Brand loyalty towards online accommodation booking platforms

This study investigates the brand loyalty of consumers in the online booking platform industry. The Mabalingwe Nature Reserve served as a case study, while the Moolla and Bisschoff brand loyalty model was used to measure brand loyalty across twelve antecedents. This study aimed to, firstly, validate the model for use in online booking platforms, secondly, to measure the reliability of the data, and finally, to measure brand loyalty across twelve antecedents in online booking platforms. Online questionnaires were distributed via an online link by the booking managers of the game reserve, and 131 responses were captured; this represented a statistically adequate sample as per the KMO measure (.741). The descriptive statistics, using a 5-point Likert scale, showed that Brand trust (4.03) and Customer satisfaction (3.96) are the most important brand loyalty antecedents, while Culture (2.34) is the least important brand loyalty antecedent in an online booking platform. Exploratory factor analysis validated the questionnaire for online booking platforms, while Cronbach’s alpha coefficient (.701) indicated that the reliability of the data is acceptable. Regarding latent variable identification, Brand quality and Brand relationship are two most important factors, respectively, explaining variance of 13.1% and 8.7%. The study culminated in a model to measure and manage brand loyalty of online booking platforms. This model can be operationalized for use by managers, researchers, and academia.


INTRODUCTION
Tourists traveling to South Africa require bookings for accommodation. The Internet has dramatically changed the way to conduct business in the past two decades, and online travel bookings via booking platforms are no exception. The once tedious process of actually contacting each accommodation provider to make a reservation for a specific date, checking availability, making deposits, and providing proof thereof is now easily performed and managed by a booking platform (Satguru Travel, 2019). These booking platforms now link the consumer with the supplier in real time (Bookeo, 2019). Formally, an online booking platform is defined as software that is used to make and manage reservations (Medium.com, 2019). But most platforms also have access to features such as real-time booking and availability, attractive call to action, social media integration (Googleflights.com, 2019), online visual calendar, multiple options to choose from (Hotels.com, 2019), comparative pricing (Trivago.com), location maps (Booking. com, 2019), and online reviews (TripAdvisor, 2019).
This study investigated booking at Mabalingwe Nature Reserve (MNR). The nature reserve is a privately owned reserve located in the Limpopo province of South Africa. The nature reserve has a strict "no day visitors" policy, which means every guest entering the nature reserve has a confirmed accommodation booking within the nature reserve. Most of these bookings are made online via booking platforms.

Brands
The traditional definition of a brand is seen as an organization's logo, sign, name, or any other invisible or visible association that represents an organization so that customers can differentiate its products from their competitors (Shen, 2018). Shen (2018) continues to state that a well-known brand would affect the customers' perceptions and has the potential to evoke comfort with their purchase process. The term "brand" is defined by the influential American Marketing Association (2019) as a name, term, design, symbol, or any other feature that identifies one seller's good or service as distinct from those of other sellers. Alan (2019), however, notes that there might be some confusion about what a real brand is. He explains his view by defining the term brand in detail and by breaking it down into the following four sub-definitions, namely: • psychological and emotional relationship one has with the organization or a brand; • type of product or service manufactured or delivered using a specific name; • identifying a product or service with a name, symbol, design term, or any other attributes, which distinguish it from competitors (Kotler & Armstrong, 2019); • organizational personality the brand represents (Tailor, 2019).
A brand name is a strategic tool that helps organizations to differentiate themselves from their competitors. Customers expect products to have branding, and they often build relationships with brands that they trust. As a result, they will frequently purchase these brands even if identical competitive branded products are available. Some customers will only purchase a particular brand and not even consider substitute products. They will display high levels of brand loyalty and rather go without a specific product, although there are acceptable alternatives on the market (Tailor, 2019). For example, in purchasing a new mobile phone, a customer will not accept viable Huawei phones but insist on buying an Apple iPhone 11 because of their brand loyalty to Apple Inc. (Geldenhuys, 2019). Similarly, some shoppers shop only at Pick n Pay for their groceries or book their accommodation via Booking.com; this shows loyalty to the retail chain.

Brand loyalty
Brand loyalty has many definitions because customers might perceive brand loyalty differently. Some authors place more emphasis on repeat purchases; others consider preferences and various degrees of dedication and commitment towards a specific brand to be loyal, while some consider customer satisfaction similar to brand loyalty. This study underpins the views of TrackMaven (2019) that brand loyalty is the tendency of consumers to purchase one brand's products over another continuously.
Gunelius (2019) states that the repeated purchase action of a brand by consumers can be conscious or unconscious buying behavior; that is because consumers trust that the brand will continue to deliver and satisfy the consumers' needs and meet their expectations. Loyal customers will not purchase a competitor's products when their preferred brand is not available. Some will even go as far as to visit multiple shops to locate their desired brand because they are only comfortable with the specific brand and trust their desired brand (Gunelius, 2019). In this regard, Carol (2019) adds that brand loyalty is a positive association that consumers have towards a particular brand's product, which they demonstrate by their repeated purchase behavior, even though a viable alternative competitive brand's product is available to select. Brand loyalty is a result of a relationship between the customer and the organization. A customer buys a specific brand from a specific organization and evaluates the purchase to see if it meets expectations. Repeated buying instills trust in the product and as soon as trust is established, customers can start building their relationship with a brand (Lamb et al., 2015). Customers then commit to the specific brand and is more likely to enter into repeat behavior of the same brand, while they are also less likely to be influenced by competitors' products or marketing strategies (Pawpa, 2018).

Measuring brand loyalty
The exact wording of Peter Drucker's seminal quotation that what gets measured, gets done, and if one cannot measure it, one cannot manage it or what gets measured, gets managed is unclear; however, whichever wording is preferred, the value of measurement in management is crystal clear (Prusak, 2010). These words of wisdom still ring true today in the management of modern businesses and likewise apply to brand loyalty as a competitive strategy (Tartaglione, Cavacece, Risso, & Granata, 2019). Only by measurements can managers characterize areas of performance and non-performance in brand loyalty. Kyriakidis and Rach (2010, p. 8) also conveyed that in 2010, astonishingly, only one in three organizations measure the performance of their brand. More recently, Gartner (2019) states that an estimated 50% of organizations engage in some form of brand performance measurement, which encompasses in brand impacts on the behavior existing customers, keeping their brand relevant, staying aligned to the changing needs of their target audience, and measuring the impact of brand investments (Gartner, 2019).
Although many brand loyalty measurement models exist, researchers agree that identifying the relevant antecedents for the specific brand is crucial (Jacoby, 1971;Klemperer, 1987 . These 26 influences were reduced to 12 antecedents of brand loyalty (see Table 1), whereafter the model was statistically validated using structural equation modeling. The model has a good fit with a CFI index of 0.82; the index exceeds 0.80 (Bentler, 1990). These brand loyalty antecedents are defined, their respective measuring criteria listed, and the descriptive statistics shown in Table 1.

AIMS
This study aims to: • determine the most preferred online booking platform of guests staying at the nature reserve; • measure brand loyalty towards online booking platforms; • provide a demographic profile of respondents.

Questionnaire
The study empirically investigates brand loyalty within online booking platforms. The twelve brand loyalty antecedents were measured using a 5-point Likert scale, while demographic information was analyzed using inferential statistics and frequencies. The questionnaire consists of two sections: • Section 1 captures demographic information.
The section also had two screening questions to ensure that only respondents who made online bookings complete the survey.
• Section 2 consists of 50 close-ended questions, which captured responses on a five-point Likert scale. The data collection period took approximately six weeks to distribute the questionnaires via an online link that was distributed by the booking manager of the nature reserve.
The questionnaire was successfully adapted, validated, and applied in several industries (such as health [generic and original medicines], banking, public services, and agriculture), but never for online booking research. Therefore, the original questionnaire was also adapted and statistically validated to measure brand loyalty of online booking platforms. The questionnaire captured the brand loyalty responses on a five-point Likert scale (where 1 = strongly disagree and 5 = strongly agree).

Study population
The study population consisted of tourists who visited the nature reserve between January 1, 2019 and August 31, 2019. Only individuals who had made an online accommodation booking were considered part of the population. No sample was drawn.

Data collection
Data were collected online via Google Forms. The data collection consisted of several steps. These are: Step 1. Obtain authorization from the nature reserve to conduct the research.
Step 2. The final questionnaire's Google Forms address was provided to the bookings manager of the reserve who sent a collective email to all the customers who stayed at the nature reserve between January 1, 2019 and August 31, 2019. This email contained an invitation letter to participate in the study and a live link on which the customers could click to transfer them to the first page of the questionnaire, where they found the letter of consent.
Step 3. After reading the letter of consent, respondents could agree by clicking on the "Yes" button to continue to the questionnaire. If they clicked on the "No" tick box, they were thanked for their time and did not receive the questionnaire to complete.
Step 4. If they agreed and gave consent for their data to be used, the questionnaire opened up, and the customers could now complete the questionnaire. Data collected from the sample Step 5. After completion, the data were automatically saved as part of the other responses. It is not possible to identify any respondent (or IP address) to any specific data entry. The data remain anonymous.

Data analysis
A quantitative research design collected data were analyzed using the Statistical Package for Social Science (version 25) (IBM SPSS, 2018). Firstly, the Kaiser, Meyer, and Olkin test was done to determine the adequacy of the sample size and the sphericity properties of the data at the significance level (p ≤ 0.05) using Bartlett's test of sphericity. Satisfactory results led toward exploratory factor analysis. The reliability of the data was determined by calculating the Cronbach's alpha coefficient. The data analysis flow and sequence are explained in Figure 1.

Research ethics
The study was submitted for scrutiny and approval to the North-West University's Faculty of Economic and Management Sciences Ethics Committee. The committee approved the study as a low-risk study and registered a formal ethics number NWU-01320-19-A4.

Descriptive statistics per brand loyalty antecedent
The data from the validated questionnaire were used to measure brand loyalty antecedents. Table  1 shows the descriptive statistics on customer satisfaction. The mean values and standard deviation values were calculated. The standard deviation provides a secondary measure; it indicates whether the respondents agree on the questions, and to what extent (Field, 2009).
The majority (39.7%) of the respondents use an online booking platform at least once every 2-3 months. 34.4% of the respondents indicated that they use an online booking platform at least once in six months. Only 3.1% indicated that they use   Interestingly, 88.5% made their booking for their previous accommodation visit, and only 11.5% of the respondents used a third party to make the booking on their behalf. Some 83.8% of the respondents indicated that they had booked accommodation for leisure. The rest used the facilities for weddings and business conferences. The position of the preferred online booking platform is shared between a local (Lekkeslaap, 34.95%) and an in-ternational (Booking.com, 33.98%) booking platform. These two platforms dominate the bookings made at the nature reserve (see Table 2).
The validity of the model to measure brand loyalty (actually to measure what it is supposed to measure) is crucial because the Moolla and Bisschoff model have been developed for fast-moving consumer goods. Although it was successfully applied in many application settings to measure brand loyalty

<1%
and has proved its worth as brand loyalty measurement instrument in the pharmaceutical industry, agriculture, services industry, pet food industry, and others, it has not done so in online booking platforms. The validity was statistically confirmed by subjecting each brand loyalty antecedent to a factor analysis to confirm if the measuring criteria, indeed, measures the specific loyalty antecedent (Moolla & Bisschoff, 2012;Salim, 2011;Wiese, 2014;Hill, 2018). The results of the statistical validation of the model for the online booking platforms appear in Table 3. Table 3 also shows the measuring criteria about each antecedent in declining order according to their factor loadings.
From the table, it is evident that their respective measuring criteria indeed measured all the loyalty antecedents because all the criteria respectively load under their initial antecedents. One criterion (B1.4, about customer service) was discarded from the analysis because its factor loading was below the required 0.40 minimum factor loading set for this study. Three antecedents consist of sub-factors. These loyalty antecedents are Switching cost, Repeat purchases, and Commitment. Their sub-factors have been labeled and indicated in brackets in Table 4.  The value of the Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy is satisfactory with eight antecedents exceeding the desired KMO value of .6; other four antecedents scored acceptable KMO values of .5 and higher (Field, 2009). Bartlett's test of sphericity is significant (p≤.05); this indicates that there are no strong relationships between the data-points (Miljko, 2017).

Reliability of results
Cronbach's alpha coefficients are calculated to determine if the data are reliable and internally consistent. The closer the Cronbach's alpha coefficient is to 1.00, the higher the reliability and the internal consistency, and a coefficient of .70 or higher is desirable (Pallant, 2010, p. 6

Exploratory factor analysis
Exploratory factor analysis requires data that is suitable for the study. The suitability of the data was tested by the Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy, and the sphericity was determined using Bartlett's test. To extract meaningful data successfully, data should have the following properties: adequate sample (KMO ≥ 0.700), low sphericity (p ≤ 0.05), and explain sufficient variance σ ≥ 60%) (IBM, 2019). These results appear in Table 5. .000 The sample adequacy is satisfactory (0.741), Bartlett's test indicates that p ≤ .05 and score is 0.000 and is lower than 0.05. This means that the data are suitable for exploratory factor analysis to determine how many embedded factors exist and if there is sufficient variance explained by the data.

Retention of factors
Almost 73% of the variance is explained among 13 potential factors. However, although all the factors have eigenvalues (a measure of explained variance) greater than one (as per the Kaiser criterion), a secondary measure called the Point of Inflection can be used complementary to determine how many factors should be retained from the rotated factor matrix. It graphically explains factor variances and is used to measure the discrepancy between the factors. The point of inflection is shown in Figure 2.
From the figure, it is clear that two points of inflection exist: Factors 3 and 9. However, the cumulative variance explained after these factors (28.7% and 59.7%, respectively) suggest that these points of inflection should be further explored. The additional variance explained (13.2%) by Factors 10-13 (after the 2 nd point of inflection) resulted in the inclusion of these factors; thus, over-ruling the evidence of the point of inflection.

DISCUSSION
In this study, a new model was developed to measure brand loyalty in specifically the online booking platform industry. The bookings made by respondents at the nature reserve were used as a case-study approach. The model was developed in three stages.
In the first stage, brand loyalty antecedents were obtained through the research study done by Moolla and Bisschoff. These antecedents were tested in a wide array of other industries and found to be applicable. This study adapted the original questionnaire based on theoretical evidence to suit measuring brand loyalty of online booking platforms. The empirical analysis of each brand loyalty antecedent and its respective measuring criteria were validated using factor analysis; this ensured that the new questionnaire is suitable to measure online booking brand loyalty. The statistical process validated all twelve antecedents. This is a first contribution to the research, a validated questionnaire that academia and managers can use to measure their brand loyalty (or even selected antecedents) of the online platforms they use to secure bookings for their resorts.
In the second stage, these brand loyalty criteria were structured into twelve brand loyalty anteced- In the third and final stage, the model has subjected the measuring criteria to exploratory factor analysis to identify 13 underlying or latent variables of brand loyalty, namely Involvement, Value, Curiousness, Dissatisfied, Economic condition, Brand quality, Brand relationship, Culture, Improvement, Commitment, Switching cost/risk, Loyalty, and Flexibility. Regarding the latent variables, Brand quality (13.1%) is the most important factor, followed by Brand relationship (8.7%). Both these variables have high reliability coefficients (α=0.86 and α=0.89, respectively), indicating that these variables can be regarded as highly constant in brand loyalty management. Economic conditions (3.0%) is the least important factor. Similarly, the most important latent variable will provide higher returns on managerial interventions than the lower order latent variables.
The integrated model appears in Figure 3.
From the above model, it is evident that all twelve original antecedents are all important in measuring the brand loyalty of online booking platforms. The model also shows the relative importance of each of the antecedents in percentage format, and it is evident that Brand trust (80.6%) and Customer satisfaction (79.2%) are two most important antecedents in booking platform loyalty. Culture is the least important antecedent in online booking platforms (46.8%). Regarding the factors, the model shows that Brand quality and Brand relationship are the most important latent variables (explaining variance of 13.1% and 8.7%, respectively). Economic conditions represent the least important latent variable (3.0%).
This model presents a practical tool that can be used by researchers, marketers, and managers to measure brand loyalty in the online booking platform industry. Measuring represents the first step towards managing brand loyalty, and with the in-crease in the number of booking platforms, and their differentiating competitive thrusts, active management of brand loyalty can ensure sustainable business from loyal online customers.

CONCLUSION
The primary aim was to measure the brand loyalty of customers towards online booking platforms in a nature reserve. This was done by adapting an existing brand loyalty model and apply it to online booking platforms. This statistically validated model measured twelve brand loyalty antecedents and identified 13 underlying (or latent variables) of brand loyalty in online booking platforms. The reliability coefficients of the data and each of the antecedents and latent variables were determined, and the sample adequacy ensured. The study culminates in a final integrated model that can be used to measure and manage brand loyalty of online booking platforms. This model integrates the relative importance of the antecedents and the latent variables in a managerial model that can be used by booking platforms targeting nature reserves in South Africa, and by managers to guide them towards listing their nature reserves on the most suitable booking platforms.