1 | P a g e Product, social media, personal and platform characteristics that influence social commerce adoption and continuance: A South African analysis Buyani Ngcobondwane 1532588 Supervisor: Lucienne Abrahams A research report submitted to the Faculty of Commerce, Law and Management, University of the Witwatersrand, in partial fulfilment of the requirements for the degree of Master of Management in the field of Digital Business Johannesburg, 2022 2 | P a g e DECLARATION I, Buyani Ngcobondwane, declare that this research report is my own work except as indicated in the references and acknowledgements. It is submitted in partial fulfilment of the requirements for the degree of Master of Management in the field of Digital Business at the University of the Witwatersrand, Johannesburg. It has not been submitted before for any degree or examination in this or any other university. Name: Buyani Ngcobondwane Signature: Signed at : Wits Business School On the 25th day of April 2022 3 | P a g e Abstract Problem: Existing research focuses on adoption and continuance separately and does not clearly understand how business, social media, personal, and platform characteristics combined influence social commerce adoption and continuance in the pre and post-adoption phases. Purpose: First, understand how a combination of factors influences social commerce adoption and continuance. Second, build a conceptual model that explains the factors that influence social commerce adoption and continuance. Third, test and validate the model. Fourth, provide recommendations to businesses and social commerce platforms. Methodology: A correlational design was used. Data was collected from 448 social media users of all demographics through an online survey. A multiple regression analysis was used to analyse the data, using adoption and continuance as dependent variables and the remaining variables as independent variables. Findings: In the pre-adoption phase, product/service recommendations, online forums and communities, trust and attitude are positively correlated with social commerce adoption. Ratings and reviews of products/services are negatively related to adoption. Brand is not significantly correlated with adoption. In the post-adoption phase, perceived usefulness and ease of use are all positively correlated with the continuance of social commerce. Satisfaction is not significantly correlated with continuance. Analysis and Conclusion: The true power of the variables studied lies in their combined influence on adoption and continuance. The novelty of this research is that it presents a dynamic and holistic perspective of social commerce adoption and continuance, which is an advance on the static and fragmented view produced in the existing literature. Keywords: Adoption, continuance, social commerce 4 | P a g e Contents Page Chapter 1: Research problem statement and background data: The social commerce opportunity ______________________________________________ 9 1.1 The research context: social commerce adoption and continuance ____ 9 1.2 The research problem statement: pre and post-adoption knowledge gaps ________________________________________________________________ 9 1.3 The purpose of the study: building the pre and post-adoption model __ 11 1.4 The research questions ________________________________________ 11 1.4.1 Main research question ______________________________________ 11 1.4.2 Research sub-questions _____________________________________ 11 1.5 The background of the study: e-commerce and social commerce symbiosis ______________________________________________________ 11 1.6 Understanding the South African context: the social commerce opportunity _____________________________________________________ 13 1.6.1. What does social commerce look like in the South African banking industry? ______________________________________________________ 14 1.6.2 What does social commerce look like in the South African clothing industry? ______________________________________________________ 14 1.6.3 What does social commerce look like in the South African ICT industry? 15 1.6.4 How do we unlock the opportunity? _____________________________ 16 Chapter 2: Literature review and conceptual model: building the model ____ 18 2.1 The building blocks of technology adoption and continuance models _ 18 2.2 Building the pre-adoption phase ________________________________ 20 2.2.1 What are the product characteristics? ___________________________ 20 2.2.2 What are the social media characteristics? _______________________ 22 2.2.3 What are the personal characteristics? __________________________ 24 2.2.4 What is intention to use and use behaviour (Adoption)? _____________ 28 5 | P a g e 2.3 Building the post-adoption phase _______________________________ 28 2.3.1 What are the platform characteristics? ___________________________ 28 2.3.2 What about satisfaction? _____________________________________ 29 2.3.3 What is continuance intention? ________________________________ 30 2.4 So, how does the proposed model work? _________________________ 32 Chapter 3: Methodology: Understanding the execution __________________ 34 3.1 The research methodology: quantitative methodological choice ______ 34 3.2 The research design: correlations _______________________________ 35 3.3 The data collection and analysis: online surveys ___________________ 35 3.4 The population and sample _____________________________________ 36 3.4.1 Population: social media users ________________________________ 36 3.4.2 Sample and sampling method: purposive ________________________ 36 3.5 The research instrument _______________________________________ 37 3.6 Procedure for data collection ___________________________________ 38 3.7 The data analysis and interpretation: multiple regression ___________ 38 3.8 The study’s limitations _________________________________________ 40 3.9 Finding validity and reliability ___________________________________ 40 3.9.1 External validity ____________________________________________ 40 3.9.2 Internal validity _____________________________________________ 41 3.9.3 Reliability _________________________________________________ 41 3.10 The ethical considerations ____________________________________ 43 Chapter 4: Findings: analysing the data _______________________________ 44 4.1 The respondents and their social commerce experience ____________ 44 4.2 What do the descriptive statistics tell us? _________________________ 44 6 | P a g e 4.3 How does a combination of factors and their relationships influence consumer adoption and continuance behaviour on social commerce platforms? ______________________________________________________ 46 4.3.1 In the pre-adoption phase ____________________________________ 46 4.3.2 In the post-adoption phase ____________________________________ 51 Chapter 5: Discussion: unpacking the findings and adjusting the model ____ 54 5.1 Unpacking the pre-adoption phase ______________________________ 54 5.1.1 The product characteristics and adoption ________________________ 54 5.1.2 The social media characteristics and adoption ____________________ 56 5.1.3 The personal characteristics and adoption _______________________ 58 5.2 Unpacking the post-adoption phase _____________________________ 59 5.2.1 The platform characteristics and continuance _____________________ 59 5.2.2 Satisfaction and continuance __________________________________ 60 5.3 The adjusted model ___________________________________________ 60 Chapter 6: Conclusion and recommendations: re-visiting and recommending _________________________________________________________________ 62 6.1 Re-visiting the questions _______________________________________ 62 6.1.1 How do product characteristics influence consumer adoption behaviour on social commerce platforms? _______________________________________ 62 6.1.2 How do social media characteristics influence consumer adoption behaviour on social commerce platforms? ____________________________ 63 6.1.3 How do personal characteristics influence consumer adoption behaviour on social commerce platforms? ____________________________________ 64 6.1.4 How do platform characteristics influence consumer continuance behaviour on social commerce platforms? ____________________________________ 65 6.1.5 How does satisfaction influence consumer continuance behaviour on social commerce platforms? ____________________________________________ 66 7 | P a g e 6.2 Understanding the adjusted model: the dynamic view and the combined strength of the variables __________________________________________ 67 6.3 Recommendations ____________________________________________ 68 6.3.1 Recommendations for businesses ______________________________ 68 6.3.2 Recommendations for Platform creators _________________________ 69 6.4 Suggestions for future research _________________________________ 71 7. References _____________________________________________________ 72 8 | P a g e List of tables Table 1: South African social media users (Kemp, 2021)……………………………. 13 Table 2: Social media followers by company and industry…………………………... 16 Table 3: Research hypotheses………………………………………………………….. 31 Table 4: Rule of Thumb for Interpreting the Size of a Correlation Coefficient (Mukaka, 2012)…………………………………………………………………………... 40 Table 5: Reliability statistics…………………………………………………………….. 42 Table 6: Inter-item correlations…………………………………………………………. 43 Table 7: Social media and social commerce experience……………………………. 44 Table 8: Descriptive Statistics………………………………………………………….. 46 Table 9: Adoption model summary …..……………………………………………….. 50 Table 10: Adoption ANOVA…..………………………………………………………… 50 Table 11: Adoption coefficients………………………………………………………… 50 Table 12: Continuance model summary……………………………………………….. 53 Table 13: Continuance ANOVA…………………………………………………………. 53 Table 14: Continuance coefficients……………………………………………………... 53 List of figures Figure 1: Social commerce combined adoption and continuance model………….. 33 Figure 2: Sample size calculator……………………………………………………….. 37 Figure 3: Adjusted Social Commerce Adoption and Continuance Model………….. 61 List of Annexures Annexure A: Participant information sheet……………………………………………. 83 Annexure B: Informed consent form…………………………………………………… 84 Annexure C: Survey questionnaire…………………………………………………….. 85 Annexure D: Ethics clearance certificate……………………………………………… 90 9 | P a g e Chapter 1: Research problem statement and background data: The social commerce opportunity 1.1 The research context: social commerce adoption and continuance This research study focuses on consumer adoption and continuance of social commerce. Social commerce in this study refers specifically to the transactions of buying and selling on social media platforms by consumers and businesses. It is worth noting that the broad definition of social commerce extends to all commercial activity that takes place on social media platforms. These commercial activities include buying, selling, advertising, and customer relations management. However, this study will only focus on the narrow definition of social commerce which focuses only on the buying and selling component, with a closer focus on the experience of buying from the consumer's perspective. The other core factors of this study are adoption and continuance. Consumer adoption refers to the consumer's intention to use the social commerce platform for purchasing goods or services. In contrast, consumer continuance is the consumer's intention to re-use the social commerce platform for purchasing goods and services. The first research objective is to understand how a number of factors and their combination influence social commerce adoption and continuance. The second research objective is to, build the combined conceptual model that explains the factors that influence consumer adoption and continuance. The third research objective is to, test and validate the newly proposed model as it relates to social commerce in the pre and post-adoption phases. The fourth objective is to, comment and provide recommendations on how businesses can advance their digital marketing capability through their social commerce platforms and how social commerce platforms can attract and retain users. 1.2 The research problem statement: pre and post-adoption knowledge gaps Existing research on consumer behaviour in e-commerce and social commerce does focus on consumer intention, consumer adoption and consumer continuance of social commerce. However, most of the existing literature studies these concepts separately; the research either focuses on adoption on its own or continuance on its own. In other words, the existing research either focuses on the pre-adoption phase 10 | P a g e on its own or the post-adoption phase on its own. Focusing on only one of these phases provides a static understanding of something more dynamic than that. Consumers’ adoption and continuance behaviour are dynamic because the same person can transition to and from one phase to the next. For example, when consumers first start using new technology, they are in the pre-adoption phase, and the adoption stage factors can influence their behaviour. However, consumer behaviour does not just end at adoption; it evolves and transitions into different phases over time. As they continue to reuse the technology, they move into the continuance phase, which also has continuance stage factors that influence their behaviour. This transitioning and shifting is not well understood because existing research provides a separate and static understanding of adoption and continuance instead of a combined dynamic view. Because of this, there is a gap in understanding how consumer adoption and continuance behaviour combine to give a pre-adoption and post-adoption explanation of consumer behaviour on social commerce platforms. This lack of understanding presents the first research gap that this study aims to address. When consumers interact with a social commerce platform, various factors influence their behaviour. Some of the main factors influencing their behaviour are product, social media, personal, and platform factors. Each of these factors has various characteristics that the user interacts with and experiences, which direct and influence their behaviour. For example, the behaviour can be the action of using social commerce to buy and sell, which refers to adoption in this study. The other behaviour can be the decision to continue revisiting and reusing social commerce to buy and sell on an ongoing basis, which in this study is continuance. With adoption being the first step in unlocking consumer value, continuance is the next step towards unlocking lifetime customer value through repeat usage. Therefore, it is essential to understand how product, social media, personal, and platform characteristics drive adoption in the pre-adoption phase and continuance in the post- adoption phase. While previous researchers have studied the influence these characteristics have on adoption and continuance behaviour individually; there is no deep understanding of how these characteristics combined influence consumer adoption and continuance. This lack of understanding presents the second gap in the literature that this study aims to address. 11 | P a g e 1.3 The purpose of the study: building the pre and post-adoption model This research study aims to develop and test a conceptual model that explains the influence of product, social media, personal and platform characteristics on consumer adoption and continuance behaviour on social commerce platforms. The data gathered on these four characteristics were analysed and used to understand their collective influence on consumers’ adoption and continuance behaviour on social commerce platforms. This insight enables the research to fulfil four purposes. First, understand how a number of factors and their combination influence social commerce adoption and continuance. Second, build the combined conceptual model that explains the factors that influence consumer adoption and continuance. Third, test and validate the newly proposed model as it relates to social commerce in the pre and post-adoption phases. Fourth, comment and provide recommendations on how businesses can advance their digital marketing capability through their social commerce platforms and how social commerce platforms can attract and retain users. 1.4 The research questions 1.4.1 Main research question How does a combination of factors and their relationships influence consumer adoption and continuance behaviour on social commerce platforms? 1.4.2 Research sub-questions a) How do product characteristics influence consumer adoption behaviour on social commerce platforms? b) How do social media characteristics influence consumer adoption behaviour on social commerce platforms? c) How do personal characteristics influence consumer adoption behaviour on social commerce platforms? d) How do platform characteristics influence consumer continuance behaviour on social commerce platforms? 1.5 The background of the study: e-commerce and social commerce symbiosis Businesses are expanding their sales channels by using e-commerce to sell their products and services online. Social commerce is a relatively new form of e- commerce that enables social media platforms to facilitate commercial activities. 12 | P a g e Social commerce could become a key component in the future of e-commerce as more social media websites and applications are evolving to become social commerce platforms. This opportunity has brought about new ways to engage with customers and meet their needs. For example, Industries used to rely on their customers coming to where they are, but this way of conducting business is changing as businesses are now meeting customers where they are. In this digital world, customers are online, and more specifically, they are on some kind of social media platform. In 2021, there were 4.2 billion active social media users globally, which is equivalent to 53.6% of the world's population (Kemp, 2021). Because a large portion of the world's consumers are on at least one social media platform, it makes commercial sense for businesses to have a presence on social media. Most companies now have some representation on a social media platform, but this is growing from solely having a presence to a channel they can offer their products and services. E-commerce is also growing; consumers are more accepting of e-commerce and are starting to purchase their goods and services online. E-commerce was already a growing norm, but this was spurred on even further by the Corona Virus pandemic (Bhatti, et al., 2020). In 2021 there were 4.6 billion internet users globally, equivalent to 59% of the world population (Kemp, 2021). Out of the global internet users between ages 16 and 64, 90.4% had visited an online store, 76.8% had purchased from an online store, and 55% had made a purchase online using their mobile phone (Kemp, 2021). The increased affordability of smartphones, increased technology capability on smartphones and the spread of smartphones have made e-commerce apps more accessible. The global growth of e-commerce and social media allows businesses to bring these two worlds together to deliver value to their customers through social commerce platforms. Therefore, it is necessary to understand how businesses can get consumers not just to show intention to use social commerce or just adopt it. Understanding what factors influence the consumers' continuance behaviour on social commerce platforms is just as important. Businesses can benefit from the symbiosis of increased social media usage and increased acceptance of e-commerce. As e-commerce adoption and social media usage grow, they are starting to merge into one in a way that benefits businesses and consumers. If a significant portion of potential and existing customers are 13 | P a g e already using social media, and a further portion is already engaging in e-commerce activities, why not merge those two by providing buying and selling opportunities for customers on the social media platforms they use. 1.6 Understanding the South African context: the social commerce opportunity In South Africa social media adoption is growing among consumers and businesses, as displayed in Table 1 and Table 2 (Kemp, 2021). South Africa had 25 million social media users in 2021, equivalent to 42% of the total South African population and 63% of the South African population between ages 15 to 64 (Kemp, 2021). South African firms have widely adopted social media and mainly use it as a platform for other marketing activities including advertising and customer relationship management, which are all within the broad definition of social commerce. However, as the social commerce capability of these social media platforms has grown, businesses have started to use them as a platform to buy and sell products/services. As a result, their following on these platforms has grown, and they have also expanded into other social media networks. Facebook, Twitter, and Instagram are among the top most popular social media platforms worldwide, and the top firms in South Africa have adopted these platforms to meet their consumers. Table 2 shows the top 16 firms in South Africa in the banking, clothing retail, household retail, and ICT industries. The increase in e-commerce access and activity in South Africa has made buying and selling online more prevalent than it used to be. As a result, consumers will continue to shift toward shopping online (Schaefer & Bulbulia, 2021). In South Africa, in 2021, there were 38 million internet users, which is equivalent to 64% of the population. Of these internet users aged between 16 and 64, 89% had visited an online store, 57% had purchased from an online store, and 38% had made a purchase online using their mobile phone (Kemp, 2021). While these numbers are lower than the global comparisons, e-commerce is becoming a fast-growing commercial activity channel in South Africa. South African social media users Facebook Twitter Instagram 23 000 000 2 300 000 5 400 000 Table 1: South African social media users (Kemp, 2021) 14 | P a g e 1.6.1. What does social commerce look like in the South African banking industry? The South African banking industry has fully embraced social media, using it as a tool for marketing and client relationships in recent years. FNB has branded itself as a leading technology and innovation bank, which has attracted more tech-savvy consumers to the brand. The deliberate branding could explain why FNB has the highest following across the three platforms. In 2018/2017, Standard Bank launched its #WhatsYourNext campaign across all of its social networks, using the hashtag to gain awareness of the campaign and drive its adoption (Standard Bank, 2019). The campaign was successful and continued to succeed as sub-campaigns under the same banner were launched, targeting specific audiences. For example, the #MyFearlessNext campaign targeted entrepreneurs to open business banking accounts and seek funding to start their businesses. FNB and Nedbank followed with their own social media campaigns, using social networks to connect with their clients. In 2018 Absa launched its chat banking services on WhatsApp, which marked the first transaction services on a social media platform offered by any of the banks in South Africa (Absa, 2019). The Chat Banking functionality allows the consumer to perform some of their daily banking functions on WhatsApp and will continue to add more functionality as the capability evolves. The same functionality has since been rolled out to Facebook messenger too. Although the other banks have not fully rolled out their chat banking functions, they are likely to follow as banking on social commerce channels continues to evolve in South Africa. 1.6.2 What does social commerce look like in the South African clothing industry? The emergence of e-commerce-only retail competitors such as Takealot.com and Superbalist has given rise to new competition in the clothing industry in South Africa. As a result, retail giants such as Mr Price and Edgars have had to become tech- savvy and adapt to the new competition. Because of this, these retailers have since amassed large followings on social media. In particular, Mr price fashion has 1 931 397 million followers across its three platforms, with the next highest being Edgars with 743 103 followers. By contrast, Superbalist has 1 281 728 social media followers, and Takealot.com has 629 780 social media followers. 15 | P a g e Mr Price currently has one of the largest social media followings in the South African retail industry. The brand has also begun using Instagram’s buying/selling functionality, allowing Instagram users to shop on the platform. Edgars has recently followed suit and made a few of its products available for purchase directly on Instagram. Smaller competitors like fashion start-ups have taken full advantage of the new social commerce capabilities. They have created relevance, awareness, and accessibility for their customers. The clothing industry is exceptionally competitive on these new social commerce platforms and will continue to be competitive as the functionality and options available to its customers increase. 1.6.3 What does social commerce look like in the South African ICT industry? The ICT industry has adopted social media and is on all of the most popular platforms as they aggressively use social media for marketing. Social media has even become an incentive offered to consumers for ICT companies looking to beat their competitors. In 2015 MTN made Twitter free to all its customers at no additional cost (MTN, 2018). In 2017 Vodacom followed, making Facebook free for all Vodacom users, meaning that they would not need to have data to use Facebook (Venktess, 2017). Furthermore, in 2018 Vodacom launched its social media command centre designed to attend to customers' needs on social media platforms (Business Tech, 2018). This command centre serves as one of Vodacom’s first attempts at social commerce by servicing its customers through social media. On the other hand, MTN announced its intention to launch its own social media platform, dubbed the WeChat of Africa (Mcleod, 2019). The MTN platform is planned to be an instant messaging platform similar to WeChat, where goods and services will also be sold. 16 | P a g e 1.6.4 How do we unlock the opportunity? Technology adoption is commonly represented in phases (Zheng, Gibson, & Gu, 2019). For example, in the pre-adoption sits intention to use and actual adoption, while in the pre-adoption phase sits continuance intention and actual continuance. In each phase, there are different factors affecting consumer behaviour. The phase by phase representation of technology adoption suggests that adoption is a dynamic process that evolves over time (Zheng, Gibson, & Gu, 2019). However, technology adoption research tends to focus on adoption and continuance separately. Studying adoption and continuance this way presents a static view of a dynamic process because it only focuses on the factors affecting one phase and not the other. This approach also provides an incremental and fragmented understanding of consumer pre and post-adoption behaviour (Humbani & Wiese, 2019). Company Industry Social media followers Facebook Twitter Instagram Standard Bank Banking 387 654 135 000 20 200 FNB Banking 1 069 733 241 600 15 900 ABSA Banking 432 129 113 600 8 803 Capitec Banking 516 923 130 800 25 900 Nedbank Banking 312 288 126 000 15 900 Edgars Clothing retail 608 803 31 000 103 300 Foschini Clothing retail 398 278 3 463 130 700 Mr. Price Clothing retail 1 317 097 105 200 509 100 Pick n Pay Household retail 1 958 102 280 000 95 500 Shoprite Household retail 1 011 113 154 000 7 265 Woolworths Household retail 1 796 193 543 000 477 700 Spar Household retail 420 693 837 6 495 Vodacom ICT 1 883 393 425 000 32 800 MTN ICT 1 448 878 1 020 000 19 900 Cell C ICT 915 442 387 000 20 000 Telkom ICT 411 697 270 000 13 700 Table 2: Social media followers by company and industry 17 | P a g e However, the reality of consumer behaviour is that it is more fluid than the static research suggests. For example, a consumer may experiment with shopping on a social commerce platform by buying a t-shirt. This purchase illustrates how the consumer starts exhibiting pre-adoption phase behaviours as influenced by the pre- adoption phase factors. Over time that same consumer may reuse social commerce and start buying other clothing items. This reuse illustrates how the same consumer can transition and start exhibiting post-adoption phase behaviours as influenced by the post-adoption phase factors. This transition from one phase to the next is dynamic, and this dynamic process cannot be explained through static models that study adoption and continuance separately. Furthermore, this lack of understanding does not equip businesses and platforms with the understanding of dynamic consumer behaviour needed to make the most of the social commerce opportunity. For businesses to truly unlock the social commerce opportunity, adoption needs to be understood. However, the actual business benefit comes from continued reuse; therefore, adoption on its own can not be seen as a success. Therefore, it is important to understand the factors that influence a consumer in the pre-adoption phase and the factors influencing that same consumer in the post-adoption phase, thus providing a combined and dynamic understanding of consumer adoption and continuance behaviour. The South African examples discussed above are just a few examples of how local companies have responded to the growth of e-commerce, social media and social commerce opportunities. These opportunities will continue to grow and provide benefits to businesses and consumers. However, the best way to unlock these benefits is to holistically understand the characteristics that influence consumer adoption and continuance of social commerce. 18 | P a g e Chapter 2: Literature review and conceptual model: building the model 2.1 The building blocks of technology adoption and continuance models Cheung, Chan, and Limayem (2005) proposed an online consumer behaviour framework which they built using a combination of established consumer technology adoption models. The models they used to develop this framework are: the Theory of Reasoned Action (TRA), the Theory of Planned Behavior (TPB), the Technology Acceptance Model (TAM) and the Expectation-Confirmation Theory (ECT) (Cheung, Chan, & Limayem, 2005). Their resulting model consists of three main components: consumer intention, consumer adoption, and consumer continuance. Furthermore, these three components are sequential as they follow on and into one another, displaying the different stages that consumers progress through when interacting with new technologies. The theory of reasoned action (TRA) proposes that a person’s behaviour is influenced by their intention to perform said behaviour, and their intention is a function of their attitude toward the behaviour (Fishbein & Ajzen, 1975). The intention is influenced by their attitude towards the behaviour, subjective norms, and perceived behavioural control. In other words, the more positive the attitude, subjective norms and perceived control, the stronger the person's intention to perform the behaviour (Fishbein & Ajzen, 1975). The theory of planned behaviour (TBP) builds on the TRA, placing behavioural intention as the central component of the theory. The TBP suggests that any action a person takes is determined by behavioural beliefs, normative beliefs and control beliefs (Ajzen, 1991). The more favourable the behavioural belief, subjective norm and perceived control, the greater the person's intention to perform the behaviour. The technology acceptance model (TAM) is a theory that explains how users accept and use technology, using cognitive factors as a basis (Davis, 1989). The TAM breaks down the cognitive factors into two types: perceived usefulness and perceived ease of use of the technology in question. Perceived usefulness refers to the user's assessment of how the technology being used improves the execution of the task. Perceived ease of use describes the user's assessment of how easy it is to learn the technology. The higher the perceived usefulness and perceived ease of use, the more likely it is that users will be willing to use the technology. The 19 | P a g e interaction of these two factors increases the user's intention to use the technology (Davis, 1989). The Expectation-confirmation theory (ECT) proposes that expectations plus perceived performance mediated by positive and negative confirmation beliefs result in dissatisfaction or satisfaction (Oliver, 1980). Simply put, if a product outperforms expectations, this leads to satisfaction, but if a product does not meet expectations, then this leads to dissatisfaction. This theory comprises expectations, performance, confirmation, and satisfaction. The expectation is what consumers use to judge performance and develop a confirmation conclusion. The negative or positive confirmation conclusion ultimately leads to a dissatisfied or satisfied user. Consumer purchase intention refers to the consumer's behavioural intention to engage in and make an online purchase. In contrast, consumer adoption is the consumer's acceptance and adoption of the technology displayed by the actual usage of the technology (Cheung, Chan, & Limayem, 2005). Consumer continuance in this research is defined as the consumer's repeated and continued use of the technology after the first time they adopted it. Consumer continuance can be associated with various positive impacts on business, including and not limited to customer loyalty and lifelong customer value. Previous literature focuses on the adoption and continuance of social commerce in a siloed approach. However, the integration between the two is not well understood and has not been studied enough. This research intends to develop and test a conceptual model that explains the characteristics that influence adoption and continuance in the pre and post-adoption phases. This will enable the research to fulfil the main objectives of the study. The first research objective is to understand how a number of factors and their combination influence social commerce adoption and continuance. The second research objective is to, build a combined conceptual model that explains the factors that influence consumer adoption and continuance. The third research objective is to, test and validate the newly proposed model as it relates to social commerce in the pre and post-adoption phases. The fourth objective is to, comment and provide recommendations on how businesses can advance their digital marketing capability through their social commerce platforms and how social commerce platforms can attract and retain users. 20 | P a g e By using social commerce, e-commerce and mobile commerce behavioural factors, this research proposes four main components that influence consumer adoption and continuance: product characteristics, social media characteristics, personal characteristics, and platform characteristics. 2.2 Building the pre-adoption phase 2.2.1 What are the product characteristics? 2.2.1.1 Price In marketing, the price of a product influences the consumers buying decision and determines whether the consumer will purchase the product or not based on their price sensitivity. Price sensitivity can be broken down into two attitudinal dimensions: price importance and price search (Shankar, Rangaswamy, & Pusateri, 1999). Price importance refers to how important price is to the consumer in their purchase decision-making process. Price search refers to a customer's inclination towards conducting a price search to find better prices. Because of the ease of online price comparisons, price search tends to intensify in online mediums, helping the consumer find the best prices online (Shankar, Rangaswamy, & Pusateri, 1999). One of the perceived advantages of e-commerce is more significant savings due to prices being lower in online mediums than in traditional ones (Chiang & Dholakia, 2003). In addition, online shopping also makes price comparisons easier as consumers go from one site to the next to compare prices, which is more challenging in traditional shopping (Chiang & Dholakia, 2003). Therefore, if the online price is perceived to be lower, this drives consumers to use online shopping platforms (Kim, Xu, & Gupta, 2012). Price perception influences a buyer's decision to purchase a product or service. This influence is because the perception of the price explains some information about the product, such as quality, value, benefit or market positioning, giving deeper meaning to the consumer (Munnukka, 2008). Hence, the price is an essential factor in decision-making, even more so for online purchases, because price provides meaning for the consumer. Based on the discussion above, there is evidence that price can influence price-sensitive customers’ online shopping adoption. Therefore, the hypothesis below is presented. 21 | P a g e • H1: Price has a positive influence on the intention to use social commerce 2.2.1.2 Product type While online shopping increases the convenience of shopping for consumers, the online shopping decision does depend on the product type that the consumer is seeking. Online shopping has two broad product types which are, experience goods and search goods (Chiang & Dholakia, 2003). A product is a search good when the product’s dominant product attributes can be known before the product is purchased. In contrast, experience goods are products that have attributes that can only be known when the consumer directly experiences the product. This means that search goods are evaluated through the external information provided before the purchase, while experience goods can primarily only be evaluated upon direct product experience (Chiang & Dholakia, 2003). Therefore, the product type influences whether the product will be bought online or through traditional channels (Cao, 2012). Previous research proves that search goods have a greater influence on online purchase intention than experience goods (Chiang & Dholakia, 2003; Singh & Srivastava, 2018). Consumers associate shopping online for experience goods with requiring higher mental effort as compared to search goods which require less mental effort (Mirhoseini, Pagé, Léger, & Sénécal, 2021). Perceived mental effort and satisfaction are negatively related, meaning the more mental effort the consumer uses, the less satisfied they become (Mirhoseini, Pagé, Léger, & Sénécal, 2021). This suggests that consumers are influenced by the product type when shopping online and are more likely to shop for search goods than experience goods. This means that product type influences online shopping adoption. Therefore, based on this, the following hypothesis is proposed: • H2: Product type influences the intention to use social commerce 2.2.1.3 Brand Brand image is the brand's position in the consumer’s mind as influenced by definitions and evaluations of the brand, the consumption experience, and marketing communications (Bilgin, 2018). Because of the inability to physically evaluate a 22 | P a g e product in online shopping, consumers use their understanding of brand image to judge the product. Their understanding of the brand image helps them make judgements about the product’s value, quality, and benefit before they make a purchase decision (Park & Lennon, 2009; Djatmiko & Pradana, 2016). Social media prominence has resulted in brands developing a more significant online presence to connect with their customers on social networking platforms. Online brands represent the offline brand through digital mediums and build relationships with consumers (Morgan-Thomas & Veloutsou, 2013). The strength of the brand relationship can drive effective commitment and consumer continuance intention (Turri, Smith, & Kemp, 2013). The consumer's attitude toward a brand also influences whether or not they would want to buy that brand’s products (Spears & Singh, 2004). Social commerce has also made it easier for small, lesser-known brands like start-ups to sell their products and services through social media. This accessibility to larger markets introduces a dynamic where small brands can compete with more prominent brands. Based on the critical role brand plays in the consumer decision-making process and the focus on social media branding, the following hypothesis is proposed: • H3: Brand has a positive influence on the intention to use social commerce 2.2.2 What are the social media characteristics? Social media characteristics refer to the activities and functionalities that have empowered consumers to generate content, share information, provide advice and share their own experiences on social media platforms (Hajli, 2015). The main social media characteristics identified by researchers are ratings and reviews; recommendations; and forums and communities (Chen, Lu, & Wang, 2017). These three social media characteristics influence the consumers’ decision-making process (Chen, Lu, & Wang, 2017). Consumers seek online reviews, ratings and recommendations for four motives: information seeking, risk reduction, quality seeking and social belonging (Constantinides & Holleschovsky, 2016). Information seeking is the phase of the purchasing process where the consumer collects and combines information from 23 | P a g e various sources. Using online reviews, ratings and recommendations as sources of information is a low-cost method for making more informed purchasing decisions (Constantinides & Holleschovsky, 2016). These forms of information seeking are additional or alternative to using brand image and price to make purchase decisions online. This information allows consumers to use other people’s experiences as information to reduce the risk of their online purchases. Participating in and bonding with online communities and forums creates a sense of social belonging. This belonging allows the consumer to trust the opinions and experiences of those members, enabling the consumer to use that information to inform their own purchasing decisions (Constantinides & Holleschovsky, 2016). When consumers decide to purchase goods or services on e-commerce platforms, they go through a process to evaluate the prices, quality and delivery of the product or service (Hu, Liu, & Zhang, 2008). Online purchases have the added complexity of uncertainty because the product or service can only be evaluated after consuming it. Therefore, consumers engage in uncertainty reduction behaviours to reduce the uncertainty and perceived risk of the product or service not meeting their expectations. One such tool they use is online ratings and reviews of previous consumers (Hu, Liu, & Zhang, 2008). Consumers also use online reviews to determine product quality and make final purchase decisions (Hu, Liu, & Zhang, 2008). Ratings and reviews allow individuals to post their product or service reviews online, giving comprehensive information about the product or service to other potential customers. This information enables better decision-making for the consumer. Seeing what other consumers have rated a product or service also increases the trust towards the item of interest and leads to increased intention to buy (Hajli, 2015). There are two product and service recommendation types: consumer-generated and system-generated. Consumer-generated recommendations are posted by social media users and are based on their experience with the product or service. System- generated recommendations are personalised recommendations based on the data collected by the platform about the user's past online behaviour (Zhang, Zhao, & Gupta, 2018). In the online world, consumers search for recommendations from other consumers of the product or service to find more information about the item of interest (Farzin & Fattahi, 2018). Senecal and Nantel (2004, p. 161) organised 24 | P a g e recommendations sources into three broad categories “(1) other consumers (e.g., relatives, friends and acquaintances), (2) human experts (e.g., salespersons, independent experts), and (3) expert systems such as recommender systems”. Recommendations influence online product choices among consumers, especially for experience goods that cannot be evaluated before purchasing, causing a higher reliance on others’ experiences (Senecal & Nantel, 2004). Platforms that provide recommendations have also been known to increase perceived usefulness among consumers (Kumar & Benbasat, 2006). Online forums and communities are groups of people who have a common interest or purpose and use the internet or social media platforms to communicate with each other. These groups use the provided platform to communicate, participate in group activities, share information and provide other forms of social support (Chen, Lu, Wang, Zhao, & Li, 2013). For example, Facebook groups, group chats and online message boards are considered online forums and communities. This trusted group of people can influence consumers’ purchase decisions by sharing their experiences and advice with group members, resulting in increased trust and intention to buy (Hajli, 2015). Therefore, based on the influence three factors have on consumers’ purchase intentions, the following hypotheses are proposed: • H4: Product/service recommendations have a positive influence on the intention to use social commerce • H5: Ratings and reviews of products and services have a positive influence on the intention to use social commerce • H6: Online forums and communities have a positive influence on the intention to use social commerce 2.2.3 What are the personal characteristics? 2.2.3.1 Trust Trust has long been regarded as one of the critical components of digital consumer behaviour, influencing the intention to use, adoption and continued use of digital channels (Cheung, Chan, & Limayem, 2005). Trust also influences attitude toward online shopping and online payment for purchases, which are essential components 25 | P a g e of e-commerce usage (Kim, Tao, Shin, & Kim, 2010; Al-Debei, Akroush, & Ashouri, 2015). In this research, trust refers to the belief that one can rely on another party’s promise to deliver on an exchange in a manner that is consistent with one’s own expectations (Hajli, Sims, Zadeh, & Richard, 2017). Due to the intangible nature of online shopping, consumers will always be faced with risks when making online shopping decisions (Kim, Ferrin, & Rao, 2008). When consumers purchase online, they are making a risky decision based on the uncertain actions of others; therefore, trust mitigates the feelings of discomfort caused by the perceived risk (Kim, Ferrin, & Rao, 2008). There are four categories of antecedents that influence consumer trust toward e- commerce: cognition-based, affect-based, experience-based, and personality- oriented trust antecedents (Kim, Ferrin, & Rao, 2008). Cognition-based trust antecedents consist of: ‘Information quality’ refers to a consumer's perception of the reliability and comprehensiveness of the information on the platform about the products and services. Consumers believe that quality information should help them make good purchasing decisions, and therefore if they believe the platform provides quality information, they are likely to view the platform as trustworthy (Kim, Ferrin, & Rao, 2008). ‘Perceived Privacy protection’ refers to a consumer's belief that the platform will protect the confidential data collected during the online transaction from misuse outside of the transaction (Kim, Ferrin, & Rao, 2008). Loss of privacy and protection of their personal data is a priority for most online consumers; therefore, if they do not believe that the platform will protect their data, they will perceive the platform as a risk and have less trust in it. ‘Perceived security protection’ refers to consumers' perception that the platform fulfils their online security requirements (Kim, Ferrin, & Rao, 2008). When consumers encounter security features and protection mechanisms on the platform, he or she perceives the platform to be fulfilling the security requirements necessary to complete the transaction safely. Therefore, consumers' perception of security protection increases their trust in the platform and decreases their perceived risk. 26 | P a g e Affect-based trust antecedents consist of: ‘Positive reputation of the selling party’ refers to the high regard and esteem that the consumer holds for the platform (Kim, Ferrin, & Rao, 2008). Based on the platform's reputation, consumers assume that the platform will continue its past behaviour in the current transaction, which influences whether or not the consumer will complete transactions on the platform. For example, suppose the platform has failed to meet its obligations in the past. In that case, the user will conclude that the platform is risky and will have less trust in the platform. On the other hand, if the platform has a history of fulfilling its obligations, the user will conclude that it is less risky and will have higher trust. Experience-based trust antecedents consist of: A consumer's ‘familiarity with the online selling party’ refers to how experienced the user is with the platform, including how well they know the platform and understand it. Familiarity is a prerequisite of trust because it provides an understanding of the platform's current actions, whereas trust focuses on beliefs about the platform's future actions (Kim, Ferrin, & Rao, 2008). Based on familiarity developed through previous experiences with the platform, users form expectations that the platform will fulfil its obligations, influencing the extent to which they trust it. Consumer personality-oriented antecedents consist of: Consumer ‘disposition to trust’ refers to a person's characteristics that inform their expectations about trustworthiness (Kim, Ferrin, & Rao, 2008). This disposition toward trust is how a person navigates and displays faith in humanity and trust towards others. We all have different cultural backgrounds, developmental experiences and personality types that inform our disposition to trust, which is why we all have a different propensity to trust. If a consumer has a high tendency to trust, they are more likely to trust the platform, whilst those with a lower tendency to trust are less likely to trust the platform. All of these antecedents influence consumers' trust in e-commerce; therefore, the following hypothesis is proposed: • H7: Trust has a positive influence on the intention to use social commerce 27 | P a g e 2.2.3.1 Attitude One’s positive or negative feelings towards a given subject play a crucial role in influencing the use of technologies and innovations (Lim & Ting, 2014). These negative or positive feelings can be defined as attitudes. Edison and Geissler (2003) studied the personal factors contributing to a person's attitude towards technology. The main personal factors that influence a positive attitude towards technology are dispositional optimism, the need for cognition and self-efficacy. Dispositional optimism refers to the extent to which a person holds positive expectations for their future and can cope with external pressures (Edison & Geissler, 2003). For example, more optimistic people have a positive attitude toward new technology. This positive attitude is because optimistic people have less anxiety toward unfamiliar technology and have more positive expectations about their personal ability to succeed in their tasks. The need for cognition refers to the degree to which a person engages in and enjoys the efforts of processing information or thinking (Edison & Geissler, 2003). Individuals who have a higher need for cognition prefer complex tasks over simple ones, whilst people who have a lower need for cognition prefer simple tasks over complex ones. Learning new technology can sometimes require high cognitive effort if the technology is not intuitive. Individuals who have a low need for cognition are more likely to have a negative attitude towards technology. This negative attitude is because they are less motivated to expend effort on cognitive tasks and are therefore less motivated to learn new technologies. Self-efficacy explains the belief an individual has about their ability to succeed or accomplish a task, which influences how they think, behave, recover from failure and motivate themselves (Edison & Geissler, 2003). People with higher self-efficacy are more likely to have a positive attitude towards technology because they display higher levels of self-motivation and resilience, enabling them to approach new technologies confidently. These factors influence attitude towards new technologies, which ultimately influences the adoption of new technology. Multiple studies have proven that positive 28 | P a g e attitudes positively influence the adoption of online shopping while negative attitudes negatively influence the adoption of online shopping (Abdul-Muhmin, 2010; Javadi, Dolatabadi, Nourbakhsh, Poursaeedi, & Asadollahi, 2012; Limbu, Wolf, & Lunsford, 2012; Chang, Chih, Liou, & Yang, 2016). Attitude influences online consumer behaviour, making attitude one of the critical components that dictate consumer behaviour. Therefore the following hypothesis is proposed: • H8: Attitude has a positive influence on the intention to use social commerce 2.2.4 What is intention to use and use behaviour (Adoption)? The theory of planned behaviour determines that performed behaviour is preceded by the intention to perform the behaviour, therefore making intention a predictor of actual behaviour (Liao, Shao, Wang, & & Chen, 1999). The behavioural intention in this study refers to the intention to adopt or use social commerce. Behavioural intention enables specific behaviours to be predicted based on the intention to perform said behaviour (Püschel, Mazzon, & Hernandez, 2010). Therefore, intention to use social commerce in this study is a proxy for actual adoption or actual use behaviour. Adoption takes place when the platform is suitable for the required task and if the platform meets or confirms the prior expectations of the user (Humbani & Wiese, 2019). Conversely, suppose the user encounters obstacles that are unsuitable for the task. In that case, this can lead to the user rejecting the platform. Similarly, if the platform does not meet or confirm their prior expectations, the user is likely to reject the platform. In this study, adoption is the final stage in the pre-adoption phase of the proposed model. 2.3 Building the post-adoption phase 2.3.1 What are the platform characteristics? 2.3.1.1 Perceived Usefulness The extent to which an individual finds the platform useful to increase their performance towards the desired end outcome is referred to as perceived usefulness (Ramayah & Ignatius, 2005). The perceived usefulness of a platform can influence whether or not one will be encouraged to revisit the platform (Bhattacherjee, 2001). 29 | P a g e Technologies that do not help complete the task may lead to consumers finding alternative solutions. Higher levels of perceived usefulness positively influence the adoption and continued use of a digital platform (Hansen, Saridakis, & Benson, 2018; Saeed & Abdinnour-Helm, 2008). Thus, the following hypothesis is proposed: • H9: Perceived usefulness has a positive influence on the continuance intention of social commerce usage 2.3.1.2 Perceived Ease of Use Perceived ease of use refers to the belief one has regarding the ease of acquiring the knowledge needed to use the technology (Kwon & Wen, 2010). Higher perceived ease of use can influence attitude towards and usage of e-commerce, social media, and other digital platforms (Kwon & Wen, 2010; Hansen, Saridakis, & Benson, 2018). When platforms are easy to use, consumers are more likely to continue using that platform, while complex platforms that take time to learn are likely to be a deterrent to continued usage. Therefore, the following hypothesis will be tested: • H10: Perceived ease of use has a positive influence on the continuance intention of social commerce usage 2.3.2 What about satisfaction? Satisfaction refers to the consumer's emotional reaction to a product or service as an evaluation of performance based on customer pleasure and contentment levels resulting from the experience (Abdul-Muhmin, 2010). Consumer behaviours are influenced by how consumers are satisfied with their experience (Wen, Prybutok, & Xu, 2011; Shang & Wu, 2017). Consumers use their previous experiences to determine whether or not they will continue to use a particular technology (Chen, Chen, & Chen, 2009). The satisfaction or disappointment of that previous experience influences their decision to continue using that technology. Although satisfaction does not fit into any of the characteristics groups, previous research suggests that satisfaction is a key influencer of continuance behaviour. Therefore, satisfaction is included as a variable in the model. Based on these findings, the following hypothesis is proposed: • H11: Satisfaction has a positive influence on the continuance intention of social commerce usage 30 | P a g e 2.3.3 What is continuance intention? Continuance intention in this study is defined as the extent to which the individual is willing to continue using social commerce in the future (Chang C. C., 2013). Continuance intention is a predictor and influencer of actual continuance behaviour (Bhattacherjee, Perols, & Sanford, 2008). 31 | P a g e Table 3: Research hypotheses Research hypotheses Phase Characteristic cluster Independent variable Dependent variable Hypothesis H# Pre- Adoption Phase Product Characteristics Price Intention to use Price has a positive influence on the intention to use social commerce H1 Product type Intention to use Product type influences the intention to use social commerce H2 Brand Intention to use Brand has a positive influence on the intention to use social commerce H3 Social Media Characteristics Product/Service Recommendations Intention to use Product/service recommendations have a positive influence on the intention to use social commerce H4 Product/Service Ratings and reviews Intention to use Ratings and reviews of products and services have a positive influence on the intention to use social commerce H5 Online Forums and communities Intention to use Online forums and communities have a positive influence on the intention to use social commerce H6 Personal Characteristics Trust Intention to use Trust has a positive influence on the intention to use social commerce H7 Attitude Intention to use Attitude has a positive influence on the intention to use social commerce H8 Post- Adoption Phase Platform Characteristics Perceived Ease of use Continuance Intention Perceived usefulness has a positive influence on the continuance intention of social commerce usage H9 Perceived Usefulness Continuance Intention Perceived ease of use has a positive influence on the continuance intention of social commerce usage H10 Satisfaction Satisfaction Continuance Intention Satisfaction has a positive influence on the continuance intention of social commerce usage H11 32 | P a g e 2.4 So, how does the proposed model work? The proposed model is split into two phases, pre-adoption and post-adoption. Consumers can occupy both the pre and post adoption phase over time because they can transition from adoption to continuance. Combining the two phases in one model provides a dynamic and holistic view of the factors influencing the consumer in the pre-adoption phase and the behaviours influencing the same person in the post-adoption phase. Although this study does not focus on the transition itself, it provides a dynamic and holistic perspective of the factors influencing consumer adoption and continuance of social commerce. In the pre-adoption phase, the main components are product characteristics, social media characteristics and personal characteristics. These characteristics influence the intention to use social commerce and therefore influence adoption. The product characteristics are price, product type and brand, all of which are related to the business offering but are evaluated from the consumer's perspective. The social media characteristics are products and service recommendations, ratings and reviews and online forums and communities, all of which are related to interactions and content on the platform but are evaluated from the consumer's perspective. Finally, the personal characteristics are trust and attitude, which the model positions as factors that are internal to the consumer and influence their decisions. In the post- adoption phase, the main components are platform characteristics and satisfaction, which influence continuance intention. The platform characteristics are perceived usefulness and ease of use, which are inherent to the platform's design but are interpreted from the consumer's perspective. This model extrapolates variables in the existing literature that influence consumer behaviour and uses them to build the social commerce adoption and continuance model. Furthermore, the model is also an essential component of the research design that guides the analysis in this study. 33 | P a g e Actual use behavior (Adoption) Post Adoption Phase Continuance intention Intention to use Product Characteristics: Price Product type Brand Social media characteristics: Product/service recommendations Ratings and reviews of products/services Online Forums and communities Personal Characteristics: Trust Attitude Platform Characteristics: Perceived usefulness Perceived ease of use Satisfaction Pre-Adoption Phase Figure 1: Social commerce adoption and continuance model 34 | P a g e Chapter 3: Methodology: Understanding the execution 3.1 The research methodology: quantitative methodological choice The literature review discussed the factors that influence consumers’ adoption and continued use of social commerce platforms, emphasising the product, social media, personal and platform characteristics. This study aimed to understand how a number of factors and their combination influence social commerce adoption and continuance, build the combined conceptual model that explains the factors that influence consumer adoption and continuance, test and validate the newly proposed model as it relates to social commerce in the pre and post-adoption phases and lastly comment and provide recommendations on how businesses can advance their digital marketing capability through their social commerce platforms and how social commerce platforms can attract and retain users. This quantitative study aimed to achieve all of these by exploring the correlations and relationships between product, social media, personal and platform characteristics and their influence on consumer adoption and continuance behaviour in social commerce. This research also proposed a social commerce adoption and continuance model that brings all these components into a pre and post-adoption understanding of the social commerce behaviour of consumers. Therefore, the research question addressed by this study is: How does a combination of factors and their relationships influence consumer adoption and continuance behaviour on social commerce platforms? In this study, a quantitative methodology was used to address the research question. Quantitative research is a research methodology in which variables are measured in quantifiable forms from samples of a population to confirm or test hypotheses (Kothari, 2004). The quantitative methodology was suitable for this study because it enables theoretical models to be tested and confirmed. It also allows for attitudes, opinions, and behaviours to be measured quantifiably. Lastly, it enables the correlation between variables to be tested. In previous studies, consumer adoption and continuance of social commerce have been studied quantitatively and qualitatively; however, most of these studies examined these two variables separately. This research study used a quantitative methodology to explore a combined adoption and continuance model. 35 | P a g e Prior to running this study, the assumptions for this quantitative methodology were as follows: • The survey responses will provide an understanding of the factors that influence adoption and continuance behaviours on social commerce platforms. • The quantitative approach will enable the combined adoption and continuance model to be tested, considering the factors that influence consumer adoption and continuance behaviours on social commerce platforms. Both assumptions were accurate and validated by the results chapter's findings. 3.2 The research design: correlations A correlational design was used for this study. Correlational research is used to establish relationships between two or more variables in the same population (Curtis, Comiskey, & Dempsey, 2015). A correlational design allows us to see whether there is a relationship between variables and the strength and direction of the relationship (Tillbrook & Crawford, 2014). In addition, it allows us to assess individual behaviour as it relates to the research topic (Tillbrook & Crawford, 2014). Because this study aimed to explore the correlations and relationships between the factors that influence social commerce adoption and continuance and propose a conceptual model, it was appropriate to use a correlational design. 3.3 The data collection and analysis: online surveys Data was collected through an online survey questionnaire hosted on survey monkey. Online surveys provide access to populations that would be difficult to access with traditional surveys (Wright, 2005). Therefore, for ease, accessibility, and convenience, an online survey was used to collect data for this study. Another benefit of online surveys is that they allow researchers to reach large audiences in a shorter time (Wright, 2005). In order to collect rich data and achieve the purpose of this study, it was important to reach a large sample size. Opting to use an online survey helped reach many people, even surpassed the desired sample size, and gave credibility to the findings of this research. Online surveys also help save costs compared to traditional paper-based surveys (Wright, 2005). It would not have been feasible or cost-efficient if all the survey instruments were printed and administered 36 | P a g e in person for such a large sample. Therefore, the online survey offered a cost- efficient alternative. The survey consisted of close-ended questions to measure responses using a six- point Likert scale. The answers ranged from: strongly disagree, disagree, slightly disagree, slightly agree, agree, to strongly agree. A Likert scale is a measurement scale with multiple categories for respondents to indicate their attitudes or feelings about the question being asked or the statement being proposed (Nemoto & Beglar, 2014). A six-point Likert scale was selected because there is no mid-point or neutral option. A mid-point option can cause some challenges in the statistical modelling and can often become a default choice by unmotivated survey respondents; therefore, excluding a mid-point allows for greater accuracy and response relevance (Nemoto & Beglar, 2014). 3.4 The population and sample 3.4.1 Population: social media users The population studied in this research were consumers of all races, genders, and ages who use social media platforms. This research did not study demographics; therefore, race, gender, and age were not used to determine the sample. Social commerce is a form of e-commerce in which goods and services are sold via social media. Therefore, the population used for this study were social media users of all races, genders, and ages. 3.4.2 Sample and sampling method: purposive Purposive sampling was used to select the sample for this research. Purposive sampling is used to select a fit for purpose sample relevant to the topic being researched (Etikan, Musa, & Alkassim, 2016). Because this study focused on people who already use social media platforms, purposive sampling was used to help select consumers that use these platforms. Figure 2 displays the Raosoft sample size calculator used to determine the sample size for this research. The margin of error, confidence level, population size and response distribution were all used to calculate the sample size for this study. A 95% confidence level was used for this study as it is widely accepted as a conservative range for confidence levels in most quantitative studies (Israel, 1992). Consequently, 37 | P a g e a 5% margin of error was used as the cut-off for this research. In January 2021, there were 25 million social media users in South Africa (Kemp, 2021). Because the population studied in this research was social media users, all 25 million were selected as the eligible population for this study. The 50% response distribution was kept by default. Using the Raosoft sample size calculator, which considers all the above factors, this research study's recommended sample size target was 385 people. 3.5 The research instrument A 47 question online survey (Annexure C) was used to gather participant responses. The survey items were adapted from various sources which have used similar construct definitions previously. Two to four items were used to measure each construct in order to avoid the survey being too lengthy or repetitive. Lengthy surveys could potentially result in participant fatigue and longer response times, therefore in order to avoid this and manage the length of the survey, it was appropriate to cap the number of items per construct in this range. The survey questions measured the main variables of the study, which were: • Product characteristics Figure 2: Sample size calculator 38 | P a g e o Price o Product type o Brand • Social media characteristics o Product and service ratings and reviews o Product and service recommendations o Online forums and communities • Personal characteristics o Trust o Attitude • Intention to use (Adoption) • Platform characteristics o Perceived usefulness o Perceived ease of use • Satisfaction • Continuance intention (Continuance) 3.6 Procedure for data collection Social media posts were used for advertising the survey to solicit responses from willing participants. The survey was also distributed to the Wits database. The surveys were completed online by 448 respondents. The respondents were required to read the introduction provided to them to give them the context of the study. They were then required to read the informed consent form and confirm that they consented to participate in the study. After completing these two sections, they could proceed to the survey questions for their completion. 3.7 The data analysis and interpretation: multiple regression Correlation refers to the extent to which two or more variables are related, explaining the size and direction of the relationship (Prematunga, 2012). Because this study explored the relationship and influence of the various components on adoption and continuance, a multiple regression correlational analysis was used to analyse and interpret the data in this study. Multiple regression tests the relationship between multiple independent variables and a single dependent variable. Multiple regression 39 | P a g e uses the independent variables to predict the dependent variable. Therefore based on the study’s intention to explore the relationship between the variables and their combined influence on social commerce adoption and continuance, a multiple regression is appropriate. Professional assistance was required to process the technical components of the data and regression through SPSS, therefore a professional quantitative consultant was recruited only for that purpose. The interpretation of the data was not done by the consultant. In the pre-adoption phase, multiple regression was done, using the intention to use social commerce as a dependent variable and the remainder of the characteristics as independent variables. Similarly, multiple regression was done in the post- adoption phase, using continuance as a dependent variable and the remainder of the characteristics as independent variables. When doing correlational analysis, the relationship's magnitude, direction, and statistical significance must be used to validate the model's predictive power (Prematunga, 2012). The magnitude or the effect size refers to a quantitative reflection of the phenomenon being studied to address the research question (Kelley & Preacher, 2012). It is important to measure the effect size because, even if the relationship is statistically significant, the effect size may be so small that it is negligible. In this study, the magnitude or effect size of the relationships is measured using Pearson’s correlation coefficient. Pearson’s correlation coefficient also measures the direction of the relationship between variables; therefore, the directions of the relationships are measured using Pearson’s correlation coefficient. The size and direction of the relationship are interpreted as displayed in Table 4. As mentioned previously, statistical significance is determined by confidence intervals and p values. The confidence interval used in this study will be 95%, and the p-value will be 0.05. 40 | P a g e 3.8 The study’s limitations • Correlation cannot prove causality; therefore, the research results do not prove a causal relationship between the variables. • Due to limited resources, a pilot study could not be conducted; therefore, the reliability of the survey questions was only tested post administering the survey. Some questions did not meet the reliability threshold; therefore, those variables were removed from the hypothesis testing. Future research can reinclude these variables using different and better items to measure. • The quantitative approach does not provide a deep understanding of why the variables influence consumer adoption and continuance. A qualitative approach can be used in future to understand why the variables influence consumer adoption and continuance. 3.9 Finding validity and reliability 3.9.1 External validity External validity is the extent to which the inferences that are drawn from a particular study and its sample apply to a broader population or other specific populations (Findley, Kikuta, & Denly, 2021). There are two qualifying criteria for external validity. They are the sample of the study, also known as population validity and the environment of the study, also referred to as ecological validity (Jhangiani, Chiang, Cuttler, & Leighton, 2019). The sampling method used for this research was purposive sampling, which targeted South African social media users only. Purposive sampling is associated with lower external validity because the criteria of the sampling method limit the generalisability to a broader population. Therefore, the population validity for this study is low. This study was not an experiment and was not conducted in a manipulated or controlled environment. Therefore, the ecological Size of Correlation Interpretation .90 to 1.00 (-.90 to – 1.00) Very high positive (negative) correlation .70 to .90 (-.70 to – .90) Strong or Large or High positive (negative) correlation .50 to .70 (-.50 to – .70) Moderate positive (negative) correlation .30 to .50 (-.30 to – .50) Small or Low positive (negative) correlation .00 to .30 (-.00 to – .30) Weak or Negligible correlation Table 4: Rule of Thumb for Interpreting the Size of a Correlation Coefficient (Mukaka, 2012) 41 | P a g e validity of this study is moderate. Overall, the external validity of this study is moderate, which makes it generalisable to some but not all populations (Jhangiani, Chiang, Cuttler, & Leighton, 2019). 3.9.2 Internal validity Internal Validity refers to the extent to which the cause and effect relationship of the variables in a study cannot be explained by other factors (Jhangiani, Chiang, Cuttler, & Leighton, 2019). Non-experimental research is often low in internal validity because it is conducted in natural settings without manipulating variables or the environment. This study is not an experiment and was not conducted in a controlled environment. Therefore, this study has low internal validity because no controls or manipulations were implemented. 3.9.3 Reliability Reliability can be described as the consistency of the measures in a study (Jhangiani, Chiang, Cuttler, & Leighton, 2019). The three types of consistency that are often considered are test-retest reliability, internal consistency, and alternate form reliability. Internal consistency was used in this study to measure the reliability of the survey items. Internal consistency measures the correlations between items to see if they measure the same construct. When items produce similar scores and have a strong correlation, they have high internal consistency and therefore measure the same construct reliably. Inter-item correlation can be used to measure the internal consistency of survey questions. Inter-item correlation examines the correlations between items and uses the average of those correlations to ensure the questions measure the same construct and are not redundant. The reliability of the scales was tested using Cronbach’s alpha to measure internal consistency. According to Bryman and Bell (2007), Cronbach alpha values of 0.70 and above are generally accepted as a rule of thumb to denote a good level of internal reliability. Clark and Watson (1995) suggest that the average interitem correlation of items should ideally be between 0.1 and 0.5. Cristobal, Flavian and Guinaliu (2007) suggest that item-total correlations above 0.3 are acceptable. The item-total correlations for this study were all above 0.3, except for price and product type. 42 | P a g e Table 5 displays the outcome of the reliability testing. Each variable is listed in the variable column, and the corresponding alpha is listed in the Cronbach’s Alpha column. Table 6 shows the inter-item correlation for each variable. As shown in Table 5, online forums and communities, Trust, Attitude, Perceived usefulness, Perceived ease of use and continuance all had a Cronbach’s alpha greater than 0.7. Product/ Service recommendations, Brand and Intention to use have alphas below 0.7 but above 0.6. No items could be considered for removal to increase the alpha. Nunnally and Bernstein (1994) suggest that alphas between 0.6 and 0.7 could be acceptable, especially in exploratory studies. Because this study is exploratory, these reliability values are acceptable. Price, Product type and Satisfaction all have alphas between 0.5 and 0.6 below the acceptable threshold. However, satisfaction has an inter-item correlation of 0.416, above the acceptable threshold of 0.3. Therefore, Satisfaction was included in the Reliability statistics Variable Cronbach's Alpha Cronbach's Alpha Based on Standardised Items N of Items Attitude 0.751 0.753 3 Brand 0.641 0.641 3 Continuance 0.818 0.818 3 Intention to use 0.676 0.677 2 Online forums and communities 0.728 0.736 4 Perceived ease of use 0.711 0.715 3 Perceived usefulness 0.748 0.748 3 Price 0.420 0.416 4 Product and service recommendations 0.648 0.650 3 Product type 0.513 0.529 4 Ratings and reviews 0.788 0.791 4 Satisfaction 0.585 0.587 2 Trust 0.727 0.731 5 Table 5: Reliability statistics 43 | P a g e analysis, whilst Price and Product type were excluded from the model and the analysis. 3.10 The ethical considerations The Wits Human Research Ethics Committee (Non-Medical) approved this research process and provided an ethics clearance certificate (annexure D) and an ethics clearance number (WBS/BA1532588/139). This study followed all the stipulated ethical standards for research as instructed by Wits University. All participants were adequately informed of the purpose of the research to ensure that they were fully aware of what they were consenting to. All participants were surveyed after they gave their consent for participation and the analysis of their responses. Inter-Item Correlations Variable Mean Minimum Maximum Range Maximum / Minimum Variance N of Items Attitude 0.504 0.479 0.532 0.053 1.111 0.001 3 Brand 0.373 0.342 0.398 0.057 1.166 0.001 3 Continuance 0.600 0.554 0.651 0.097 1.176 0.002 3 Intention to use 0.512 0.512 0.512 0.000 1.000 0.000 2 Online forums and communities 0.411 0.312 0.578 0.266 1.853 0.008 4 Perceived ease of use 0.455 0.432 0.495 0.063 1.145 0.001 3 Perceived usefulness 0.498 0.474 0.537 0.064 1.134 0.001 3 Price 0.151 -0.044 0.481 0.525 -10.888 0.047 4 Product and service recommendations 0.383 0.307 0.441 0.134 1.438 0.004 3 Product type 0.219 0.054 0.274 0.220 5.062 0.006 4 Ratings and reviews 0.486 0.394 0.550 0.155 1.394 0.004 4 Satisfaction 0.416 0.416 0.416 0.000 1.000 0.000 2 Trust 0.353 0.248 0.510 0.262 2.057 0.009 5 Table 6: Inter-item correlations 44 | P a g e Chapter 4: Findings: analysing the data 4.1 The respondents and their social commerce experience As displayed in Table 7, 448 respondents completed the online survey. Of the 448 respondents, only one person indicated that they had never used social media before. This respondent was excluded from all hypothesis testing based on the participation criteria. Therefore, the hypothesis testing is based on the 447 respondents who indicated they had used social media before. Of the 448 respondents, 61% (273) indicated they had used social media to purchase products/services before, meaning they have some social commerce experience. Furthermore, of the 448, 60% (267) indicated that they use social media to purchase products or services, meaning they actively use social commerce for some of their shopping needs. Based on this, most of the participants in this study have some experience with making purchases on their social media platforms, and a large majority continue to use social media platforms to make purchases. 4.2 What do the descriptive statistics tell us? The descriptive statistics table (Table 8) indicates the average responses and standard deviations for all the variables measured in the survey. The survey measured 13 variables; however, because the questions measuring price and product type did not pass the reliability thresholds, they were excluded from the Social media and social commerce experience Item Response Frequency Percent Have you used social media before? Yes 447 99.8% No 1 0.2% Do you use social media to purchase products or services? Yes 267 59.6% No 181 40.4% Have you used social media to purchase any products or services before? Yes 273 60.9% No 175 39.1% Table 7: Social media and social commerce experience 45 | P a g e analysis. Therefore, the table displays the 11 variables that were analysed from the survey responses. Each variable had a minimum of 2 questions that measured it, so all the responses for each variable were combined to give an overall mean measure of that variable. For example, if attitude had four questions, the responses from all four questions were combined to give an average score for attitude overall. Each question was measured on a 6-point Likert scale, with 1 strongly disagree to 6 strongly agree. Thus, the mean score displays the average answer between 1 and 6 from all respondents for each variable. None of the survey items was reverse coded. Therefore the higher the mean is between 1 and 6, the higher the respondents perceived that variable and the lower the mean, the lower the respondents perceived that variable. In other words, this is an indication of how important each construct is to the respondents. Out of the 11 constructs, 7 have a mean greater than 4 out of 6, which means these seven constructs were perceived highly and indicate that they hold some importance to the respondents. The seven constructs were brand, perceived ease of use, perceived usefulness, recommendations, ratings and reviews, satisfaction, and continuance. Of these seven constructs, brand was perceived as the highest, with a mean of 4.65 out of 6. All the questions that measured brand were used to measure the importance of brand to the respondent and the respondent's brand sensitivity. Therefore, this mean score suggests that the respondents are quite brand sensitive and conscious of the brands they purchase from. The remaining four variables were perceived moderately because their mean responses were slightly above 3 out of 6. The four variables that were perceived as moderate are attitude, intention to use, online forums and communities and trust. Interestingly, two of these four variables, trust and attitude, are personal characteristics. Of all the variables, trust was perceived as the lowest because it had a mean score of 3.2 out of 6. This low mean score suggests that the respondents in this survey have lower trust in social commerce sites. Another interesting observation is that intention to use was perceived lower than continuance. The questions that measured intention to use were directed towards measuring intention to use social commerce currently. The continuance questions in the survey were directed toward future reuse. This disparity suggests that the respondents currently have less intention to use social commerce than they will in 46 | P a g e the future. The higher mean for continuance suggests that the respondents acknowledge that they will likely need to use and reuse social commerce in the future. 4.3 How does a combination of factors and their relationships influence consumer adoption and continuance behaviour on social commerce platforms? 4.3.1 In the pre-adoption phase A multiple regression was used to test the pre-adoption phase relationships. The regression outcomes are displayed in Table 9: Adoption model summary, Table 10: Adoption ANOVA and Table 11: Adoption coefficients. The following variables were used as independent variables: brand, product and service recommendations; ratings and reviews; online forums and communities; trust and attitude. Intention to use was used as the dependent variable. The model summary (Table 9) describes how the independent and dependent variables are related to each other. The values in the R and R square columns are important for interpretation. The value in the R column indicates the relationship between the independent and dependent variables combined. In this instance, R = 0.778 indicates a strong positive correlation, which suggests that the model is a relatively good predictor of intention to use social commerce. R square describes Descriptive Statistics Variable N Minimum Maximum Mean Std. Deviation Attitude 448 1 6 3.9687 0.95304 Brand 448 1 6 4.6548 0.87774 Continuance 448 1 6 4.2485 1.08878 Intention to use 448 1 6 3.9163 1.12189 Online forums and communities 448 1 6 3.7718 0.87467 Perceived ease of use 448 1 6 4.5677 0.89597 Perceived usefulness 448 1 6 4.1562 1.01193 Product service recommendations 448 1 6 4.2515 0.88585 Ratings and reviews 448 1 6 4.6501 0.9116 Trust 448 1 6 3.2134 0.89292 Satisfaction 448 1 6 4.2321 0.94046 Table 8: Descriptive Statistics 47 | P a g e how much of the variation in the dependent variable can be explained by the independent variables. In this case, the combination of the six independent variables explained 60.5% of the variance in intention to use social commerce. According to Ellis and Steyn (2003), the effect size of this value renders it practically important. The model summary confirmed that the model is a relatively good predictor of intention to use social commerce. It also confirmed that the independent variables explain a large portion of the variance in the dependent variable. The ANOVA (Table 10) describes whether the findings in the model summary are statistically significant using an analysis of variance. In other words, it indicates if the model is a statistically significant predictor of intention to use social commerce. The value in the Sig. column is the significance value. In this case, p<.001, which is less than the 5% margin of error set out for this research study. Therefore, the model is a statistically significant predictor of intention to use social commerce. The ANOVA confirmed that the model is statistically significant. The coefficients table (Table 11) indicates the extent to which each independent variable contributes to the model. The B column under unstandardized coefficients indicates the correlation between each independent variable and the dependent variable. The value in the Sig. column indicates the statistical significance of the correlation. Analysis of the independent variables shows that only 5 of the predictors were statistically significant in the model. These were Product and service recommendations (B=0.142, p <0.05), Ratings and reviews, (B= -0.110, p <0.05); Online forums and communities (B=0.127, p <0.05), Trust (B=0.265, p <0.05) and Attitude (B= 0.607, p <0.05). Brand (B=0.014, p >0.05) is the only variable that was not statistically significant in the model. Product and service recommendations, online forums and communities, trust, and attitude positively correlate with the intention to use social commerce. Ratings and reviews is the only variable that negatively correlates with the intention to use social commerce. This result is interesting considering that respondents perceived ratings and reviews so highly, with a mean response of 4.5 out of 6. The strength of these correlations is also important to analyse. Product and service recommendations, ratings and reviews, online forums and communities, and trust have a weak correlation with intention to use because they all have a correlation 48 | P a g e coefficient of less than 0.3. Notably, product and service recommendations was also among the variables that were perceived highly, but its relation to the intention to use social commerce is weak. Attitude has the highest correlation with intention to use; however, because it is below 0.7, it falls in the moderate range. This finding is also interesting because attitude was one of the variables respondents perceived low; however, it has the highest correlation with intention to use social commerce. All of these correlations suggest that these variables on their own have a weak relationship with the intention to use social commerce. However, they have a strong relationship with the intention to use social commerce when combined. The outcome of these correlations answers the following research questions: How do product characteristics influence consumer adoption behaviour on social commerce platforms? • Price – Not tested • Product type – Not tested • Brand – Not statistically significant. H3 is not supported. How do social media characteristics influence consumer adoption behaviour on social commerce platforms? • As product and service recommendations increase by one unit, the intention to use social commerce increases by 0.142 units. H4 is supported. • As ratings and reviews increase by one unit, the intention to use social commerce decreases by -0.110 units. H5 is not supported. • As online forums and communities increases by one unit, the intention to use social commerce increases by 0.127 units. H6 is supported. How do personal characteristics influence consumer adoption behaviour on social commerce platforms? • As trust increases by one unit, the intention to use social commerce increases by 0.265 units. H7 is supported. • As attitude increases by one unit, the intention to use social commerce increases by 0.607 units.H8 is supported. 49 | P a g e Validating the pre-adoption phase In summary, hypotheses 4, 6, 7 and 8 were all supported, while 3 and 5 were not supported. In order to meet the regression assumptions, 6 outliers were removed from the analysis. The assumptions of the homoscedasticity, linearity and normality of residuals were tested and met. The overall regression was statistically significant (F(6, 435) 110.829, p<0.01). The combination of independent variables explained 60.5% of the variance in the dependent variable in the pre-adoption phase of the model. 50 | P a g e Table 11: Adoption coefficients Table 10: Adoption ANOVA ANOVAa Model Sum of Squares df Mean Square F Sig. 1 Regression 322.881 6 53.814 110.829 <.001b Residual 211.216 435 0.486 Total 534.097 441 a. Dependent Variable: Intention to use b. Predictors: (Constant), Attitude, Brand, Ratings and reviews, Online Forums and communities, Trust, Product Service Recommendations Model Summaryb Model R R Square Adjusted R Square Std. Error of the Estimate Durbin-Watson 1 .778a 0.605 0.599 0.69682 0.004 a. Predictors: (Constant), Attitude, Brand, Ratings and reviews, Online Forums and communities, Trust, Product Service Recommendations b. Dependent Variable: Intention to use Coefficientsa Model Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta 1 (Constant) 0.035 0.245 0.141 0.888 Brand 0.014 0.039 0.011 0.358 0.721 Product Service Recommendations 0.142 0.058 0.115 2.443 0.015 Ratings and reviews -0.110 0.048 -0.091 -2.267 0.024 Online Forums and communities 0.127 0.052 0.101 2.435 0.015 Trust 0.265 0.054 0.216 4.916 0.000 Attitude 0.607 0.049 0.527 12.419 0.000 a. Dependent Variable: Intention to use Table 9: Adoption model summary 51 | P a g e 4.3.2 In the post-adoption phase A multiple regression was used to test the post-adoption phase relationships. Table 12: Continuance model summary, Table 13: Continuance ANOVA and Table 14: Continuance coefficients display the regression results. The following variables were used as independent variables: perceived ease of use, perceived usefulness and satisfaction. Continuance intention was used as the dependent variable. The model summary (Table 12) shows that R = 0.722 indicates a strong positive correlation, suggesting that the model is a relatively good predictor of intention to continue using social commerce. The combination of the three independent variables explained 52.2% of the variance in intention to continue using social commerce. As mentioned in the pre-adoption phase, the effect size of this value renders it practically important. The value in the Sig. column in the ANOVA table (Table 13) is p<.001, which is less than the 5% margin of error set out for this research study. Therefore, the model is a statistically significant predictor of the intention to continue using social commerce. Analysis of the independent variables (Table 14) shows that only 2 of the predictors were at the 5% level. These were, Perceived ease of use (B=0.356, p <0.05) and Perceived usefulness (B=0.500, p <0.05). Satisfaction (B=0.060, p >0.05) is not statistically significant in the model. Perceived ease of use and perceived usefulness have a positive correlation with continuance. The correlations in the post-adoption phase are stronger than the pre- adoption phase correlations. Perceived ease of use and usefulness have a moderate correlation with continuance intention because they both have a correlation coefficient of less than 0.7. Similar to the pre-adoption phase, these correlations suggest that these variables on their own have a moderate relationship with the intention to continue using social commerce. However, when combined, they have a strong relationship with the intention to continue using social commerce. The outcome of these correlati