An Engel–Kollat–Blackwell model application on restaurant clientele purchase decision-making processes in commercial eateries in Kakamega County, Kenya

ABSTRACT Restaurant clientele have a task of making critical decisions concerning their choices of food and drinks. The decision-making process involves, firstly, the selection of the restaurant in which to have a meal and, secondly, the selection of food items from the menu to constitute their meals. Thus, it requires cognitive and physical effort for restaurant clientele to make decisions concerning their restaurant choices and ultimately a final purchase decision. This study applied the Engel–Kollat–Blackwell-consumer purchase decision model to investigate restaurant clientele’s purchase decision-making process among commercial eateries in Kakamega County in Kenya. For the purpose of this study, a five-stage customised EKB model was applied. The study employed a descriptive research survey design, while Fisher’s formula was utilised, giving a sample size of 384 respondents. The study results show a high positive correlation (R = 0.892) between the study variables. The study results indicate that 79.6% of the total variation in consumer ratings was because of the restaurant customer purchase decision-making process, in addition to the overall regression model predicting the outcome variable (p < 0.05). In relation to the contribution of studied predictor variables, a manual search gave the greatest contribution (B = 0.545, p < 0.001, t = 9.105), while sense of belonging/ class of the eatery gave the least contribution (B = −0.584, p < 0.001, t = −12.505). To yield positive results, eatery managers will be required to adjust these predictor variables depending on the contribution of each on the outcome variables, given the predictor variables to be tested account for the most variance in the outcome variable.


Introduction
The purchase decision-making process encompasses making a selection from a series of choices presented and available to the prospective customer.Restaurant clientele, just like any other consumer, have a series of choices in the process of making critical decisions concerning their choice and consumption of services and products.The decision-making process therefore involves, first, the selection of the restaurant in which to have a meal and, second, the selection of food items from the menu to constitute the meal.According to Wilkie (1994), it requires a great deal of cognitive and physical effort from restaurant clientele in making decisions concerning their restaurant choice, and ultimately a final purchase decision of food and drinks.It therefore implies that the customer has to compare all the information about each of the parameters that constitutes the yardstick upon which their ultimate selection is based.Thus, the customer goes through a series of steps which involve decisionmaking before the actual buying behaviour witnessed during menu item selection and consumption (Longart et al., 2010).
Restaurant clientele's purchasing decision-making is therefore not an instant process, but a lengthy, complex, physiological and psychological process that begins prior to patronising any restaurant outlet.It first starts with the arousal of the desire to eat, the gathering of relevant information on the availability of outlets and food choices from which to make the best selection, making a choice based on the available information, purchasing, consumption and lastly post-consumption.Thus, the "purchasing" process begins well in advance, with a prepurchase stage and comprising a sequence of challenging tasks (Chen & Dubinsky, 2003).This may be explained by the fragmented nature of the restaurant industry not only in Kenya, but in many countries of the world, comprising a myriad of establishments.Attempts to explain the process of consumer purchasing were made by researchers who yielded the Engel-Kollat-Blackwell (EKB) model (Blackwell et al., 2006).
According to Lillicrap and Cousins (2014), there are two main reasons why consumers eat out.The first is convenience: for example, a meal or drink during work breaks.People may also eat out for leisure, which is a more deliberate and thoughtful decision-making process as compared to meal breaks in busy schedules.Studies have shown that eating out is a high-involvement purchasing decision-making process, and that restaurants need highly sensitive employees to be able to ignite ABSTRACT: Restaurant clientele have a task of making critical decisions concerning their choices of food and drinks.The decision-making process involves, firstly, the selection of the restaurant in which to have a meal and, secondly, the selection of food items from the menu to constitute their meals.Thus, it requires cognitive and physical effort for restaurant clientele to make decisions concerning their restaurant choices and ultimately a final purchase decision.This study applied the Engel-Kollat-Blackwell-consumer purchase decision model to investigate restaurant clientele's purchase decision-making process among commercial eateries in Kakamega County in Kenya.For the purpose of this study, a five-stage customised EKB model was applied.The study employed a descriptive research survey design, while Fisher's formula was utilised, giving a sample size of 384 respondents.The study results show a high positive correlation (R = 0.892) between the study variables.The study results indicate that 79.6% of the total variation in consumer ratings was because of the restaurant customer purchase decision-making process, in addition to the overall regression model predicting the outcome variable (p < 0.05).In relation to the contribution of studied predictor variables, a manual search gave the greatest contribution (B = 0.545, p < 0.001, t = 9.105), while sense of belonging/ class of the eatery gave the least contribution (B = −0.584,p < 0.001, t = −12.505).To yield positive results, eatery managers will be required to adjust these predictor variables depending on the contribution of each on the outcome variables, given the predictor variables to be tested account for the most variance in the outcome variable.KEYWORDS: alternatives, choice, consumer, food and drinks, menu, purchase the desire not only to choose the restaurant, but also the food items from the menu as well as positive post-purchase behaviour (Sparks et al., 1997).Thus, restaurant clientele's involvement in the purchase decision-making process is considered a significant element with considerable influence on the ultimate purchase decisions of food and drinks in a restaurant set-up.
Studies on this topic can be envisaged from a cognitive perspective in which the restaurant clientele shows mainly clear behaviour.In the general cognitive model, restaurant clientele research on the purchasing decision-making process can be organised into three areas: information processing, consumer culture theory and behavioural decision research, but they are intertwined to make a whole (Bartels & Johnson, 2015).Nonetheless, there are two categories of restaurant consumers: high-involvement consumers, who tend to actively gather and disseminate information on food, drinks and services that they are interested in; and low-involvement consumers, who seldom contribute to the dissemination of information on either food, drinks or services (Hong, 2015).Even so, both categories of customers will eventually make choices from the alternatives available.
Generally, consumers' ultimate decisions to purchase is thought of as being a process that involves details of where they buy, how they buy, what they buy and how much they buy (Kotler et al., 1998).Thus before making the final decision, the consumer goes through a series of the available options from which an ultimate choice is eventually made.This also applies to restaurant consumers who must make a decision about where to dine, how to make the final choice from a list of food and drink offers and how much to buy, thus leading to a number of interrelated factors, which this study sought to investigate, since establishing the whys of the restaurant consumer decisionmaking process is not an easy task (Kotler et al., 1998).

literature survey
According to Kotler et. al. (1998), a restaurant consumer buying process begins long before the purchase and continues long after the purchase and consumption of goods and services.Thus, to gather sufficient information, it is important for hospitality professionals to understand the entire purchasing process, which is a complex rather than a simplistic purchasing decisionmaking process (Kotler et. al., 1998).Thus for purposes of this study, the consumer process model was applied in an attempt to understand the restaurant clientele purchasing decision-making process in commercial eateries in Kakamega town of Kakamega County in Kenya.
According to Blackwell and colleagues (2006), the EKB model embodies "a road map" of consumers that marketers and managers can use to help guide the product and/or service mix, communication and sales strategies, and the same was applied in this study.The applied EKB model traces the psychological state of discrete restaurant clients from the point at which they become aware of the possibility of satisfying a hungerrelated need by purchasing and consuming food or drink, up to their post-purchase evaluation stage.Nonetheless, a number of factors, which are related to meal experience, control this process (Lillicrap & Cousins, 2014).
In view of current models of the consumer decision-making process, the process comprises of a series of undertakings with some limitations in terms of the theoretical background (Erasmus et al., 2001;Bartels & Johnson, 2015).From the perspective of Pham and Higgins (2005), the EKB model is critiqued for its limitations in addressing the motivational aspect of the decisionmaking process, a fundamental limitation of the original model.Nevertheless, despite criticisms and limitations, the EKB model has indeed yielded fruits in the provision of a useful guide about consumers, enabling sales people and executives to provide guidance on the product and/or service mix, communiqué and rummage sale stratagems (Blackwell et al., 2006).According to Blackwell et al. (2006), the consumer decision-making process is a seven-step sequential flow, beginning with need recognition and ending with divestment as shown in a flow diagram (Figure 1).
To describe the interaction of the model, the consumers' needs cannot be adequately addressed until they are properly outlined, thus the need-recognition stage.This is a crucial stage associated with aspects affecting either the desired and/or the actual state of an individual potential restaurant client.Second is the information-search stage, which presents several aspects for consideration (Hoyer & MacInnis, 2003).These may include a directional-type search (internal or external), an informationtype search and a structure-based search (alternative-based versus attribute-based) (Hauser & Wernerfelt, 1990).The third stage comprises the evaluation of alternatives, which revolves around narrowing down to the available range of alternatives.The fourth stage encompasses the actual purchase practice based on the evaluation of alternatives in stage three, and thus involves an examination of attributes to make the final purchase decision.Once the purchase has been executed, then the fifth  2006) stage (consumption) starts either on site or off the premises.The sixth stage involves reviewing the decision for purchase and consumption rules, options taken, attitudes and preferences (post-consumption evaluation).Lastly, the seventh stage is about a post-choice and consumption assessment.
Even though there are a small number of studies on the topic, they are limited in the application of the EKB model for explaining the restaurant clientele purchase decision-making process as many of the studies have revolved around the consumer purchase decision-making process in a generalised perspective.Thus, this study specifically focused on the restaurant clientele purchase decision-making process, supported by the applicable principles of the model.Moreover, numerous studies have contradicted the EKB model, including Olshavsky and Granbois (1979), who postulate that in many instances, the majority of the purchases of consumer goods are quick and so elaborate decision-making processes may seldom occur.Rickwood and White (2009) and Pham and Higgins (2005) dispute that the application of the EKB model is more appropriate in the purchasing processes of goods than services.On the other hand, Erasmus, Boshoff and Rousseau (2001) and Bruner and Pomazal (1988) maintained that the purchase process is never linear as exhibited in the model, because of a multiplicity of factors and undertakings.Nonetheless, although it is established that linear processes rarely occur in reality, it is also important to simplify the process to provide a guide to how it occurs.These inadequacies in research constituted the reasons for initiating this study for purposes of attempting to fill the gaps.

Materials and methods
This study sought to apply the EKB model to investigate the restaurant clientele purchase decision-making process among commercial eateries in Kakamega County in Kenya.For the purposes of this study, a five-stage customised EKB model was applied.Needs recognition (customers' initiative to visit the eatery) formed the first stage of the model, followed by information search (customer source of information about the eatery).Moreover, an evaluation of alternatives (customer basis for evaluating the menu and/or alternatives) formed the third stage, followed by purchase (customer basis of final choice of food/drink from the menu) and finally post-purchase behaviour (customer post-purchase behaviour) (Hauser & Wernerfelt, 1990;Blackwell et al., 2006).
These constructs were further operationalised to yield hunger and/or thirst, sense of belonging/class of the eatery, norm to patronise the eatery and food specialty prepared in the eatery (customers' initiative to visit the eatery).On the other hand, social networks (e.g.Facebook, Twitter, etc.), media sharing networks (e.g.Instagram, etc.), discussion forums (e.g.WhatsApp, Telegram, etc.), bookmarking and content curation sites (e.g.Pinterest, Reddit, Diigo, Delicious, etc.), consumer review networks (e.g.TripAdvisor, etc.), colleagues, friends and family members, and a physical search around town constituted the customer sources of information about the eatery.Customer basis for evaluating the menu and/or alternatives yielded food/ drink appearance (colour), aroma (smell), taste (sweetness, sourness and pepper heat), texture (elasticity, hardness, consistency) and temperature at service (hot, room temperature, cold).Nonetheless, customer basis of the final choice of food/drink from the menu yielded service level/speed, price charged, promotions (photographs of the food on the menu, food presentation, terms used to describe the food, etc.), state of the restaurant (hygiene, cleanliness, beauty) and the human factor (waiters, other guests).Lastly, customer post-purchase behaviour yielded ideas of complimented verbally, tipped, shared the experience on social media and planned to visit again.
The study employed a descriptive research survey design.According to Whaley et al. (2014), this design is concerned with describing situations based on respondents' responses and examining their relationships.Consequently, academic disciplines, especially the social sciences and psychology use the descriptive research survey design method to obtain a general overview of the subject, and therefore this fits in its application to this study (Gall et al., 2007).This design was chosen based on its ability and appropriateness to yield accurate information on the purchase decision-making process (Kothari, 2010).
For purposes of this study, Kakamega County formed the study area, while commercial eateries in Kakamega town formed the study population.To produce accurate results, Fisher's formula (1935) was applied, giving a sample size of 384 respondents.Moreover, a systematic simple random sampling was applied in the selection of the commercial eateries (30 out of 53 outlets), as well as in the selection of the study respondents.Structured questionnaires were self-administered, and thereafter the collected data was entered into spreadsheets, cleaned and analysed.The study yielded the results presented below.

Results
This study investigated restaurant clientele's purchase decisionmaking process in commercial eateries in Kakamega County in Kenya.Out of the 384 self-administered questionnaires, only 333 were returned and the study recorded a response rate of 87%.Prior to final analysis, a diagnostic test was carried out based on collinearity statistics (Table 1).
The variables under test gave variance inflation factors (vIF) and tolerance values, which were used to determine multi-collinearity.Thus, customer initiative to visit the eatery gave 0.919, 1.089, customer source of information about the eatery was 0.830, 1.204, customer basis of evaluating the menu and alternatives was at 0.837, 1.195, customer basis of final choice of food/drink from the menu had 0.869, 1.150, and customer post-purchase behaviour yielded 0.754, 1.325.In order to describe the pattern of variables under investigation, descriptive statistics were run and produced results showing that the majority (51.4%) were female respondents compared to 48.6% males.
Further, the study gave results showing that the majority (43.2%) of respondents had tertiary education, 40.5% had finished high school and 16.2% had a primary school level of education.According to the study results, a majority of the respondents attained higher levels in education as compared to the few (16.2%) who only managed to attain a basic primary level of education Table 2. Furthermore, the study results on the type of visit gave a majority of respondents (41.4%) indicating that they revisit, 33.03% were first timers, while the smallest in this category (25.53%) comprised regular customers.Nonetheless, to find out the relationship between the restaurant customers purchase decision-making process and the resultant rating of the variables under investigation, a multiple regression was carried out and the output tabulated (Tables 3, 4 & 5).
From the model summary, the study gave an R-value of 0.892, R-squared value of 0.796 and adjusted R-squared of 0.780 respectively.Moreover, the study yielded the ANOvA results (Table 4).
The ANOvA table of results gave the significance (p) value of <0.001 (i.e., p < 0.05. Finally, to establish the contribution of each predictor variable, the study gave the multiple regression coefficients (Table 5).

Discussion
The study used 384 self-administered questionnaires.However, only 333 questionnaires were appropriately filled in and returned.The study thus achieved a response rate of 86.7%.According to Jackson (2009), Kothari (2010), O'Leary (2014) and Stehlik-Barry and Babivec (2017), a response rate of over 60% is acceptable for a survey study, and therefore validates the response rate that was achieved here.
To determine multi-collinearity, a diagnostic test was carried out which gave variance inflation factors (vIF) and tolerance values.The study results show vIF values less than 10 and tolerance values greater than 0.100 and less than 1.00 for all the study constructs.Since vIF values were ≥1 and ≤10, this implies that there was no multi-collinearity (Jackson, 2009;Kothari, 2010;O'Leary, 2014;Stehlik-Barry & Babivec, 2017).Moreover, since the tolerance values are greater than 0.100 and less than 1.00, it implies that there was also no multi-collinearity (Jackson, 2009;Pallant, 2010;Kothari, 2010;O'Leary, 2014;Stehlik-Barry & Babivec, 2017).
The study further sought to investigate respondents' demographic factors including gender, education level and respondent visits.Thus, the gender of respondents yielded results showing that the majority (51.4%) of respondents were female, suggesting that the sampled eateries were patronised by more female than male patrons.Investigations on respondents' education levels on the other hand gave results showing that a majority (43.2%) of clientele patronising the sampled eateries attained higher education levels as compared to the few (16.2%) who only managed to attain a basic primary level of education.Finally, the study yielded results showing that a majority of respondents were loyal clientele, hence maintaining over 74.4% of loyal customers is a clear indication of a high satisfaction level of patrons in the sampled eateries.A high level of service quality is the only sure way to attract repeat customers in not only hospitality organisations, but also all types of organisations globally (Uedufy, 2023).
The study further sought to investigate the relationship between the restaurant purchase decision-making process and the ratings thereof.The EKB model was adopted and customised for the study, which yielded five constructs that were further operationalised to give 25 predictor variables.Multiple regression was applied and gave a high positive correlation (R = 0.892) between the study variables.Conversely, the results depict that 79.6% of the total variation in consumer ratings was as a result of the restaurant customer purchase decisionmaking process, while the 20.4% variation may be as a result of other factors outside the scope of this study.Nonetheless, the study results indicate that the overall regression model predicts the outcome variable since the study-attained results with the significance (p) value of 0.00, thus p < 0.05.
For purposes of establishing the contribution of each of the 25 predictor variables, the study results show that a physical search around the town had the greatest contribution (B = 0.545, p < 0.001, t = 9.105).This is followed by food/drink texture/elasticity/hardness/consistency (B = 0.291, p < 0.001, t = 8.319) and third, hunger and/or thirst (B = 0.230, p < 0.001, t = 6.249) respectively.On the other hand though, a sense of belonging/class of the eatery gave the least contribution (B = −0.584,p < 0.001, t = −12.505),followed by price charged (B = −0.487,p < 0.001, t = −7.630),and third, colleagues, friends and family members (B = −0.190,p < 0.001, t = −7.297)respectively.Whereas 12 predictor variables in this study gave a positive contribution, a majority of the predictor variables (13) gave negative contributions to the relationship between the variables in the proposed restaurant customer purchase decision-making model.

Conclusion
This study sought to investigate the suitability of the application of the EKB model in the restaurant clientele purchase decisionmaking process.The adopted model constituted five constructs that yielded 25 predictor variables.Thus, the study applied multiple regression to establish the magnitude of their effect on the general restaurant clientele purchase decision-making process.According to Garson (2014) and Stehlik-Barry and Babivec (2017), multiple regression is useful when testing for relationships and thus was applied in this study.The primary goal of multiple regression was to test the applicability of the EKB model.Thus, based on the established magnitude of the effect of each predictor variable under investigation, the eatery managers can adjust these predictor variables depending on the contribution of each on the outcome variable, given the predictor variables to be tested account for the most variance in the outcome variable (R-squared) (Garson, 2014;Stehlik-Barry & Babivec, 2017).Therefore multiple regression was chosen since it is the best applicable technique that uses an algorithm to select the best grouping of predictor variables that account for the most variance in the outcome variable (R-squared) (Garson, 2014;Stehlik-Barry & Babivec, 2017).Thus, in order of their established magnitude of effect, the 25 predictor variables gave varied contributions on the basis of the resultant B, t-and p-values from the multiple regression analysis coefficients output.Table 6 shows the predictor variables according to their contribution to the restaurant clientele purchase decision-making process in accordance with the EKB model.

TABLe 6 :
Level of individual predictor contribution to the restaurant purchase decision-making model Promotions (photographs of the food on the menu, food presentation, terms used to describe the food etc.) Simon O. Were -https://orcid.org/0000-0002-7055-6296