Dissertation Methodology on the Impact of Digital Marketing Reviews on Restaurant Reputation and Customer Behaviour

Introduction

This research adopts a quantitative design through performing a choice-based conjoint analysis to study the effects that digital marketing reviews impose on a restaurant's reputation as well as consumer behaviours. Essentially, this model compares between the influence that digital reviews and actual dining experiences have on the customers' behaviours when selecting their preferred restaurants in addition to its reputation. Additionally, the impact of digital reviews is also determined with respect to external aspects that may determine clients' behaviour such as TV and radio advertisements, the restaurants' official websites, word-of-mouth, and Google searches.
Prior to conducting this analysis, the actual dining experiences that customers consider when selecting their preferred restaurants are identified. These factors are taken to be the greatest determinants of customer behaviours as observed in their preferences and choices, and eventually affecting a restaurant's reputation. A presumption is made that customers read digital reviews to acquire more comprehensive details about the restaurant's food and services' quality, and their general dining experiences. Drawing reference from the DINESERV model, five critical aspects, i.e., value, price, service quality, atmosphere and food quality are identified that together with the number of digital reviews have an influence on a restaurant's ability to fully satisfy its clients (Kim, NG & Kim, 2009).

Is your time best spent reading someone else’s essay? Get a 100% original essay FROM A CERTIFIED WRITER!

Save 25%

To understand the effect that digital reviews have on the restaurant's reputation and clients' behaviours, this approach will essentially weigh the average number of consumers who use these reviews in making their choices for preferred dining joints against the averages of the other five factors. These other criteria or factors, that is, quality of food, atmosphere, quality of service, price and value are assumed to directly influence the consumers' contentment and restaurant's reputation while digital reviews have an indirect influence. This is because the five alternative criteria represent the actual consumers' experiences within the restaurant, which determine the number and nature of digital reviews given by these customers after dining in the eateries. It is presupposed that the effect is two-way since the restaurants' management teams are likely to consider these reviews when adjusting their prices and food quality among other aspects.

The choice-based conjoint model, labelled as a trade-off analysis, will be based on two conjectures, that is, the above-mentioned critical factors are defined as a collection of service levels, and, secondly, these levels delineate the vital aspects from the client's viewpoints. Thus, this approach considers the customers' decision-making process, whereby, participants in the study will be made to select between two dissimilar combinations. Additionally, this methodology assists in comprehending how clients trade off one restaurant element against another. Through different analytic combinations highlighted in a questionnaire, the study participants are requested to pick a restaurant from two alternatives. There will also be a "neither of them" alternative. The combinations used in this analysis are presented to the participants one at a time and encompass each aspect stated earlier with such levels as excellent, very good, good, fair and poor. With regards to the order of presentation of these aspects, digital reviews will be presented last.

Sample

American nationals who have visited restaurants more than once and use digital reviews when selecting their preferred restaurants make up the target sample population. On the basis of a study conducted by Orme (2010), who recommends a sample size between 150 and 1200 subjects, 300 respondents will be recruited for this research through a web-based survey devised in the MTurk application.

According to Orme (2010), to determine the smallest sample size, the following equation is employed:

500 paocWhereby: r is the number of participants, a is the number of activities, o is the number of options per activity while c is the number of study levels. For that reason, with two options, five activities and five levels being presented to the subjects, then the minimum number of respondents, r, for this research is calculated as:

500 r525

r= 250

The 300 respondents selected to participate in this research is within the range recommended by Orme (2010), with the 50 additional respondents being an allowance given for invalid responses and surveys that may occur in the course of the study.

MTurk enhances the researcher's ability to diversify their respondents in addition to a seamless access to them. Essentially, this research's link is posted in the participant's web portals through the MTurk application, where they can choose to engage in the study. Questionnaires, which are the main data collection platform in this study, are also administered through this MTurk application.

Questionnaire

In the questionnaire, queries are presented in a structured form that comprises of 18 items that are presented in four parts. The first section indicates the research objectives while the second one focuses on gaining more details about the participants' preferences and choices of restaurants. Moreover, in the second section, a set of three succeeding questions are stated, whereby each question determines the participants' responses which will still be considered in the ultimate data. A 7-point Likert scale is used to gauge the respondents' preferences that relate to the specific attributes being inquired, with 7 equating to "extremely vital" while 1 equates to "extremely insignificant".

In the third question, various contexts are presented to the participants where they are required to select one restaurant from two alternatives with an additional "none of them" option also available. These contexts signify the key elements of the choice-based conjoint analysis, which are designed using Question-Pro. Like MTurk, Question-Pro is also an online survey application that provides an instinctive expert interface for generating opinion poll questions that are circulated through websites or e-mails. Moreover, this platform incorporates review and analysis tools for the responses given. Lastly, the participants' demographic information is collected in the fourth section of the questionnaire. The above features of this questionnaire are illustrated in the appendix.

The five aspects stated earlier, other than digital reviews, which are considered as key elements of the restaurants' reputation since they directly determine the kind of customer experiences at various restaurants are manipulated at five levels, with four of these aspects shown as a star. In this line, 5, 4, 3, 2 and 1-star signify exceptional, good, average, poor and very poor quality respectively. These five levels determine the respondents' choices in each of the five aspects. Yelp.com, a site that lets consumers give their reviews for each eatery is used to obtain the number of reviews left and price ranges of products and services offered by these eateries.

Reliability and Clarity Checks

Data that pertains the number of digital reviews is validated using the pilot test. Through this test, 60 subjects are selected using the MTurk platform and requested to leave their comments on the methodology employed during this research. Responses from these subjects are analysed by deleting those that fell short of answering the succeeding questions and the incomplete ones. Integration of the levels of these digital reviews into the conjoint model then follows, albeit dependent entirely on the findings from this test. The pilot test feedback form is meant to corroborate that the number of digital reviews linked to every variable's level.

To attest the set of questions asked in the questionnaire, Cronbach's coefficient alpha is adopted. This measure tests the internal consistency of these questions, which represent this study's scale. The threshold of this consistency as recommended by Pallant (2013) is 0.7. If the value falls between 0.7 and 0.8, then the internal consistency is acceptable although any value above 0.8 would be desirable. For this study, the SPSS statistics are used to compute Cronbach's coefficient.

The consistency of responses given by the respondents is tested through administering the conjoint questions again after they have completed filling the questionnaire. This can be in the form of another questionnaire but due to time constraints, the second set of questions will be indicated in the same questionnaire as the first set. This test will gauge the validity scale of the chosen participants in adopting the entire evaluative approach in addition to their ability to answer similar queries on several occasions through the hold-out sample process, whereby, an arbitrary sample from the data set is withdrawn and not utilized in the conjoint analysis. This presents an impartial approach to check the respondents' reliability owing to the absence of filters and adjustments to the withheld data sample. At least 75% of the subjects should give similar responses to the first set of questions during the repeated run to corroborate the entire exercise.

In such a case when online simulation data is being used, it may prove to be problematic to assess the reliability. Nonetheless, validity studies indicate the capability of conjoint models to accurately project subjects' behaviour. Validity is conjectured to be the correspondence between projected and experiential preference measures for participants in the real market setting. Consequently, a face validity check is conducted in this research to elucidate on whether the subjects comprehend the sets of choices at their disposal. Face validity represents the degree to which the conjoint analysis adequately tests the respondents' behaviour. The choice-based conjoint model employed in this research should be descriptive of the larger population that it simulates. The pilot study is then employed to check the clarity of the conjoint analysis combinations in addition to comparing findings from this choice-based model and rankings to test the validity of option sets.

Attesting the accuracy of the levels of digital reviews, elimination of defective review scores in the data gathered and corroborating this data make up the main objectives of conducting the pilot study. The responses of subjects who failed their succeeding questions are removed from the ultimate data. The levels of these digital reviews are only deemed to be acceptable for the research if the pilot study indicates that they are accurate. Moreover, the participants are also requested to elucidate on the practicality of the conjoint combinations, with their comments and propositions being deliberated when checking the combinations. Investigating the most critical restaurant aspects from the five elements stated earlier that describe the customers' direct dining experiences, a ranking question is posed to the subjects where they specify their choices from the least to the most important.

Ultimate Data Collection

As stated earlier, 300 subjects have been selected to participate in responding research questions posed to them through an online questionnaire. Analysis of the final data will be conducted through the SPSS and Question-Pro applications. With regards to statistics, the questions asked essentially fall either under demographic or descriptive statistics.

The descriptive data includes the number of digital reviews, quality of service and food, restaurant's rating, price and atmosphere within the restaurants' premises. The subjects are required to rank each of these...