Product recommendation is a common form of personalisation used in various marketing communication channels. Chinchanachokchai et al. (2021) define recommendation marketing as the process by which firms provide product or service recommendations to current and prospective customers based on their personal interests or purchase history. Personalised product recommendations enable retailers to make more targeted product recommendations to consumers, reducing the associated product screening and product evaluation costs (Zhang et al., 2011). In the field of marketing, several scholars have investigated the categorisation of recommendation types. For example, Ansari et al. (2000) argued that mainstream historical preference-based recommendation types can be classified into collaborative and content filtering. Gai and Klesse (2019) classified recommendation types into item-based and user-based recommendations. Kim and Rao (2023) divide online recommendation methods into simultaneous recommendation and sequential recommendation. Lewin (1935) points out that setting goals creates tension that can only be eliminated by goal achievement. By recommending a collection of different products, a completion impulse can be created in the consumer. The completion impulse is triggered by an incomplete set of products, which triggers the desire to accomplish a goal and arguably removes the discomfort of failure (Bauer et al., 2022). Therefore, in this article, we categorise the forms of product recommendation into two categories based on different categories of products: separate recommendation and set recommendation.

Separate recommendation is the recommendation of separate products to consumers (Zhao & Xia, 2021). In separate evaluation, each option is presented and evaluated separately (Bazerman et al., 1999). If the focal options are already attractive in separate evaluation, then subjecting these options to joint evaluation will hurt their attractiveness (Hsee & Leclerc, 1998). Mogilner et al. (2013) found that consumers were more satisfied with what they chose when different product options were presented at the same time than when the options were presented separately in sequence. In separate recommendations, firms do not intentionally assign a specific link between separate products (Zhao & Xia et al., 2021). For example, in a study conducted by Zhao and Xia et al. (2021) burgers and fries are presented separately and individually (even though there may also be some complementary relationship between the products themselves), and it was found that compared with joint presentation, individual presentation leads to lower product evaluation.

Set recommendation is a combination of products from different categories that are recommended to consumers via some connection (see Appendix 1 for details). This has become one of the most common tools for retaining customers and stimulating product usage. When induced by marketing stimuli, people are motivated to pursue consumption-related goals (Cheema & Bagchi, 2011), and set recommendations are the products category derived under the goals. For example, running gear includes items such as running shoes, quick-drying apparel and fanny packs, and the items in the set are closely related to the needs of runners. Multiple goods within the set may also be linked by some goal (Spaid, 2018). Consumers are therefore motivated to purchase products they might not otherwise have considered (Bauer et al., 2022). In the study done by Carey (2008), results indicated that sets can motivate collecting behaviour by increasing the social value of an item to the collector. Evers et al. (2014) found that the collection-matching effect increased consumer diversity seeking. Barasz et al. (2017) demonstrated that even when no explicit goals were provided, the arbitrary combinations of items or tasks as part of an apparent ‘set’ (pseudo-set), either through visual representations or written descriptions, motivate people to reach a perceived point of completion. Zhao and Xia (2021) found that presenting products together was more likely to provide consumers with solutions to problems than presenting separate products and describing their attributes. Bauer et al. (2022) further demonstrated that marketers can increase the likelihood of cross-selling products by using them as part of a set, set of products or other goals to be accomplished.

Building on the literature, numerous studies examine the positive effect of collecting the complete set (Barasz et al., 2017; Bauer et al., 2022). As recommended products may not be the products that consumers initially need to purchase, dividing the dependent variable into initial product purchase intention and cross-buying intention should not be overlooked, hence included in this study. On the one hand, before purchasing all the products in the collection, consumers begin with an established interest in one or more products (otherwise they would not have paid attention to the collection), we call this sentiment or the initial intention to purchase (Zhu & Chang, 2015). On the other hand, when consumers buy an initial product, they begin to think that they need the other products as well. To finalise this situation, they tend to buy complementary products, a desire for instinctive directness called the Diderot effect (Çakaröz et al., 2022). Cross-buying refers to the behaviour of customers who purchase other products and services in addition to those of their existing service providers (Ngobo, 2004). Cross-buying customers may become more loyal and purchase multiple items across product categories (Min et al., 2016); cross-buying is therefore beneficial to merchants and the key to successful cross-selling in inducing subsequent purchases by consumers (Kim & Tanford, 2021). Because of the efforts of marketers, when products that consumers initially want to buy are assembled into a collection with other products and recommended to consumers, those products that would not have been considered receive the attention of consumers, which creates the intention to cross-buy other products. Thus, we propose hypothesis 1:

Goals significantly impact decision-making and guide consumer choice and behaviour (Bagozzi & Dholakia, 1999). In the hierarchy of goals, there are three types of goals: superordinate goals, focal goals and secondary goals (Vatllacher & Weger, 1985). They are arranged in order from the top down in the structure. Superordinate goals answer the question: why do I want to achieve the goal I am pursuing? Focal goals are located in the middle of the goal hierarchy and answer the question: what is the goal I am pursuing? Secondary goals are located at the bottom of the structure and answer the question: how can I achieve the goal I am pursuing? An individual’s behaviour is thought to be controlled by a goal at the middle level of the goal hierarchy (the focal goal), with the initial motivation for engaging in the focal goal coming from the superior goal it serves (e.g. to make one’s appearance look better), and the goals at the lower levels of the structure providing the means to operate to achieve the focal goal (e.g. exercising, dieting) (Bagozzi & Dholakia, 1999; Vatllacher & Weger, 1985). We can imagine the higher-level (high-level) goals as ends or standards, the lower-level goals as operational levels and the focal goal at the centre of the goal structure, with the focal goal being rationalised through the higher- and lower-level goals, which are shown to be interconnected through inferential relationships (Bagozzi & Dholakia, 1999). Therefore, individuals make decisions with many goals, but one goal dominates the framing process and that goal is the focal goal (Friedman et al., 2018).

In the current research on the focal goal in marketing, the goal of products serving consumers is referred to as the focal goal, and the consumer’s decision to purchase a particular product depends on the importance given to the ‘goal that the product serves’ (Friedman et al., 2018; Lee & Chu, 2022). Consumption is about achieving a goal and consuming a product to fulfil a need; such as a consumer’s decision to ‘buy shirts and overalls’ (a sub-goal) is derived from the ‘goal of dressing for work’ (a focal goal) (Friedman et al., 2018). Therefore, the purchase behaviour (sub-goal) is derived from the focal goal and the purchased product serves the focal goal within the consumer. Focal goal is, therefore, a combination of motivation and activated cognitive structures. Thus, a person’s motivation and actions are influenced by the ‘focal goal’, which changes as cognition evolves (Dong et al., 2023). Hence, we propose hypothesis 2:

According to regulatory focus theory (Higgins, 1997), people have two distinct motivational management systems that govern how they pursue goals: promotion focus and prevention focus. These two self-regulatory systems can be either chronic tendencies or situationally induced in individuals (Aaker & Lee, 2001; Higgins et al., 2001). Individuals who are promotion focussed are sensitive to the presence or absence of positive outcomes (e.g. gained and unearned), are motivated by proximity and are oriented toward the pursuit of achievement and growth consistent with a desired state (Camacho et al., 2003). Prevention focussed individuals are more sensitive to the presence or absence of negative outcomes (losses and non-losses), are typically avoidance motivated and have a security or vigilance orientation consistent with their ought state (Higgins, 1997). Therefore, according to regulatory focus theory, when focal goal importance is high, individuals will exhibit a promotion focus and thus come close to collecting all the products in the set, meaning that a set recommendation is beneficial to the initial product purchase intention and cross-buying intention. However, when focal goal importance is low, individuals will exhibit a preventive focus and avoid the financial loss or loss of energy that set purchasing brings to them. In that case, the set recommendation inhibits initial product purchase intention and cross-buying intention.

Construal level theory (Trope & Liberman, 2003) indicates that the level of abstraction at which people make mental representations of cognitive objects is classified as high or low level of construal. When adopting a high level of construal, people use abstract, essential and general features to represent events, whereas, when initiating a low level of construal, they adopt concrete, surface and local features to represent events (Labroo & Patrick, 2009). In this research, set recommendation is a joint presentation of different products, which is a detailed and concrete description of the thing, a low level of construal, and separate recommendation is a separate and independent presentation of different products, which is an abstract overview of the thing, a high level of explanation. Therefore, this article argues that when focal goal importance is high, it drives consumers to have enough patience to understand the related products within the set, and a low level of construal will occur, and the recommendation for the products should be a set recommendation. When focal goal importance is low, the recommendation for the different products will take the form of abstract representation, and a high level of construal will occur. In that case, separate recommendation is better.

Based on the above analyses, this research concludes that there is an interaction effect of focal goal importance and the product recommendation type on consumer’s initial product purchase intention and cross-buying intention. Thus, we propose hypothesis 3:

Human memory capacity and perceived information level are very limited, and if an individual receives an amount of information that exceeds their information capacity in a certain period, this sudden influx of information will produce a certain ‘load’ on their cognitive system, termed the cognitive load (Sweller, 1988). The level of cognitive load refers to the level of mental energy required to process information (Sweller, 1988). When competition for cognitive resources is high, individuals experience high volume of interference, and this distraction interferes with problem-related thinking (Jeong & Hwang, 2015). Pantoja et al. (2016) found that cognitive load affects consumers’ attitude towards brands. It reduces psychological resources and weakens spontaneous mental simulation (Cian et al., 2020). Therefore, this research argues that when focal goal importance is low, the various types of product information appearing in the set recommendation will cause consumers to experience information overload selection difficulties, increasing their cognitive load and reducing their purchase intention for the initial product. Therefore, we propose hypothesis 4a:

Incompleteness is an unsettling and irremediable feeling that one’s behaviour or experience is not ‘just right’ and underlies many of the symptoms of obsessive-compulsive disorder (OCD) (Summerfeldt, 2004). According to self-completion theory, it occurs when people realise that they are falling short of what is required and is an aversive self-evaluative state (Gollwitzer, 1987). The ‘incompleteness effect’ draws from psychological studies on the human inclination to seek completion, applying this concept to consumer behavior. Zeigarnik (1927) posits that tasks left unfinished are more likely to be recalled than those that have been completed. An important sub-dimension of consumer purchase planning is completeness, which refers to the degree of specificity of each part of the course of action, and may result in more complete plans for consumer tasks that are of high importance or difficult to achieve (Bagozzi & Dholakia, 1999). Abramowitz et al. (2010) suggest that perceived incompleteness, which focusses on symmetry, completeness and the need for things to be ‘just right’, is associated with OCD. Cheema and Bagchi (2011) further increased perceived incompleteness by presenting a visualisation of a collection, increasing people’s drive to complete and leading to a greater likelihood of collection completion. People often have an innate need for visual completeness, and once people think in terms of collections, they are motivated to reach a satisfactory and complete endpoint (Barasz et al., 2017). Complete perceptions are processed more easily and quickly than incomplete judgements, especially when facing more complex multi-attribute products in more challenging decision-making environments (Harahap et al., 2018). The intrinsic need for visual completeness drives people to spontaneously perceive complete images from incomplete displays (Gerbino, 2020). As a result, consumers are increasingly interested in unused product categories, and the proportion of customers purchasing the full range of products is many times higher than in traditional marketing settings (Bauer et al., 2022). Ladeira et al. (2023) propose a display incompleteness effect model to analyse the impact of product information processing. Therefore, this research argues that when focal goal importance is high, consumers perceive extreme incompleteness within themselves if they do not purchase the full range of items within the set, and that this perception of incompleteness increases as focal goal importance increases, which in turn affects cross-buying intention. Therefore, hypothesis 4b is proposed:

The theoretical framework and the research guideline of this research is presented in Figure 1.

Study 1 builds on previous research (Barasz et al., 2017; Bauer et al., 2022) and explores the effect of set recommendations (vs. separate recommendation) on consumers’ initial product purchase intention and cross-buying intention, with the aim of validating H1. This article hypothesises that people will have higher initial product purchase intention and cross-buying intention when products in different categories are set recommended (vs. separate recommended). We manipulated the form of product recommendation as detailed in Appendix 2.

The purpose of Study 2 was to explore the positive effect of high (vs. low) focal goal (health) importance on the purchase intention and cross-buying intention toward the relevant health product.

The purpose of Study 3 was to explore the interaction effect between product recommendation type and focal goal importance on consumers’ initial product purchase intention and cross-buying intention and to verify the mediating mechanism of cognitive load on initial product purchase intention and the mediating mechanism of perceived incompleteness on cross-buying intention.

The purpose of Study 4 was also to explore the interaction effect of product recommendation type and focal goal importance on consumers’ initial product purchase intention and cross-buying intention, as well as their underlying mechanisms. Different from Study 3, Study 4 tests the hypotheses using a field study and a different consumer category, which enables the testing of the generalisability of Study 3 results to a new sector.

This article does not contain any studies involving human participants or animals performed by any of the authors.

Firstly, the findings of this study can inform shopping platforms on whether to choose set or separate recommendation when recommending products to different users. Personality traits may be particularly important in online shopping behaviour (Wang & Keh, 2017). Consumers typically spend 3 s – 7 s on a product, which is known as the first moment of truth in the industry (Han et al., 2022). Sohn and Ko (2021) proposed a framework for integrating personalisation into complementary product recommendations to suggest compatible products based on product relationships and user preferences. As a result, the user habits and recent browsing volume can be used to infer the level of importance that users attach to a focal goal, and when users attach high importance to focal goal, the push of related products can be selected as a set recommendation. However, when users attach low importance to a focal goal, it may be more prudent to select separate recommendations for the push of related products.

Secondly, the importance that consumers place on the focal goal served by a product can change. Firms can increase the importance of the consumer focal goal served by a product through advertising and other promotional means. Psycholinguistic research suggests that goal-derived categories do not come to mind naturally but must be prompted (Gentner & Kurtz, 2006; Gibbert & Mazursky, 2009). For example, by promoting the importance of running, the focal goal importance of ‘running’ may be increased across the whole population, which in turn increases the sales of a collection of running-related products (e.g. running shoes, jerseys, fanny packs). For example, by promoting the better life brought by ‘smart’, Yummi increases consumers’ purchase intention to buy smart home appliances. Sellers should also develop related advertising and marketing programmes based on product characteristics and customise marketing recommendations from the perspective of consumers (Li et al., 2023).

Thirdly, the study can provide a reference for product design and advertising design of enterprises. When constructing set of products from different categories, businessmen attempt to increase the complementarity between products and enhance their information processing fluency to reduce cognitive load. It is crucial for marketing managers to create visual content that consumers enjoy at first glance (Overgoor et al., 2022). Purpose-driven brand building is operationalised through responsible leadership to benefit consumers (Enslin et al., 2023). Hence, empirical research on what aspects of visual content can contribute to the information-processing fluency that recommendations can generate is necessary to help companies use visual content more effectively on social media.

The limitations of this research mainly lie in the following areas. Firstly, as the materials chosen for study 1, 2 and 3 were protein powder and multivitamin, which are both health products, there are fewer varieties and limitations in the range of product category choices. In the future, the scope of experimental products may be expanded beyond personal care products to further explore a variety of products in other categories. Secondly, this research only focusses on one-time purchases. A future research avenue worth pursuing is to investigate how the relationship between product recommendation types and primary goals, as well as the significance of initial purchase intentions versus cross-buying intentions, may evolve over time within various experimental contexts. Thirdly, in this research we measured initial product purchase and cross-buying intentions, while consumers’ reactions to overall evaluation remain to be explored.