Best Track 56641
Best Track 56641
Best Track 56641
Keywords
Consumer behavioural evolvement, behavioural segmentation, promotion, exploration and
exploitation, retailing, store scanner data.
Abstract
Consumers treat and react to promotions diversely in terms of their promotion proneness and variety
seeking tendency. This paper investigates 589 consumers with 169678 transaction records in the US
salt-snack market ranging from year 4 to year 7 in IRI Market Dataset for exploring how their
purchase behaviours evolve in the consumer life cycle. A set of algorithms is presented to process store
scanner data for measuring consumers promotion proneness and variety seeking tendency, which are
then used for conducting clustering analysis. Four types of purchase behaviours including
Promotion averse, Bargain hunters, Opportunistic explorers, and Opportunistic exploiters
are identified and assigned accordingly for each consumer in each year. Even though consumers
purchase behaviours will dynamically and freely shift within or among those four behavioural
segments over time, two clear behavioural evolvement patterns can be identified statistically in the
consumer life cycle. In the US salt-snack market, some consumers who are initially not sensitive to
promotions will gradually evolve to take advantage of promotion to try alternatives for extending
their market knowledge. Some explorers promotion proneness will gradually increase and outweigh
their variety seeking tendency over time, thus, those opportunistic explorers become inclined to buy
any brands on promotion. In contrast, consumers who initially prefer to buy their familiar brands will
become motivated to buy any brands, particularly the brands they are familiar with, on promotion.
After a certain period of exploitation with promotions, their variety seeking tendency will gradually
increase and slightly outweigh their promotion proneness, thus, those consumers will particularly
keen to take advantage of promotion to explore the US salt-snack market over time. We discuss the
implications of these findings in terms of managerial recommendations regarding promotion activities
for retailers to increase the rate of response to promotions and offer suggestions for future research.
1. Introduction
According to solid research in marketing, consumers made more than 70% of their
purchase decisions in store(Babej and Pollak, 2007).Messages delivered at point-of-purchase
thus are expected to have the best chance to affect consumers purchase behavioursand
motivate consumers to behave as retailers desire(Babej and Pollak, 2007). In order to achieve
marketing objectives through influencing consumer behaviours, retailers select and combine
types of promotions includingin-store advertisement, value increasing sales promotions,
and value adding sales promotions. Even though in-store promotions were found to have
positive effects on sales performances, the issue of low promotion response rates is always a
serious problem confronted by retailers (Gilbert and Fackaria, 2002). Consumers have
different views and reactions toward promotions. The promotion proneness, which is
defined as thewillingness a consumer would like to accept promotions, is essential to be
considered when targeting consumers for providing sales promotions.
International Trade & Academic Research Conference (ITARC), 3-4th November 2014 UK 43
The Business & Management Review, Volume 5 Number 3 November 2014
International Trade & Academic Research Conference (ITARC), 3-4th November 2014 UK 44
The Business & Management Review, Volume 5 Number 3 November 2014
the increase of market knowledge, which thus can be quantified by using the value of
information from purchases. In other words, if consumers want to learn then their
motivation to seek variety follows an inverted U-shape. The motivation of obtaining
information from exploration varies with consumers knowledge in a product market due in
part to consumers ability to distinguish between brands (Heilmanet al., 2000).
Inexperienced consumers with limited knowledge about a product market have little
incentive to obtain market information through exploring because they cannot really
distinguish between brands in the product market (Heilmanet al., 2000). Consumers
motivation for exploration will gradually increase since they will sample brands to pursuit
and acquire knowledge about alternatives and gain experiences in the product market
(Erdem and Keane, 1996; Heilmanet al., 2000). When consumers first enter a product market,
they inclined to buy big and famous brands that they are familiar with, in the product
market (Heilmanet al., 2000).As experiences are gained from trying and repeatedly purchase
big brands over a period of time, consumerswill expand their exploration range and try
small and unfamous brands in the product market (Heilmanet al., 2000). Market knowledge
acquired from brands sampling enables consumers to be able to differentiate brands in the
product market (Heilmanet al., 2000). Until consumers gained sufficient experiences and
knowledge in a product category from their exploration and exploitation activities, they will
become loyal to their preferred brands (Lodorfoset al., 2006; Heilmanet al., 2000). The
motivation for exploring will then gradually decrease to 0 as consumers are relatively
certain about brands in the product market and extra information is no longer perceived to
be as valuable as previously (Heilmanet al., 2000).
In this research, we adapted Chens (2004) measurement of information value in
financial market to apply in a frequently purchased consumer goods market for measuring
consumers value of information from purchases. Following the literature review, we
describe an approach for measuring promotion proneness and variety seeking tendency and
an approach for segmenting consumers in US salt-snack market based on their purchase
behaviours in the method section.
3. Method
3.1 Dataset for analysis
The IRI marketing dataset was used for analysis in this study. The panel dataset and store
dataset in salt-snackmarket ranging from year 4-7 were combined for processing. For the
measurement of promotion proneness and variety seeking tendency, the sufficient
transaction records are required for each consumer. In this study, 595 consumers, who had
at least 34 purchase records in each of year 4-7, were selected. While, among the selected 595
consumers, 6 consumers were rejected sincemore than 10% of their salt-snack purchases are
associated with missing brand promotion information. In the final dataset, 589 consumers
associated with 169678 purchase records were selected for further analysis. The number of
selected consumers accounts for around 19% of consumers who have purchase records in
the consecutivelyfour years. Consumers variety seeking reactions to promotions were
measured by their expected utility for taking advantage of promotion to explore in a
product market in terms of promotion proneness and variety seeking tendency. The
following sub-section explains the measurements of promotion proneness and variety
seeking tendency.
3.2 Behavioural measurements
3.2.1 Promotion proneness prevalence of promotion
International Trade & Academic Research Conference (ITARC), 3-4th November 2014 UK 45
The Business & Management Review, Volume 5 Number 3 November 2014
In IRI dataset, the information of advertising, point of display, and price reduction
are available for each product in each retail store. In this study, the data about those
promotion approaches in salt-snack market from year 4-7 were used for measuring
consumers promotion proneness. In each retail store, at least one brand is promoted with at
least a promotion approach every week. Therefore, consumers have opportunities to buy
promoted salt-snacks in all their purchases. This means that when we calculate the
prevalence of promotion, which is defined as the number of purchases of promoted
products relative to the total number of purchases, we are also measuring how prone
consumers are to make use of promotions. The prevalence of promotion is calculated by
using formula 1 as:
International Trade & Academic Research Conference (ITARC), 3-4th November 2014 UK 46
The Business & Management Review, Volume 5 Number 3 November 2014
year 4 to their most current purchase are processed. The knowledge about a product market
was measured by using formula 2:
Knowledge about a product market I M p
The number of brands tried by a consumer in his / her purchase life cycle n
The total number of brands available in the product market in the consumer s purchase life cycle N
(2)
Formula 2: Knowledge about a product market
According to Chen (2004), the unit value of information of trying brands in the product
market is log 2 I M p . The formula for calculating the associated unit value of
information is:
n
Unit value of information log 2 I M p log 2 (3)
N
Formula 3: Unit value of information
When I M p =1, log 2 I M p = 0. The unit value of information for trying alternatives in
a product market is zero when consumers already have full knowledge about the product
market. When consumers certain about a product market, the extra information about the
product market have no value for them and they will not be expected to do exploration
activities to further extending their market knowledge. I M p Varies between 0 and 1.
Theoretically, the more the I M p approaches to 1, the less valuable the information is.
When I M p approaches to zero, log 2 I M p approaches infinity. In that case, the unit
value of information obtained from trying alternatives is very high and consumers are
expected to do information search by trying alternatives. However, in practice, since new
consumers are not able to differentiate brands in the product market due to the lack of
market knowledge, they are normally insensitive to the brands in the market and will not try
alternatives in the product market. Consumers variety seeking for learning purpose is
determined by their knowledge about a product market (Heilmanet al., 2000). When
consumers do not have any knowledge about the product market, their value of information
will be zero even though their unit value of information will approach to infinity. With the
increase of market knowledge, the capability of differentiation among brands increases,
thus, the value of information will increase accordingly. With sufficient knowledge was
obtained, extra information about the product market will no longer be as valuable as
before, thus, the value of information will gradually decrease until approach to zero when
all brands in the product market are tried by a consumer. The consumers with none value of
information are expected to stick with their preferred brands in their purchases later on. In
general, the value of information from information search for learning purposes in
purchases presents an inverted U-shape with the increase of product market knowledge,
which is consistent with and can be explained by the motivation of variety seeking in
dynamic choice process proposed by Heilmanet al. (2000).
In this study, we measure the value of information from exploration activities by
using the quantified consumers knowledge about a product market to multiply the
International Trade & Academic Research Conference (ITARC), 3-4th November 2014 UK 47
The Business & Management Review, Volume 5 Number 3 November 2014
associated unit value of information. Formula 4 shows the calculation of the value of
information from purchases:
Value of information from purchases
Knowledge about a product market The unit value of information
n n
I M p log 2 I M p log 2 (4)
N N
Formula 4: Value of information from purchases
International Trade & Academic Research Conference (ITARC), 3-4th November 2014 UK 48
The Business & Management Review, Volume 5 Number 3 November 2014
differences in their promotion proneness and variety seeking tendency. Consumers in the
Red segment are not prone to make use of promotions in US salt-snack market since they
have the lowest number of purchases of promoted products relative to the total number of
their purchases. In other words, with the same level of variety seeking tendency, those
consumers have the lowest probability to respond to promotions since the provision of
promotion may result in negative brand image of the promoted brands/products for them.
They basically do not want to buy promoted salt-snacks (particularly true for the brands
they are not familiar with). We named the consumers in this segment as Promotion averse
on the basis of their behavioural characteristics.
On the contrary, consumers in the Magenta segment have the highest value in
promotion prevalence and are prone to take advantage of promotions to maximize their
material benefits in purchases. With the same level of variety seeking tendency, those
consumers have the highest probability to buy the promoted salt-snacks in US market. We
named the consumers in this segment as Bargain hunters as they tend to go shopping for
bargains.
Consumers in Blue and Green segments have medium level of promotion
proneness. They are differentiated from promotion averse and bargain hunters in promotion
proneness. For distinguishing the consumers in Blue segment from consumers in Green
segment, consumers variety seeking tendency plays a critical role even though consumers
in Green segment are slightly more sensitive to promotions than consumers in Blue
segment. Consumers in Blue segment have higher variety seeking tendency than the
consumers in Green segment. We named the consumers in Blue segment as
Opportunistic explorers and consumers in Green segment as Opportunistic exploiters.
Opportunistic explorers are prone to take advantage of promotion to explore in the product
market for extending their market knowledge. With the same level of promotion proneness,
opportunistic explorers have the highest value of information from purchases and thus are
highly motivated to try alternatives for information search in the salt-snack market.
On the contrary, opportunistic exploiters have low value of information and are
prone to take advantage of promotion to repeatedly buy their preferred brands. With the
same level of promotion proneness, opportunistic exploiters have the lowest value of
information from purchases and thus are not motivated to explore in the salt-snack market.
International Trade & Academic Research Conference (ITARC), 3-4th November 2014 UK 49
The Business & Management Review, Volume 5 Number 3 November 2014
International Trade & Academic Research Conference (ITARC), 3-4th November 2014 UK 50
The Business & Management Review, Volume 5 Number 3 November 2014
In general, the number of promotion averse decreases since many of them evolved to
taking advantage of promotion to extend their market knowledge after they first enter the
US salt-snack market. The behavioural evolvement dynamics suggest that consumers are
expected to become more and more sensitive to promotions over time in their purchase life
cycles.Similarly, the number of opportunistic exploiters also decreases since many of them
either directly transferred tobe opportunistic explorers or indirectly evolved to be
opportunistic explorers through bargain hunters. As the time evolves, the number of
opportunistic explorers and the number of bargain hunters increase. Consumers
dynamically transferred between those two behavioural segments by making the trade-off
between promotion proneness and variety seeking tendency.
In order to find out how and why consumers behaviours evolve from one segment
to the other segment over time, 9 consumers, whose behavioural evolvement processes from
year 4 to year 5 are presented in Figure 3, are selected. In Figure 3, each line represents a
typical behavioural evolvement type, which is presented by one of nine selected consumers.
When consumers first enter a salt-snack market, some of them tend to purchase big
brands in the salt-snack market regardless of the promotions(Heilmanet al., 2000).The
increase of promotion sensitivities and variety seeking tendency, which are resulted from
trying big brands to extending market knowledge, make some of promotion averse became
opportunistic explorers over time. Those consumers gradually inclined to take advantage of
promotion to explore among the salt-snack market to extend their market knowledge in
International Trade & Academic Research Conference (ITARC), 3-4th November 2014 UK 51
The Business & Management Review, Volume 5 Number 3 November 2014
order to reduce risks from trying small brands. Since promotions of interesting brands are
not always available, consumers high variety seeking tendency motivate them to try
alternatives even without any promotions. However, in the evolvement process, consumers
variety seeking tendency is not necessarily to continuously increase. Some consumers, who
originally have high variety seeking tendency and evolve from promotion averse to
opportunistic explorers, took advantage of promotion to repeatedly buy their preferred
brands for a certain period when their expected costs from trying alternatives exceed their
expected benefits obtained from information search(Heilmanet al., 2000). Since those
consumers still have high variety seeking tendency, they inclined to re-take advantage of
promotion to try alternatives when their expected benefits from exploration exceed their
expected costs.
Consumers differ from their exploration needs and learning rates. After a certain
period of exploration, some consumers are able to differentiate brands in the product market
and know which purchase decisions can most satisfy their needs (Heilman et al., 2000).
Around 31% of opportunistic explorers thus became more and more sensitive to promotions
and prefer to buy any salt-snacks on promotion. Promotion proneness replaces variety
seeking tendency as the critical criterion in determining their segmentation. In other words,
they became bargain hunters.
The second typical behavioural evolvement dynamic starts from the opportunistic
exploiters. Consumers are regarded as opportunistic exploiters if they have low variety
seeking tendency for learning purpose and incline to take advantage of promotion to
repeatedly buy a subset of brands. When consumers first enter a product market, they
cannot differentiate brands in the product market. Some of them inclined to take advantage
of promotion to repeatedly buy big brands in the salt-snack market(Heilmanet al., 2000).On
the contrary, consumers, who have sufficient market knowledge and a set of preferred
brands in a product market, are also regarded as opportunistic exploiters if they repeatedly
buy and loyal to their preferred brands. However, in reactive environment, true brand loyal
consumers are supposed to be very limited or even non-existent in salt-snack market. Even
though 19% of consumers are consistently to be regarded as opportunistic exploiters in both
year 4 and year 5, they also tried alternatives with or without promotions during the two
consecutive years. The increase of variety seeking tendency resulted from occasionally
trying new brands in the salt-snack market motivated opportunistic exploiters to further
extend their market knowledge from exploration. They thus evolved to be opportunistic
explorers directly or even indirectly through bargain hunters over time depending on their
expected benefits and costs from exploration.
When the expected benefits for trying alternatives sufficiently exceed the expected costs,
opportunistic exploiters inclined to explore in the salt-snack market even without taking
advantage of promotion. Their variety seeking tendency rapidly increased, which makes
them to directly evolve to be opportunistic explorers over time. While, when the expected
costs from exploration exceed the expected benefits, opportunistic exploiters inclined to
repeatedly buy a subset of their preferred brands on promotion. Only when the alternatives
are on promotion, which reduces the risks from trying them, opportunistic exploiters are
motivated to try the promoted alternatives since the expected costs are reduced by the
provision of promotions. In other words, those opportunistic exploiters became bargain
hunters. The variety seeking tendency of bargain hunters increases with the increase of
market knowledge obtained from trying promoted alternatives. When variety seeking
outweighs the promotion proneness in purchase decision making, bargain hunters inclined
International Trade & Academic Research Conference (ITARC), 3-4th November 2014 UK 52
The Business & Management Review, Volume 5 Number 3 November 2014
to try alternatives even without promotions when expected benefits of exploration exceed
the expected costs. Thus, they became opportunistic explorers over time. While, if the
expected costs of trying promoted alternatives exceed the expected benefits, bargain hunters
inclined to repeatedly buy a subset of their preferred brands even without promotions for a
certain period, which results in a decreased promotion proneness and variety seeking
tendency if new brands are introduced to the product market. As for some of those bargain
hunters with high variety seeking tendency, in this circumstances, they also became
opportunistic explorers when their promotion proneness and variety seeking tendency are
slightly decreased. Their exploration activities are resumed when their expected exploration
benefits exceed their expected exploration costs.
5. Conclusions and recommendations
This study segments consumers into four segments as Promotion averse,
Opportunistic explorers, Opportunistic exploiters, and Bargain hunters in terms of
their promotion proneness and variety seeking tendency. In order to increase response rate
of promotions, providing the tailored promotions for consumers in each segment based on
their purchase behaviours is necessary. Among those four behavioural segments, promotion
averse is not the target of retailers for providing promotions. On the contrary, any
promotions are suggested to be provided for the bargain hunters since they incline to take
advantage of promotions to optimize their monetary value in purchases. Bargain hunters are
the main target for retailers to satisfy marketing needs by using marketing interventions.
Opportunistic explorers incline to extend their market knowledge from trying alternatives
on promotion. Any promotions about the new brands/products are thus suggested to be
provided to opportunistic explorers. The promotions of new brands/ products are expected
to be able to motivate them to explore among brands with less expected costs. However,
since some opportunistic explorers may repeatedly purchase their favourite brands for
ascertain period, the promotions of their preferred brands are also suggested to be provided.
Similarly, the promotions of opportunistic exploiters preferred brands are suggested to be
provided to them since they prefer to stick with their current best choices on promotion.
Those promotions are expected to motivate them to buy their preferred brands/products
with less cost and thus will increase the sales of the brands/products.
Even though the behavioural evolvement dynamics are identified and the general
probabilities of behavioural evolvement from one segment to the other are suggested in this
study, we still cannot precisely predict consumers purchase behaviours in the future. In
order to overcome this limitation, a predictive modelling will be developed for predicting
consumers purchase behaviours by using their associated promotion proneness and variety
seeking tendency. The behavioural evolvement dynamics and the predictive modeling in
other product markets will also be explored to find out how generalize the identified
behavioural evolvement dynamics and the developed predictive modeling in salt-snack
market are in future research.
6. References
Audibert, J. Y., Munos, R., and Szepesvari, C. (2009) Exploration-exploitation trade-off
using variance estimates in multi-armed bandits. Theoretical Computer Science. 410. p.
1876-1902.
Babej, M. E., and Pollak, T. (2007) In-store advertising: Coming of age? [Online] Available from :
http://www.forbes.com/2007/10/29/unsolicited-advice-supermarkets-
oped_meb_1030advice.html [Accessed: 10 June 2014]
International Trade & Academic Research Conference (ITARC), 3-4th November 2014 UK 53
The Business & Management Review, Volume 5 Number 3 November 2014
Cerrito, P. B. (2005). Comparison of Enterprise Miner and SAS/Stat for data mining. Louisville:
SAS Institute Inc.
Che, H., Erdem, T., and nc , T. (2013) Consumer learning and evolution of consumer brand
preferences. Pennsylvania: University of Pennsylvania.
Chen, J. (2004) Generalized entropy theory of information and market patterns. Corporate
Finance Review. p. 23-32.
Collica, R. S. (2011). Customer segmentation and clustering using SAS Enterprise Miner. 2nd Ed.
Cary: SAS Institute Inc.
Erdem, T., and Keane, M. P. (1996) Decision-making under uncertainty: Capturing dynamic
brand choice processes in turbulent consumer goods markets. Marketing Science. 15(1).
p. 1-20.
Gupta, A. K., Smith, K. G., and Shalley, C. E. (2006) The interplay between exploration and
exploitation. Academy of Management Journal. 49(4). p. 693706.
Gilbert, D. C., and Jackaria, N. (2002) The efficacy of sales promotions in UK supermarkets:
A consumer view. International Journal of Retail & Distribution Management. 30(6). p.
315-322.
Heilman, C. M., Bowman, D., and Wright, G. P. (2000) The evolution of brand preferences
and choice behaviors of consumers new to a market. Journal of Marketing Research.
37(2). p. 139-155.
Lodorfos, G. N., Mulvana, K. L., and Temperley, J. (2006) Consumer behaviour: Experience,
price, trust and subjective norms in the OTC pharmaceutical market. Innovative
Marketing. 2(3). p. 41-66.
SAS Institute Inc. (1999) SAS/STAT user's guide. The FASTCLUS Procedure. Cary: SAS
Institute Inc.
Teunter, L. H. (2002) Analysis of sales promotion effects on household purchase behavior. A Thesis
Submitted in partial fulfilment of the Requirements of Erasmus University for the
Degree of Doctor of Philosophy. Rotterdam: Erasmus University.
International Trade & Academic Research Conference (ITARC), 3-4th November 2014 UK 54