Keywords

1 Introduction

With the development of mobile technology, people could shop on various devices, e.g., tablet, PC, laptop, and smart phone. According to IBM, during 2015 Valentine’s week, the mobile devices (tablets and smart phones) contribute to 46.5 % of online traffic and the conversion rate of tablets is catching up with PCs [1]. Despite the popularity of mobile commerce, it is found that consumers tend to behave differently when shopping on different devices. For example, Han et al. [2] found that consumers bought more distinct products after adopting iPad. It was also found that the types of products people bought online also varied by devices. Approximately 50 % of PC and tablet owners purchased clothing and accessories on their devices, compared to just 28 % of mobile phone users; 49 % of PC owners purchased electronics on their PC, compared to 41 % of tablet owners, and just 21 % of mobile phone users. It motives us to investigate whether and why consumer’s preference and decision might be influenced by the device they are using.

Why do people purchase different products with different devices? Prior literature indicates that many factors could contribute to this phenomenon, e.g., information accessibility, device portability, device use context and trust in the platform [3]. Despite these factors, one potentially underestimated factor is the nature of the device and the concepts the device represents. A series of empirical studies have demonstrated that abstract concepts activated by unrelated tasks would play an important role in human judgment and decision making. For example, using words such as “support” and “share” to construct sentences can activate the concept of “cooperation”, leading people to sacrifice personal benefits for the public good subsequently [4]. Similarly, consumers are more likely to choose a high-priced option after being exposed to words invoking prestige goals in an irrelevant sentence correction task [5]. Although people could shop on both tablets and PCs and sometimes use the devices interchangeably, they use the two devices quite differently in their daily life. Tablets are mainly for entertainment purpose while PCs are more for work purpose. Therefore, a device might convey specific information, which may activate different goals and thus influence consumer judgment and decision making.

Hence, this article seeks to enhance the understanding of how the device affects consumer decision system (reliance on feelings vs. reliance on logic during decision making tasks) and product choice (hedonic products vs. utilitarian products) from a goal-activation perspective. In particular, we predict that as people typically use tablets for entertainment, tablets may activate experiential goals; in contrast, as people typically use PCs for in-depth thinking and rigorous work, PCs may activate instrumental goals. Prior literature suggests that these goals will influence which system (feelings vs. logic) would dominate in the decision making process [6]. Therefore, we further predict that on tablets, consumers are more likely to rely on feelings than those using PCs. Consequently, they are more likely to choose a hedonic option.

The remaining paper is organized as follows. First, we review literature on consumer purchasing behavior across devices, goal-activation and consumer decision system theories. Drawing on these theories, we develop our hypotheses, followed by the description of methodology and the report of data analysis results. Lastly, contributions and limitations of this study are discussed.

2 Literature Review

2.1 Consumer Behavior with Different Devices

As smartphones and tablets are increasingly growing in popularity, it becomes imperative for firms and researchers to understand how consumers browse and make decisions on various devices. In particular, tablets provide a balance between portability and navigational convenience. They become an importance channel for e-commerce. Prior literature indicates that at the macro level tablet channel acts as a substitute for the PC channel but a complement for the smartphone channel [7]. At the micro level, mobile-device consumers are found to be more likely to undertake simple decision-making tasks [8]. Moreover, they tend to browse products casually, which leads to the purchase of more impulse and diverse products [7]. However, it is unknown that whether people would make different decisions on different devices when facing the same choice set.

Many factors might contribute to difference in browsing patterns and purchasing behavior on different devices. For example, Bang et al. [9] identified usability and ubiquity as core features that distinguished the mobile channel from traditional channel. Other differences between tablets and PCs include platform safety, interaction modes (touchscreen vs. mouse), use context [3, 10, 11]. Tablets are believed to be less safe but provide better sensory experience than PCs. Despite these features, a tablet is usually regarded as a hedonic product while a PC is usually regarded as a utilitarian product. Thus, they may activate different shopping goals and mindset when consumer shop with them. As prior literature suggests that consumer decision and choice are driven by goals, in this study, we try to understand the effects of the device (tablets vs. PCs) from a goal-activation perspective. We aim to investigate whether tablets and PCs could trigger different shopping goals and subsequently influence consumer decision process and final choice.

2.2 Goal-Activation Theory

Goals are an increasingly important concept in consumer behavior. Goals can influence what information we attend to, what attribute we use to make a decision and what product category is considered [12]. The Goal-Activation Theory views goals as knowledge structures in long-term memory [4, 13]. As goals are not purely independent but interconnected with concepts in mind, exposure to any of these concepts can spontaneously trigger the goal, which in turn guides subsequent behavior [14]. For example, Zhou and Pham [15] found that evaluating financial products such as individual stocks in trading accounts would trigger a promotion focus, whereas evaluating financial products such as mutual funds in retirement accounts would trigger a prevention focus. Similarly, using words such as “memory” and “impression” to construct sentence could activate different information processing goals (memorization vs. impression formation), thus leading to different information processing pattern [16]. These and similar findings all provide support for the underlying notion that goals are cognitive structures that can be activated on exposure to associated concepts stored in memory, thus influencing subsequent behavior.

In particular, experiential goals and instrumental goals are two types of goals consumers hold when making purchase decisions [17]. Driven by experiential goals, consumers seek for experiential satisfaction, which leads to preference for luxuries, hedonic products and affect-rich products. On the contrary, driven by functional goals, consumers seek for utility, which leads to necessities, utilitarian products and affect-poor products [18]. As people usually use tablets for entertainment and use PCs or laptops for work, the stereotype of tablet is hedonic and the stereotype of PC or laptop is utilitarian. Hence, the device may unconsciously trigger different shopping goals, which would guide consumer decision process and influence their preference and choice.

2.3 Consumer Decision System

How do consumers choose whether to have a rich, creamy Häagen Dazs ice cream as a dessert or a healthy but perhaps less tasty bowl of fresh juice? It is decided either in a cognitive, logic-based way or in an affective, feeling-based way [6, 19, 20]. When people make decisions relying on cognition and logic, they tend to assess, weigh, and combine attribute information into an overall evaluative judgment. In contrast, when people make decisions relying on feelings, they tend to inspect their momentary feelings. The dominant decision system (feelings or logic) will determine consumer choice. Whether people rely on feelings or rely on logic is dependent on several factors, which could be grouped into five categories, i.e., the salience of the feelings, the representativeness of the feelings for the target, the relevance of the feelings to the judgment, the evaluative malleability of the judgment, and the level of processing intensity [21]. In particular, it has been found that consumers are more likely to rely on their feelings in making a decision when their processing resources are limited or when they have an experiential consumption goal. In contrast, consumers are more likely to engage in cognitive, logic-based decision making when they have an instrumental consumption goal [6, 22]. Therefore, as the device might trigger different goals, it would lead to different level of reliance on feelings when consumers are making decisions.

Decision system has several impacts on consumer judgement and preference. For example, prior literature suggests that compared to judgments based on the logic-based system, judgments based on the affective system tend to be rendered faster, more polarized, more holistic, more context dependent and less sensitive to numerical quantities [23, 24]. Moreover, reliance on feelings will lead consumers to an affectively-superior product and luxurious option [17]. Therefore, decision system is a helpful way to explain consumer judgment and decision.

3 Hypothesis Development

Drawing on goal-activation theory and consumer decision system theory, we propose our research model as below (Fig. 1).

Fig. 1.
figure 1

Research model

Full size image

Because of the marketing positioning of PC and tablet as well as individual device usage, people may think a tablet is more experiential and a PC is more instrumental. Flurry [25] reported that people spent more than 70 % of their time on games (32 %), social networking (more than 25 %) and other entertainment tasks (e.g., photo, watch video) (more than 15 %) when using mobile devices. Time spent on utility and productivity only took around 12 %. As our initial perceptions of objects, both conscious and unconscious, are stored in memory and are simulated or played back on subsequent encounters, we may naturally activate experiential goals when exposed to tablets. Similarly, as PCs and laptops are utility tools, people may naturally link them with work, in-depth thinking and responsibility, which further activate instrumental goals. Prior study indicates that these goals would influence consumer decision system, i.e., among consumers that have experiential motive (compared to consumers with instrumental motive), feelings will play a more important role in consumer decision making [6]. Therefore, people making decisions on tablets, who have stronger experiential motive, will be more likely to rely on feelings. When holding the experiential tablet, consumers might naturally step into a “relax” mode, which constrains their deliberate thinking. Therefore, we propose that

H1: Compared with consumers who make decisions on PCs, consumers making decisions on tablets are more likely to rely on feelings.

Furthermore, as with other cognitive structures, the extent to which object exposure evokes a goal and guides subsequent behavior appears to differ across people as a function of the strength of the goal–object association [26]. For example, Sengupta and Zhou [27] found that hedonically tempting food, such as chocolate could activate a promotion focus and influence subsequent food choice in a completely unrelated choice task. However, this effect only existed among people who are impulsive eaters. They argued that this was because only for impulsive eaters, there was an automatic link between hedonically tempting food and promotion focus. Similarly, people may hold different perception towards the digital devices. For users who also work on tablets and have fun on PCs or laptops, their perception of device stereotype is much weaker. Hence, the automatic link between the device and experiential goals is weaker. As a consequence, the device would not have much impact on their decision system. Therefore, we propose that

H2: The effect of the device on the decision system is moderated by perception of device stereotype. That is, for consumers who have stronger device stereotype beliefs, the effect would be stronger.

We further investigate how the device changes consumer preference and choice. Recent years saw a surge of interest in consumer hedonic vs. utilitarian choice. Hedonic products are ones whose consumption is primarily characterized by an affective and sensory experience of aesthetic or sensual pleasure, fantasy and fun; while utilitarian products are ones whose consumption is more cognitively driven, instrumental, and accomplishes a functional or practical task [31]. One product could have both utilitarian benefits, i.e., functional, instrumental, and practical benefits of consumption offerings, as well as hedonic benefits, i.e., aesthetic, experiential, and enjoyment-related benefits [17, 28, 29]. As we posit that tablets could activate experiential consumption goals while PCs could activate functional consumption goals, it is conceivable that consumers using tablets might have a higher chance to select a hedonic option than consumers using PCs or laptops. Therefore, we propose that

H3: Consumers making decisions on tablets are more likely to choose a hedonic option than consumers making decisions on PCs.

Despite the direct goal-activation effect, we also investigate the role of the decision system, which has always been recognized as an important antecedent in consumer decision making. Prior literature suggests that even when experiential goals are activated, consumers might not choose to indulge. For example, Kivetz and Simonson [30] found that consumers attached greater weight to the utilitarian (versus hedonic) dimension unless they believed that they had “earned” the right to indulge. Similarly, Okada [31] found that choice of hedonic option typically required much effort for justification when the “logic” system dominated. As the decision system theory suggests that relying on feelings leads to more holistic and impulsive decisions and the device would influence the dominance of the two systems (feelings vs. logic), we propose that

H4: The effect of the device on consumer preference for hedonic (vs. utilitarian) products is mediated by the decision system.

4 Methodology

4.1 Experimental Design

An experiment was conducted to test the hypotheses. To investigate the effects of the device on consumer decision system and choice, two responsive webpages were developed to present the hedonic product (Hotel A) and utilitarian product (Hotel B). Responsive web design is an approach aimed at crafting sites to provide an optimal viewing experience—easy reading and navigation with a minimum of resizing, panning, and scrolling—across a wide range of devices (from desktop computer monitors to mobile phones). Subjects were asked to browse information about the two hotels either on a PC or on a tablet and make a choice between the two.

On each page, we provided a short description about the hotel, eight images (four of them presented hotel facilities and four of them presented room facilities), textual location, facility and service information, as well as scores on general services, location, room facilities, food and drinks, activities, hotel facilities. Prior literature suggests that consumers judge the utility of a hotel based on location characteristics (e.g., near places of interest, near public transportation, near downtown) and service characteristics (e.g., hotel class, number of internal amenities) [32]. Therefore, Hotel A, which represented the hedonically-superior option, was featured as newly-built, uniquely-decorated, but far from downtown and providing limited services and room space; Hotel B, which represented the utility-superior option, was featured as less modern and aesthetic, but with good location and services and large room space.

4.2 Sample and Experimental Procedures

We recruited 80 participants from Amazon Mechanical Turk in total. Prior literature suggests that Amazon Mechanical Turk is a reliable way for data collection [33]. 40 participants were asked to complete the task on a PC or a laptop and 40 participants were asked to complete the task on a tablet. We added verification code on our experiment website, i.e., if the participants were not using the device as required to complete the task, they could not access the webpages and subsequent survey link.

The participants were instructed to browse two hotel webpages and make a choice between the two in the scenario that they would be travelling to Singapore. After they made their decision, they would be redirected to an online questionnaire. In the questionnaire, we first measured their hotel choice and how they made the decision, i.e., relying on feelings or relying on logic, adapted from Chang and Pham [34]. This was followed by a four-item scale measuring arousal [35] and a three-item scale measuring task involvement [36], captured as control variables. Lastly, we asked participants how they perceived the two hotels (utilitarian or hedonic) and how they perceived the device stereotype, i.e., to what extent they categorized PC as a utilitarian product and tablet as a hedonic product.

5 Data Analysis

We first checked whether the two hotels differ in terms of hedonic/utilitarian characteristics. Before asking the participants to evaluate the hotels, we gave definitions on hedonic product and utilitarian product. On a seven-point scale, participants indicated the extent to which they categorized Hotel A as hedonic (1 = “utilitarian,” and 7 = “hedonic”). We repeated the same question for Hotel B. The t-test showed that Hotel A was a more hedonic option compared with Hotel B (5.18 vs. 4.19, t(79) = 3.891, p < 0.05). Self-reported task involvement, arousal, device stereotype did not differ between the tablet group and PC/laptop group (all p > 0.05).

We first analyzed whether the device had an impact on the decision system. The decision system was measured on two 7-point agree-disagree items [34]: (1) I made my decision of which hotel to stay based on overall feelings towards the hotel; (2) I made my decision of which hotel to stay based on the logical balance of pros and cons of living in the hotel. Responses to these two items were combined into a composite scale in which lower scores indicated greater reliance on logic assessment and higher scores indicated greater reliance on feelings. ANOVA result shows that on tablets (mean = 4.28), participants were more likely to rely on feelings than participants who completed the task on PCs (mean = 3.73, F(1,78) = 9.354, p < 0.05). As consumer decision system is also contingent on situational emotional state, we did ANCOVA with arousal (α = 0.75) and task involvement (α = 0.847) as control variables. Result showed that after controlling for these factors, the device still had an impact on decision system (F(1,76) = 8.125, p = 0.012 < 0.05). Therefore, H1 was supported.

To further investigate whether the impact of the device on the decision system was dependent on device stereotype, we conducted regression analysis. Device stereotype was measured on two 7-point items: (1) To what extent do you think laptop/PC is Utilitarian/Hedonic? (1-utilitarian, 7-hedonic) (2) To what extent do you think tablet is Utilitarian/Hedonic? (1-utilitarian, 7-hedonic). A composite scale was calculated in which lower scores indicated stronger device stereotype perception. Results of regression showed a significant main effect of device (b = −1.49, t(76) = −3.68, p < 0.05) and interaction effect of the device and device stereotype on the decision system (b = 1.23, t(76) = 2.98, p < 0.05). No significant result of device stereotype on the decision system was detected. A median-split regression showed that when people had stronger stereotype perception, the effect of the device on the decision system would be stronger, which supported H2.

A logistic regression was performed to ascertain the effect of the device on preference for hedonic (vs. utilitarian) products. The result showed that consumers who made decisions on tablets were more likely to choose the hedonic option (p < 0.05). Therefore, H3 was also supported.

Lastly, we tested the mediation effect of the decision system. We performed 5000 bootstrap resamples using Preacher and Hayes’s [37] SPSS macro, as recommended by Zhao et al. [38], to test the indirect path (i.e., the path from the device to choice via decision system). The results showed that the device influenced the decision system (b = −0.55, p < 0.05) and the choice (b = −0.99, p = 0.03 < 0.05) and the decision system influenced the choice (b = 0.80, p < 0.05). Given that the bias-corrected 95 % confidence interval did not include zero (−1.18 to −0.06) and that the significance of the effect of the device was reduced (b = *−0.99 to b = −0.67) after including the decision system, we concluded that the decision system mediates the effect of the device on the choice.

6 Discussion

6.1 Contributions

This study offers several theoretical contributions. First, it shows the device has an impact on consumer preference. Though industry data and prior research indicate that product categories purchased from tablets and PCs are quite different, it is largely due to value of the products, trust in mobile payment, product accessibility on both channels, etc. We provide a new perspective to understand this phenomenon, i.e., a goal-activation perspective. In particular, we show that even when the above factors are kept constant, the device still has an impact on consumer choice. Second, we explain this effect with decision system theory. Our study shows that tablets activate experiential goals, which further lead to feeling-based thinking and hedonic choice; in contrast, PCs activate instrumental goals, which further lead to logic-based thinking and utilitarian choice. It enriches the literature on decision system theory by showing that decision system could not only be influenced by salience of feelings and information processing intensity, but also by contextual factors, such as the device.

Our findings are also important for practitioners. As we found that consumers concerned more about hedonic dimensions on tablets, retailer should highlight the hedonic aspects of their products or present more hedonic products on mobile channels. Our findings indicate that mobile platforms might be a better channel for hedonic product marketing. In addition, when designing for different portals, a more entertaining portal might benefit mobile users more while a more functional portal might benefit PC users more. That is, affective design would benefit the mobile channel more. For consumers, they may need to be cautious that tablets may lead to hedonic seeking, impulsivity and self-regulation failure, especially for those who often play games and use entertainment apps on tablets.

6.2 Limitations and Future Directions

There are some limitations of this study. First, we recruited our subjects from Amazon Mechanical Turk. Since we required subjects to complete the task on tablet in one condition and participants could choose whether to complete the task voluntarily, there might be a selection bias although our result showed that there was not much difference in the two groups in terms of device usage. Second, we propose that the device influences the decision system and the choice because different devices (i.e., tablets and PCs) activate different concepts represented in mind. In this study we use device stereotype to capture “device-consumption goal” link strength. However, a more rigorous way is to use IAT test [39]. In addition, this study did not consider cross-device browsing and purchasing behavior although nowadays people typically own more than one digital devices and use them interchangeably for online shopping.

As our experiment provides initial evidence that the device affects the consumer decision system, we could further investigate whether more affective interface design would benefit the mobile channel more. For example, we could investigate whether providing the same product information with different webpage designs would change consumer judgment. We expect to see that the device would moderate the effect of webpage aesthetics on product evaluation. In particular, the positive effect of webpage aesthetics on product evaluation would be more salient on tablets. In addition, as feeling-based processing and logic-based processing differ in terms of speed, consistency and regulation focus, we consider collecting objective data to investigate whether the effect of the device on the choice could be explained by these factors. We could also investigate whether on mobile channels, consumers are more novelty-seeking, risk-tolerant, and impulsive.