probability of preferring another brand in another category (that is, Pj/Pk is
independent of the remaining probabilities). The premise of the IIA assumption is the independent and
homoscedastic disturbance terms of the basic model in equation (1).
The parameter estimates of the MNL model provide only the direction of the effect of the
independent variables on the dependent (response) variable, but estimates do not represent either the actual
magnitude of change nor probabilities.
The magnitudes of the coefficients of MNL model are difficult to interpret. Thus, either we compute
partial effects, as in equation (3), or compute differences in probabilities. These results are easily obtained
by comparing fitted probabilities after multinomial logit estimation. The fitted probabilities can be used for
prediction purposes: for each observation i, the outcome with the highest estimated probability is the
predicted outcome. This can be used to obtain a percent correctly predicted, by category if desired
(Wooldridge, 2002).
Therefore, differentiating equation (1) with respect to the explanatory variables provides partial
effects of the explanatory variables given as:
J
P ( y j / X )
hK exp( X K )
.........................................................(3)
P( y j / X ) jK K 1
X K g( X , )
Where hk is the k element of h and
th
The marginal effects or marginal probabilities are functions of the probability itself and measure the
expected change in probability of a particular category with respect to a unit change in an independent
variable from the mean (Wooldridge, 2002).
Regarding the overall model fit, chi-square (i.e. the LR (likelihood ratio) test for the current model
compared to the null model) and the McFadden‟s Pseudo R-square will be used to validate the regression
output.
Table 3.1: Respondents Demographic characteristics Inferential and Descriptive Analysis Result
Inferential and Descriptive Analysis of Demographic Characterstics of Sampled Respondents
The statistical summary provided in Table 4.1 shows that male respondents 82% (216) is higher than
that of female-headed households 18% (45). The mean age of a typical respondents is about 38 years with
the youngest being 24 and the oldest 60 years old. Majority of respondent attained first degree and above
54% (140), 26% (69) attained primary to secondary education, 18% (48) have attained college diploma and
only 2% (4) have not attained formal education. On average, each respondent has a family size of three;
however the range varies from one to a maximum of nine.
Table 3-2: Inferential and Descriptive Analysis Result for Variables that Determine Respondents' Beer
Brand Preference
Descriptive Analysis and Non-Parameteric Tests By Using One-ANOVA and Pearson chi2 Test for Determinants Variables of Beer Brand Preference
The descriptive result shows that majority of 54% (142) respondents enjoyed good quality beer and
18% (47) enjoy best while only 25% (66) of the respondents perceive their preferred beer is of normal
quality, thus benefited. The remaining 2% (6) of the respondents perceive the beer they consume is of
inferior quality, thus not benefited. The Pearson chi2 and one-way ANOVA test for the perceived beer
The graph below depicted that respondents percentage beer preference for St. George, Habesha,
Others (Meta, Bedele, Harar & Dashen), and Walia is 39.85% (104) , 28.35% (74), 19.16% (50), and
12.64% (33) respectively.
19.16%
39.85%
12.64%
28.35%
Others(DashenHararBedeleMeta) Waliya
Habasha StGeorge
The result from these study is comparable with beers previous market share. Data collected by
Fortune in 2014 indicates 38% market share hold by BGI (St George), Heineken which owns Walia, Harar
and Bedele breweries holds 35% of the market and Diageo owner of Meta Abo Brwery, and Dashen
Brewery S.C takes third stage with 27% market share.
The graph below depicted that respondents consumption volume for 1 to 4, 5 to 10 and above 10
beers at a time is 70.5%, 24.14%, and 5.36% respectively. the Person chi2 test result shows that
respondents consumption volume level have statistically significant effect on preferred beer. Whereas, the
5.364%
24.14%
70.5%
The graph below depicted that respondents mean consumption year for St. George, Others (Meta,
Bedele, Harar & Dashen), Habesha and Walia is is 15, 6.56, 2.27, and 2 respectively. Moreover; the mean
consumption year of a typical respondents is about 8 years with the shortest period of consumption being 1
and the longest period 43 years. Moreover; the one-way ANOVA test result showed consumption year do
have statistically significant effect on respondents beer preference.
3.4 Multinomial Logit Estimation Result for Determinants of Beer Brand Preference
The likelihood ratio statistics as indicated by chi2 statistics are highly significant (P <0.0000),
suggesting the model has a strong explanatory power. We tested whether the assumption of IIA holds in
our model using the Hausman tests. The result consistently indicates that the assumption is not violated
and hence application of multinomial logit model is appropriate. The Pseudo R2 is 0.68, indicating the
specification fits the data well the variables included in the model explain 68% of the variation in the
respondents beer brand preference. The maximum likelihood estimate for the multinomial logistic
As indicated earlier, the parameter estimates of the MNL model provide only the direction of the
effect of the independent variables on the dependent variable: estimates do not represent actual magnitude
of change or probabilities. Thus, the marginal effects from the MNL, which measure the expected change
in probability of a particular category with respect to a unit change in an independent variable, are reported
and discussed. In all cases the estimated coefficients should be compared with the base category. Table 3.5
presents the marginal effects along with the levels of statistical significance.
Sex 0.0936 0.5414 0.8700 -0.0983 0.5628 0.2780 -0.0521 0.7858 *0.107 0.0567 0.0661 0.3910
Age 0.0193 0.0269 ***0.002 -0.0078 0.0287 0.5770 -0.0028 0.0337 0.3590 -0.0087 0.0039 **0.028
Familysize -0.2033 0.5680 ***0.002 -0.1213 0.5831 ***0.003 0.0595 0.6583 0.7050 0.2651 0.8019 ***0.001
Marital Status 0.0234 0.1449 0.3670 -0.0023 0.1742 0.2770 -0.0582 0.3982 ***0.000 0.0371 0.0221 *0.093
Educational level 0.0236 0.2357 0.2260 0.0124 0.2376 0.2450 0.0119 0.3608 0.1320 -0.0479 0.0343 0.1620
Percieved Beer Quality 0.1465 0.3086 *0.052 0.0540 0.3341 0.7670 0.0349 0.5028 0.2140 0.0577 0.0458 0.2080
Percieved Beer Price -0.0572 0.2839 0.3250 -0.0090 0.2930 0.6700 -0.0164 0.3897 0.4870 -0.0318 0.0413 0.4410
Percieved Social Benefit 0.2465 0.3055 ***0.000 0.1303 0.3152 0.7440 0.0048 0.3885 0.2800 0.1114 0.0442 **0.012
Advertisment Effect 0.0423 0.2228 0.2300 0.0339 0.2380 *0.102 0.0403 0.3352 *0.051 -0.0360 0.0337 0.2850
Situational Influence 0.0859 0.2781 *0.071 0.0803 0.3004 **0.011 0.0138 0.3831 **0.027 0.1000 0.0405 **0.014
Peer Influence 0.0404 0.2584 **0.025 0.1078 0.3089 ***0.006 -0.0475 0.3591 0.2390 -0.1007 0.0397 **0.011
Source: Own computation from own Survey
Note: ***1% significance level, **5% significance level, *10% significance level
The result indicated that respondents sex decreases the probability of preferring Walia and have no
statistically significant effect on St. George, Habesha and Others (Meta, Bedele, Harar & Dashen)
preference. Thus being female decrease the probability of preferring Walia by 0.05. The result is consistent
with Hartford et al., (1983) finding which involving demographics and drinking behaviors, males tend to
drink in larger quantities in same sex groups, whereas women drink with mixed crowds or with a male.
Moreover; an increase in age by one year significantly increases the probability of preferring St.
George 0.02 where as it decreases the probability of preferring Habesha by 0.03 and Others (Meta, Bedele,
Harar & Dashen) 0.01. According to Blackwell, Miniard & Engal (2006), understanding consumers‟ needs
in marketing analysis is related to age. For instance, Bennett, (2002) study group between ages 25 and 34
prefer to drink standard domestic beer drinkers whereas 35 to 44 year old choose light beer. Older people
drink more than younger people (Midanik et al., 1994).
The result also revealed that being married decrease the probability of preferring Walia. Whereas it
increase the probability of preferring Other beers (Meta, Bedele, Harar & Dashen); and have no
statistically significant effect on St. George and Habesha preference. Thus being married increases the
probability of preferring Others beer (Meta, Bedele, Harar & Dashen) by 0.04 and it decrease the
probability of preferring Walia by 0.06. On the other hand, a change in respondents family size by one
significantly decrease the probability of preferring St. George by 0.2, Habesha by 0.12 where as it
increases the probability of preferring Others (Meta, Bedele, Harar & Dashen) by 0.27.
Educational status of respondents increases the probability of preferring Walia at 15% of
significancy level; and have no statistically significant effect on St. George, Habesha and Others beer
(Meta, Bedele, Harar & Dashen) preference. Thus; educational status increases the probability of
preferring Walia beer by 0.012. This result shows that the educated segment of respondents preferred
Walia more than other. The finding is consistent with Michman, (2003) in which educational achievement
explained purchasing decisions and it was also closely associated with occupation and economic
circumstance. Moreover; Wells & Prensky (1996) claimed that education and occupation might affect the
consumer behavior process of evaluating and choosing of products. For instance, the working class will
choose products based on function and comfort rather than what is trendy, also most of them will not take
risks to try new products. Therefore; the more marketers understand the consumer demographics, the more
they can build the attitude to their brand in order to response the specific requirements of consumers.
Perceived beer quality increase the probability of preferring St. George only and a change in
perceived beer quality scale increases the probability of preferring St. George by 0.15. Perceived beer
4.2 Recommendations
Based on the findings discussed above the following recommendation forwarded:
The company can take a market segmentation strategy and design their products in a manner that
make the products appeal to different categories of individuals. The managers should appreciate the
influence of personal factors on customer satisfaction. In so doing, they should implement a product design
strategy that appeals to greater number of customers.
5. REFERENCE
Aaker, D. (1991& 1996) Managing Brand Equity (New York, NY: The Free Press).
Ali, M. J. (2014). Factors Influencing Marketing of Alcoholic Beverages In Kenya: A Case Study of East African
Breweries. European Journal of Business and Social Sciences. Vol. 3 (2), 122 – 153.
Allison, Ralph I., and Kenneth P. Uhl (1964), “Influence of Beer Brand Identification on Taste Perception,” Journal
of Marketing Research, 1(3), 33-39.
Assmus, Gert, John U. Farley, and Donald R. Lehmann (1984), “How Advertising Affects Sales: Meta- Analysis of
Econometric Results,” Journal of Marketing Research, 21(1), 65-74.
Ayanwale, A.B, Alimini, T and Ayalambe, M.A (2005). The influence of advertising on consumer brand preference.
Journal of Social Science 10(1), 9-16
Babor, T. F., and Mendelson, J. H., Greenberg, I., and Kuehnle, J. (1978), Experimental analysis of the „happy hour:
effects of purchase price on alcohol consumption‟, Psychopharmocology, Vol. 58, pp. 35-41.
Bagwell, Kyle (2007), “The Economic Analysis of Advertising,” Handbook of Industrial Organization, 3, 1701-
1844.
Bain, Joe S. (1950), “Workable Competition in Oligopoly: Theoretical Considerations and Some Empirical
Evidence,” American Economic Review, 40(2), 35-47.
Bandura, A. (1977). Social Learning Theory. New York: Prentice Hall.
Bearden, W. O. and Etzel, M. J. (1982), Reference group influence on product and brand purchase decisions,
Journal of Consumer Research, Vol. 9, No. 2, pp. 183-194.
Becker, Gary S., and Kevin M. Murphy (1988), “A Theory of Rational Addiction,” Journal of Political Economy,
96(4), 675-700.
Belk, R. (1974), An exploratory assessment of situational effects in buyer behavior, Journal of Marketing Research,
Vol. 11, pp. 156-163.
Berkman, Harold W., Jay D. Lindquist, and M. Joseph Sirgy (1997), Consumer Behavior: Concepts and Marketing
Strategy, NTC Business Books, Lincolnwood, IL.
Bovee, L.C, Thill, V.J, Dovel, G.P and Wood, M.B (1995) Advertising Excellence Boston:
Bruijin, B. (2011). Alcohol Marketing Practice In Africa: Findings From the Gambia, Ghana, Nigeria and Uganda.
Utretch, NE: World Health Organization
Casswell, S. (2004, November/December). Alcohol brands in young peoples' everyday lives: New developments in
marketing. Alcohol and Alcoholism, 39(6), 471-476. Retrieved September 15, 2007, from EBSCOhost
database.