Abstract
Anecdotal evidence and extant research show that consumers can prefer both user-designed and designer-designed products. However, the factors that moderate such preferences are not well understood. The authors posit power distance belief (PDB) as a moderator such that low-PDB consumers prefer user-designed to designer-designed products because they identify more with user-driven companies. In contrast, high-PDB consumers prefer designer-designed to user-designed products due to their stronger trust in designer-driven companies. Six studies examining power distance belief at both the country and individual levels provide convergent support for the proposed moderating effect of PDB and underlying mechanisms. Furthermore, the authors demonstrate that the interaction between design philosophy and PDB is more likely for low-complexity than high-complexity products.
Marketing executives have traditionally sought to create successful product designs using expert designers who are seen as a core capability for many companies (Leonard-Barton 1992). For example, expert designers have been the raison d’etre of many iconic brands in the fashion, automotive, and furniture industries. Even in education, curricula being designed by experts is deemed as a distinction (Fuchs et al. 2013). However, more recently, a growing number of companies have encouraged users to participate in the product design process using techniques such as crowdsourcing and cocreation (e.g., Hoyer et al. 2010; Von Hippel 2005). As an example, Threadless, an online apparel brand, has sold t-shirts with designs created or selected by its registered users, a community of over two million. Lego solicits product ideas from its crowdsourcing platform and has marketed 23 user-designed toy sets chosen on the basis of customer votes. Lego regards these user creators as design partners, giving them the right to make the final product approval, a share in the profits, and recognition. London-based online furniture retailer Made.com also relies on designs submitted by users and commercializes the designs that receive the most votes from the user community.
According to eYeka, an online crowdsourcing platform, 85% of the “100 Best Global Brands” have used crowdsourcing for product design with several successes reported (Roth, Pétavy, and Céré 2015). In a user-design approach, consumers participate exclusively in a specific stage of the product or service development process, which consists of three distinct stages: ideation, selection, and development (Hoyer et al. 2010). User participation can lead to new-to-the-world products or improvements on and variants of existing products. More importantly, many of these companies actively publicize the people that designed the product. It leads us to the question of whether and how this communication would benefit the firm. This article examines the impact of consumers’ perceptions—specifically regarding whether users or experts designed a product/service—on their preferences.
Previous research generally shows a user-designed approach benefits a firm’s performance and improves participating consumers’ perceptions and preferences. For a firm, a user-design approach may bring better and more timely products to market at a lower cost and risk (e.g., Cheng and Huizingh 2014; Fuchs and Schreier 2011; Hoyer et al. 2010; Menguc, Auh, and Yannopoulos 2014; Von Hippel 2005). Among consumers, participating in the design phase can result in greater enjoyment, feelings of achievement, and better-quality inferences about the product (Dellaert and Stremersch 2005; Franke and Schreier 2010; Franke, Schreier, and Kaiser 2010; Moreau, Bonney, and Herd 2011). Beyond the participating consumers and firms, there may also be positive outcomes among nonparticipating consumers—consumers who did not participate in the design process but are aware of whether the product was designed by other users versus experts. Specifically, we examine nonparticipating consumers’ preference for products designed by other users versus products designed by engineers and designers within the firm.
A review of prior empirical research on consumer preference between user- versus designer-design philosophy suggests contradictory patterns in empirical results. Some studies indicate that consumers have a higher preference for user-designed products because they believe that firms adopting a user-design philosophy are more customer-oriented (Fuchs and Schreier 2011), have a higher ability to innovate (Schreier, Fuchs, and Dahl 2012), and produce a better-quality product (Nishikawa et al. 2017). However, other studies show reversed or attenuated effects of the user-design approach on consumer preferences depending on factors such as product type (Fuchs et al. 2013; Schreier, Fuchs, and Dahl 2012), community openness (Dahl, Fuchs, and Schreier 2015), perceived similarity with user community (Dahl, Fuchs, and Schreier 2015), and consumers’ comprehension level of product information (Ratneshwar and Chaiken 1991). Theoretically, this raises the interesting issue of whether there is a theoretically driven factor that moderates nonparticipating consumers’ preference for user-designed versus designer-designed products. We posit power distance belief (PDB) as such a moderator (Song, Zhang, and Mittal 2017): “the extent to which a society and an individual accepts and views as inevitable or functional human inequality in power, wealth, or prestige” (Oyserman 2006, p. 353). Put another way, PDB is the extent to which people expect and accept that power is distributed unequally (Hofstede 2001; Zhang, Winterich, and Mittal 2010). People with low PDB prefer having a higher level of autonomy and ability to influence decisions (Brockner et al. 2001). Consequently, low-PDB consumers should prefer user-designed products more than designer-designed products because they are more likely to identify with products made by companies empowering consumers like themselves. In contrast, high-PDB consumers prefer designer-designed products over user-designed products due to their higher trust in products designed by experts. Because people with high PDB respect legitimate power and expertise, they are more likely to perceive expert designers within a company to have higher levels of competence in designing products than lay users. Also, we explore an additional boundary condition on the moderating role of PDB. When product complexity increases, the low-PDB consumers’ preference for user-designed products is attenuated because consumers’ trust in experts is more salient for high-complexity products.
We make several theoretical contributions to the literature. First, our work improves extant understanding of consumers’ relative preference for user- versus designer-designed products by examining the role of a “belief system that embodies the endorsement of economic and social inequality/equality” (Dahl, Fuchs, and Schreier 2015, p. 1987). Prior research investigating the moderators of consumers’ preference for user- versus expert-designed products has not focused on consumer-specific factors. Second, we explicate a novel psychological process—consumer trust—that explains consumers’ preference for designer-designed products among high-PDB consumers. Our results suggest there are multiple motivations underlying consumers’ preference for a user- versus designer-designed approach. Third, we augment Paharia and Swaminathan (2019) in several ways. Theoretically, we specify a very different and more proximal mediating process underlying the PDB effects. Managerially, we identify a new boundary condition in determining when the effect of PDB will more likely occur. Finally, our research advances the PDB literature by demonstrating the critical role played by PDB in consumers’ preference for user- versus designer-designed products. Furthermore, by utilizing both cultural- and individual-level PDB, we confirm its wide level of applicability in marketing research and practice (Lalwani and Forcum 2016).
Literature Review and Hypothesis Development
Objective and Psychological Consequences of a User-Design Approach
A user-design approach engages customers in the design process through a variety of web-based platforms and social media to generate and facilitate product ideas (Schreier, Fuchs, and Dahl 2012). This approach can affect both firm and consumer outcomes.
Firms utilizing a user-design approach tend to exhibit (1) improved product novelty (Poetz and Schreier 2012) and user value (Magnusson, Matthing, and Kristensson 2003) and (2) increased efficiency by reducing both the likelihood of product failure (Ogawa and Piller 2006) and time to market (Fang 2008). In a case study of a furniture company, the aggregate sales of the user-designed products were five to eight times greater than sales of the designer-designed counterparts (Nishikawa, Schreier, and Ogawa 2013). Surely, however, the positive effects of a user-design approach may differ on the basis of industry- and firm-specific factors such as a firm’s technological capability (Cui and Wu 2016), firm size, technological turbulence, and the level of technology in the industry (Chang and Taylor 2016).
Regarding consumers, early studies focused on participating consumers—those who participated in the design processes. Thus, participating consumers may reap psychological benefits and experience greater subjective value by (1) increasing the fit between their preference and the resulting product (Dellaert and Stremersch 2005), (2) enhancing enjoyment (Franke and Schreier 2010), and (3) seeing their effort reflected in product value (Moreau, Bonney, and Herd 2011). Participating consumers also show greater feelings of accomplishment (Franke, Schreier, and Kaiser 2010), psychological ownership (Fuchs, Prandelli, and Schreier 2010), and self-integration with the product (Troye and Supphellen 2012).
Relevant to this article, past research suggests some benefits for nonparticipating customers—those who use a user-designed product without participating in the design process. These studies are summarized in Table 1, and they show that nonparticipating consumers make quality inferences and form judgments about firms’ innovation ability and customer orientation on the basis of whether a product is designed by users versus designers. Nonparticipating consumers also experience feelings of empowerment. Building on this line of research, the current research aims to understand nonparticipating consumers’ preference for user- versus designer-designed products.
Literature Review on User Design.
Nonparticipating Consumers’ Preferences for User- Versus Designer-Designed Products
Research suggests that nonparticipating consumers tend to prefer user-designed products over designer-designed products. Fuchs and Schreier (2011) argue that nonparticipating consumers perceive firms with a user-design approach as being more customer oriented. Consequently, nonparticipating consumers have more favorable corporate attitudes and higher purchase intentions toward such firms’ products. Consumers also perceive firms adopting a user-design approach as having access to a larger number of diverse, unconstrained, and specific ideas from the user community (Schreier, Fuchs, and Dahl 2012). Supporting this, Nishikawa et al. (2017) found that consumers view the quality of user-designed products to be higher because they believe they are based on ideas that address their needs more effectively. Dahl, Fuchs, and Schreier (2015) showed nonparticipating consumers who observe other consumers designing the product feel more empowered by being vicariously involved in the design process, which enhances nonparticipating consumers’ identification with the product.
However, a careful reading of the literature suggests that the very opposite might also be true: some studies show a higher preference for designer-designed products among nonparticipating consumers. Such consumers are more likely to prefer designer-designed products when they believe that a high level of expertise is required to design the product (Schreier, Fuchs, and Dahl 2012). For example, when consumers’ comprehension of the product description is manipulated to be low, they evaluate a product designed by an expert (vs. an ordinary person) more favorably (Ratneshwar and Chaiken 1991). Similarly, when product complexity of unfamiliar brands is high, nonparticipating consumers prefer designer-designed brands over user-designed brands, perceiving user-designed brands as having lower innovation ability (Liljedal 2016; Schreier, Fuchs, and Dahl 2012). In the context of the luxury fashion industry, nonparticipating consumers show a lower preference for brands labeled as user-designed, viewing them as lower in quality and as failing to signal high status (Fuchs et al. 2013).
Previous research also suggests an important boundary condition of the positive effect of user design among nonparticipating consumers. When consumers feel dissimilar to participating users or when they see little or no chance to be part of the design community because firms only invite selected users, nonparticipating consumers’ preference for user-designed products weakens (Dahl, Fuchs, and Schreier 2015). Similarly, if consumers are unfamiliar with user innovation (Schreier, Fuchs, and Dahl 2012), they are less likely to prefer user-designed products. These findings suggest that nonparticipating consumers’ preference for user-designed products is not absolute. Rather, it may be contingent on specific factors.
To quantitatively examine this issue, we conducted a meta-analysis of previous studies of nonparticipating consumers’ preference for user- and designer-designed products. We examined individual effect sizes from nine studies that examined this issue. 1 The meta-analysis process is fully detailed in Web Appendix A. The combined effect size of nonparticipating consumers’ preference for user- versus designer-designed products is statistically nonsignificant (r = .18, 95% CI [–.05, .41]). There is no difference in nonparticipating consumers’ preference for user- versus designer-designed products. I2 was estimated at 95.5% [93.14%, 97.61%], suggesting a high degree of heterogeneity among studies. Conceptually, this implies there are specific moderators of consumers’ preference for user- versus designer-designed products. Prior research has examined moderators pertaining to product category (i.e., complexity, luxury fashion items) and user–community relationship (i.e., perceived familiarity, community openness). Building on this research, we consider consumer-specific characteristics as moderators. Specifically, we examine the important role of consumers’ PDB.
The Moderating Role of Power Distance Belief
Researchers have widely used PDB, the extent to which an individual accepts and expects inequality in society, to explain a variety of consumer behaviors (Lalwani and Forcum 2016; Zhang, Winterich, and Mittal 2010). We posit that PDB moderates the relationship between design philosophy and consumers’ preference and behavioral intention. Our theoretical argument is based on recent research showing the effect of country- and individual-level power distance (i.e., PD and PDB, respectively) on several consumer outcomes, such as impulsive buying (Zhang, Winterich, and Mittal 2010), status consumption (Gao, Winterich, and Zhang 2016), brand choice (Lalwani and Forcum 2016), and charitable behavior (Winterich and Zhang 2014). As a construct, PDB can be measured at the country and individual level and be made temporarily accessible through priming tasks. PDB scores also exhibit large variation within a society, suggesting its utility for firms and managers (Lian, Ferris, and Brown 2012).
Theoretically, PDB may influence how individuals perceive and react to authority, such as a person with much knowledge or experience in some field whose opinion is therefore reliable. High-PDB individuals are more likely to respect, defer to, and trust people who have authority vested in them, such as supervisors (Bochner and Hesketh 1994; Kirkman et al. 2009). In marketing, Pornpitakpan and Francis (2000) find that advertisements presented by authority figures (e.g., dermatologist endorsers with source expertise) are favored more by consumers from a high-PDB culture (Thailand) than those from a low-PDB culture (Canada).
Conversely, low-PDB individuals de-emphasize authority and are more likely to prefer nonauthoritarian environments that emphasize decision making by those lacking power and authority (Hofstede 2001). Hui, Au, and Fock (2004) find that the relationship between empowerment and job satisfaction is stronger in countries with lower PDB scores. Brockner et al. (2001) observe that people’s dissatisfaction from a lack of participation in decision making is stronger in low-PDB cultures (United States) than in high-PDB cultures (China). More relevant to the current research, Chan, Yim, and Lam (2010) find that low-PDB customers are better cocreators of value, especially when firms also empower their employees.
Using this argument, we predict that the preference for products associated with firms using experts and professional designers will be higher among high-PDB consumers, who are more likely to respect and trust designers due to such designers’ expertise and authority. In product development, professional designers, who are likely to have higher expertise than product users, are perceived as more authoritative than product users (Schreier, Fuchs, and Dahl 2012). In contrast, low-PDB consumers are likely to have a stronger preference for products from firms with a user-design approach that empowers consumers to have more of a say in developing new products. Formally, we propose:
Identification as the Mediator for Low-PDB Consumers
Our theoretical argument is based on two potential mechanisms, which we test subsequently. Among low-PDB consumers, the preference for user-designed products is enhanced through identification with the firm. Consumer identification is “perception of oneness with or belongingness to an organization, where an individual defines him or herself in terms of the organization(s)” (Haumann et al. 2014, p. 81). A user-design approach empowers nonparticipating consumers by increasing their perception that they have a say in the product development process. This empowerment is critical for low-PDB consumers who prefer a higher level of autonomy and a perceived ability to influence decisions among consumers (Brockner et al. 2001). Thus, consumer empowerment through a user-design approach increases low-PDB consumers’ identification with user-driven companies. This process of identification undergirds Dahl, Fuchs, and Schreier’s (2015) “user design → perceived empowerment → identification → preference” framework that explains consumers’ preference for user-designed products. Our core argument is that this process is stronger among low-PDB customers.
In contrast, high-PDB consumers tend to regard product design as emanating from professional designers rather than consumers (Chan, Yim, and Lam 2010). Thus, empowering consumers to exert more influence on a company’s product design activities is unlikely to be important for these customers. As such, high-PDB consumers are likely to have a lower need for identification with the company. Thus, we propose that identification should not mediate the relationship between design philosophy and product preference among high-PDB consumers. Accordingly, we propose:
Trust as the Mediator for High-PDB Consumers
Among consumers with high PDB, we posit trust as a mechanism that explains their differential preference for products with a designer-design philosophy. Consumer trust is defined as consumers’ belief in a firm’s ability and competence to perform a specific task under specific circumstances (Mayer, Davis, and Schoorman 1995; Sitkin and Roth 1993). Consumer trust can potentially explain the effectiveness of different marketing strategies on consumer behavior. For instance, low-price policies increase consumers’ trust in sellers’ ability, which in turn enhances consumers’ purchase intention (White and Yuan 2012). Similarly, a firm’s investments in website design enhance consumers’ trust in the firms’ abilities, which increases consumers’ online purchase intention (Schlosser, White, and Lloyd 2006). Our core argument is consistent with prior research showing that trust engendered through competence mediates the effect of a website’s visual design on loyalty among high-PDB consumers but not among low-PDB consumers (Cyr 2008). Similarly, Williams, Whyte, and Green (1966) found that high-PDB employees are more likely to trust a supervisor with higher technical competency, whereas low-PDB employees base their evaluation of a supervisor on the supervisor’s benevolence.
When products are designed by expert designers associated with a firm, high-PDB consumers are more likely to trust the company responsible for the product design. This occurs because internal designers, presumably with high levels of design expertise, will engender more trust from high-PDB consumers due to the designers’ expertise and perceived authority. Thus, we theorize that consumer trust engendered by a designer-designed approach will mediate high-PDB consumers’ preference for designer-designed products. However, this may not be the case for low-PDB consumers, as these consumers are less likely to rely on competence or expertise as a source of consumer trust. Accordingly:
Product Complexity as a Boundary Condition
An important contribution of our research is that it extends Dahl, Fuchs, and Schreier’s (2015) identification framework and suggests consumer trust as an additional mechanism explaining high-PDB consumers’ preference for designer-designed products. In addition, we also contribute to the extant literature by identifying product complexity as a theoretically and managerially relevant boundary condition of the mediating role of trust.
Product complexity is the extent to which consumers perceive a product to be difficult to design (Schreier, Fuchs, and Dahl 2012). The process of designing a complex product requires a wide variety of skills and expert knowledge of technology, materials, and processes (Novak and Eppinger 2001). As product complexity increases, consumers are more likely to rely on experienced parties or professional designers to ensure design integrity and satisfactory product experience. For example, when evaluating complex products, the difference in consumers’ judgments of products made in highly versus newly industrialized countries becomes more pronounced (Ahmed, d’Astous, and Eljabri 2002). Similarly, online communication with experts has a stronger influence on consumers’ purchase behavior for complex products than for simple products (Adjei, Noble, and Noble 2010). The user innovation literature has shown that the positive effect of customer participation on new product performance in high-tech industries is smaller than in low-tech industries (Chang and Taylor 2016), presumably because consumers prefer user-designed products only when product complexity is low (Schreier, Fuchs, and Dahl 2012).
Considering this line of research, we propose that the joint effect of design philosophy and PDB is contingent on product complexity. When product complexity is low, we expect the same interactive effect posited earlier: low-PDB consumers would prefer user-designed products because of higher identification with companies, whereas high-PDB consumers would prefer designer-designed products due to stronger trust in designer-driven companies. In contrast, when product complexity is high, design expertise becomes more central in achieving design success. Thus, both low-PDB and high-PDB consumers would demonstrate stronger trust in products designed by the professional designers who are perceived to have more expertise. So, low-PDB consumers may place higher trust in a designer-driven company while also being more likely to identify with a user-driven company. These conflicting forces would offset each other and ultimately result in no significant preference between user-designed and designer-designed products for low-PDB consumers. This is consistent with the findings that show attenuation of a positive user-design effect when the product complexity increases in studies using consumers from low-PDB cultures (Schreier, Fuchs, and Dahl 2012). Therefore, the interaction effect between design philosophy and PDB would diminish for complex products, although high-PDB consumers’ preference for designer-designed products would remain high. Thus:
Overview of Studies
We test these hypotheses across six studies. Study 1 replicates Dahl, Fuchs, and Schreier’s (2015) t-shirt study in the United States and China and explores the underlying psychological processes. Study 2 is a field test with real purchase behavior. Study 3 provides a causal replication of Studies 1 and 2 by priming PDB. Study 4 tests product complexity as the boundary condition for the interaction between design philosophy and PDB. Finally, Studies 5a and 5b use two sets of firm data and test the association between country-level PDB and the market shares of user-designed and designer-designed products to provide generalizable evidence. The results of Studies 1–4 are summarized in Table 2.
Summary of Results.
Study 1: Cross-Country Study
We sought to test H1 by comparing consumers from high- and low-PD countries. Hofstede (2001) suggested that countries with high PD scores tend to have higher proportions of high-PDB consumers than countries with low PD scores. We expect that user- (designer-) designed products would be preferred more by consumers in a low- (high-) PD country.
Method
Participants and design
This study was a 2 (design philosophy: designer design vs. user design) × 2 (PDB: low vs. high) between-subjects design. Participants from the United States (PD index = 40) and China (PD index = 80) represent low-PDB consumers and high-PDB consumers, respectively. We recruited a total of 391 participants. The 195 U.S. participants were recruited from Amazon’s Mechanical Turk (MTurk) (age range: 20 to 73 years; Mage = 36.38 years, SD = 11.13; 48.2% were female) and the 196 Chinese participants were recruited from Wenjuanxing (www.wjx.cn), an online survey platform (age range: 22 to 65 years; Mage = 35.30 years, SD = 9.94; 51.5% were female). There are no significant differences between the U.S. participants and Chinese participants in terms of their age (p > .30) and gender (p > .50).
Procedure
We carefully followed and replicated the procedure of Study 1 in Dahl, Fuchs, and Schreier (2015) and added some new measures to their questionnaire. Participants in the two countries first read about two firms, completed a short survey about the firms, and provided measures and demographic information. When collecting the data in China, the questionnaire was first translated into Chinese and then back-translated into English by independent translators.
Manipulation of design philosophy
Participants viewed pictures of two unisex t-shirts from two firms: Firm A and Firm B (the real brand names were blinded). We told them that one firm depended exclusively on internal designers to generate new product ideas (designer-designed condition), whereas the other firm exclusively relied on new product ideas from its user community (user-designed condition). Thus, the between-subjects factor of design philosophy was whether Firm A was described as designer-driven and Firm B as user-driven, or vice versa (i.e., Firm A was user-driven and Firm B was designer-driven). The detailed stimuli and measures of Study 1 can be found in Web Appendix B.
Measures
We then asked participants to indicate product preference, identification with firms, and relative trust. Product preference was measured with three items (e.g., “I would more likely buy a t-shirt from” Firm A = 1, and Firm B = 7; α = .96; Dahl, Fuchs, and Schreier 2015). Identification measures were preceded by four items (e.g., “I feel more connected with” Firm A = 1, and Firm B = 7; α = .96; Dahl, Fuchs, and Schreier 2015). Trust was measured using the following three items: “Which company is a more reliable company?” “Which firm can be counted on to produce good t-shirts?” and “Which firm is more trustworthy?” (Firm A = 1, and Firm B = 7; α = .90; Doney and Canon 1997; Garbarino and Johnson 1999).
Participants then answered manipulation check items. Design philosophy was measured with two items (“Which firm adopts the design philosophy of designed by the user community?” “which firm adopts the design philosophy of designed by designers (R)?” Firm A = 1, and Firm B = 7; r = .83). For PDB, we used an eight-item scale (e.g., “As citizens we should put high value on conformity,” 1 = “strongly disagree,” and 7 = “strongly agree”; α = .71; Hofstede 2001). As expected, Chinese participants demonstrated a higher level of PDB than the U.S. participants (MUS_(low_PDB) = 3.56 vs. MChina_(high_PDB) = 4.03; F(1, 387) = 30.80, p < .001, ηp 2 = .07).
Participants also provided several measures of alternative explanations such as self-construal, familiarity, and product complexity, as suggested in the literature (Schreier, Fuchs, and Dahl 2012; Zhang et al. 2010). Self-construal was measured using the 24 items in Singelis (1994). We used the difference scores between independent (α = .90) and interdependent self (α = .86) as an index of self-construal. We measured familiarity with user design using two items (e.g., “How familiar are you with the design philosophy of designed by user-community?” 1 = “very unfamiliar,” and 7 = “very familiar;” r = .87). Product complexity was measured using two items (e.g., “How complex is it to design t-shirts?” 1 = “not complex at all,” and 7 = “very complex;” r = .89; Schreier, Fuchs, and Dahl 2012).
Results
Manipulation check
A 2 × 2 ANOVA on design philosophy measures revealed only a significant main effect of design philosophy (Mdesigner = 2.34, Muser = 5.87; F(1, 387) = 645.60, p < .001, ηp 2 = .63). Neither the main effect of PDB nor the PDB × design philosophy interaction was significant (ps > .10). Thus, the design philosophy manipulation was successful.
Alternative explanations
The 2 × 2 ANOVAs on familiarity and complexity with design philosophy and PDB revealed no significant effects (ps > .10). This enables us to rule out familiarity with design philosophy and perceived product complexity as alternative explanations. As for self-construal, Chinese participants indeed demonstrated a lower level of individual self-construal than U.S. participants (MUS = .35, MChina = −.77; F(1, 387) = 89.01, p < .001, ηp 2 = .19), which is consistent with results of extant research on self-construal. However, the following results remained almost the same regardless of whether we used self-construal as the covariate. The detailed results with self-construal as the covariate are reported in Web Appendix C.
Product preference
We conducted a two-way ANOVA to assess the impact of design philosophy and PDB on product preference. The main effects of both design philosophy and PDB were not statistically significant (ps > .30), but their interaction was (F(1, 387) = 16.88, p < .001, ηp 2 = .04). A planned contrast (see Figure 1) shows that low-PDB participants prefer user-designed products to designer-designed products (Mdesigner = 3.57 vs. Muser = 4.38; F(1, 387) = 8.32, p = .004, ηp 2 = .04). In contrast, high-PDB participants prefer designer-designed products to user-designed products (Mdesigner = 4.55, Muser = 3.73; F(1, 387) = 8.56, p = .004, ηp 2 = .04). Thus, H1 is fully supported.
Product preference in the United States and China (Study 1).
Mediator (identification)
A 2 × 2 ANOVA on identification showed a significant main effect of design philosophy (F(1, 387) = 23.94, p < .001, ηp 2 = .06) and a significant interactive effect between design philosophy and PDB (F(1, 387) = 10.86, p = .001, ηp 2 = .03). Low-PDB participants identified more strongly with Firm B when it was described as user-driven rather than designer-driven (Mdesigner = 3.36, Muser = 4.80; F(1, 387) = 33.44, p < .001, ηp 2 = .15). Among high-PDB participants, there was no difference in identification with Firm B when it was described as user-driven versus when it was described as designer-driven (Mdesigner = 3.85, Muser = 4.13; F(1, 387) = 1.28, p = .26).
Mediator (trust)
A 2 × 2 ANOVA on trust revealed a significant main effect of design philosophy (F(1, 387) = 23.80, p < .001, η2 = .06). More importantly, the interactive effect between design philosophy and PDB (F(1, 387) = 6.24, p = .013, η2 = .02) was statistically significant. High-PDB participants demonstrated a higher level of trust when Firm B was described as designer-driven versus when it was described as user-driven (Mdesigner = 4.62, Muser = 3.50; F(1, 387) = 27.27, p < .001, ηp 2 = .12). In contrast, low-PDB participants demonstrated a similar level of trust when Firm B was described as designer-driven rather than user-driven (Mdesigner = 4.19, Muser = 3.83; F(1, 387) = 2.83, p = .093).
Moderated mediation
H2a and H2b predict that among participants with low (high) PDB, identification (trust) would mediate the effect of design philosophy on consumers’ preference. We conducted a moderated mediation analysis using Model 8 in PROCESS with 5,000 bootstraps (Hayes 2015). An index of moderated mediation was significant for identification (95% CI [−1.14, −.27]) and trust (95% CI [−.59, −.07]). Furthermore, consistent with H2a and H2b, we find differential mediation through trust and identification. Among low-PDB participants, identification mediated the positive effect of design philosophy on product preference (95% CI [.53, 1.21]), whereas trust did not (95% CI [−.38, .03]). In contrast, among high-PDB participants, the relationship between design philosophy and product preference was mediated by trust (95% CI [−.71, −.29]) but not by identification (95% CI [−.10, .46]). In summary, H2a and H2b are supported.
Discussion
Study 1 supports our prediction that low-PDB participants prefer user-designed products to designer-designed products, and this preference is mediated by identification with the user-driven company. It further shows that high-PDB participants prefer designer-designed products to user-designed products because of their higher trust in the designer-driven company. This study provides evidence for our predicted effects from consumers in different cultures. In the next study, we test the hypothesis in real business.
Study 2: Field Experiment
Method
Participants and design
This study adopted a 2 (design philosophy: designer design vs. user design) × 2 (PDB: low vs. high) between-subjects design. Participants were 252 Chinese consumers recruited in a coffee shop located in Dalian, China (age range: 18 to 58 years; Mage = 29.12 years, SD = 7.71; 56% were female).
Procedure
Participants were randomly assigned to the four cells. First, an experimenter dressed in an employee uniform intercepted customers as they entered the coffee shop and asked them to read an advertisement about a campaign (PDB manipulation). The experimenter invited participants to sign their names on the advertisement to show their support to the campaign and provide demographic information. As appreciation for their support, participants received a coupon for a new product at the coffee shop (design philosophy manipulation). Lastly, another experimenter at the counter observed whether participants used the coupon to buy the new product.
PDB manipulation
We developed two advertisements named “Cultural Awareness Month” (see Web Appendix D; the English version is also available in Web Appendix G) to manipulate PDB. Specifically, we called on the participants to support the “Cultural Awareness Month” campaign to encourage people’s mutual understanding of different cultures. The low- (high-) PDB advertisement described the topic of the month as “the appreciation of values in an equal (hierarchical) society” and inspired people to identify with equal (hierarchical) values. To reinforce the manipulation, participants signed their names on the bottom of the advertisement to show their support for the campaign.
Design philosophy manipulation
At the beginning of 2020, the manager of the coffee shop decided to introduce a new product to the market, an Italian-style chicken toast that had never previously been sold in the store. We developed two versions of coupons for the toast (see Web Appendix D). In the designer-design (vs. user-design) condition, the new product was described as “ideated and designed by a professional designer who works for the shop [vs. consumers].” All other information on the coupon was identical in both conditions, such as the colored picture of the product, product description, price (“regularly priced at 35 RMB, for a limited time at 25 RMB”), and the expiration date of the coupons (“valid for the date of issue only”). Participants randomly received one of the two coupons.
Measures
We asked participants to provide demographic information. We also recorded their actual purchase behavior for the new product (0 = did not buy; 1 = did buy), which served as our dependent variable.
Results
Purchase behavior
We conducted a logistic regression using the actual purchase as the dependent variable. The independent variables were design philosophy, PDB, and their interaction. The main effects of design philosophy (Wald’s χ2 = 1.15, p > .20) and PDB (Wald’s χ2 = .20, p > .60) were not significant. However, the design philosophy × PDB interaction was statistically significant (β = −2.12, Wald’s χ2 = 15.67, p < .001, ηp 2 = .06). As predicted, in the low-PDB condition, participants were more likely to buy the toast when it was described as user-designed rather than designer-designed (designer-designed = 32.8%, user-designed = 50.8%; β = .74, SE = .36, Wald’s χ2 = 4.22, p = .04, ηp 2 = .03). In contrast, those in the high-PDB condition were more likely to buy the product when it was described as designer-designed rather than user-designed (designer-designed = 55.7%, user-designed = 24.2%; β = −1.37, SE = .39, Wald’s χ2 = 12.20, p < .001, ηp 2 = .10).
Discussion
By directly manipulating PDB and using a different product category, Study 2 presents causal evidence. Moreover, by examining actual purchase behavior in a real shopping setting, Study 2 enhances the external validity of our research. Managerially, the realistic manipulation of PDB and the communication of design philosophy in this study could be easily adopted by marketing managers.
Study 3: Manipulating PDB
Study 3 has two objectives. First, one might argue that the effects of Study 1 are driven by other cultural dimensions (e.g., uncertainty avoidance; see Lee, Garbarino, and Lerman 2007) or other consumers’ perceptions (e.g., perceived customer orientation, see Fuchs and Schreier 2011). We test these alternative explanations in this study. Second, Paharia and Swaminathan (2019) found that empowerment and quality explain the moderating role of PDB on user versus designer design, whereas we propose identification and trust as the underlying mediators. In this study, we examine all these mediators simultaneously to investigate their relationship.
Method
Participants and design
This study was a 2 (design philosophy: designer design vs. user design) × 2 (PDB: low vs. high) between-subjects design. Participants were 453 U.S. consumers recruited from MTurk (age range: 19 to 76 years; Mage = 40.32 years, SD = 13.03; 56.5% were female).
Procedure
Participants were randomly assigned to one of four experimental conditions. After completing the PDB manipulation, they read information about a company’s design philosophy and a product that the company had recently marketed. Then they filled out a short survey measuring the dependent variable, mediators, and other variables.
PDB manipulation
We used the sentence completion task from Zhang, Winterich, and Mittal (2010) to manipulate PDB. Participants formed sentences from sets of scrambled words. In the low- (high-) PDB condition, they completed ten sentences related to social equality (hierarchy).
Design philosophy manipulation
Following Schreier, Fuchs, and Dahl (2012), we informed participants that “Company A is a company that specializes in breakfast cereals. As with many firms nowadays, this company has an online user community.” Participants were then exposed to a color picture of “a product that has recently been marketed by the company,” which contained a box of cereal without any brand logos. In the designer-design condition, participants read “new products of Company A are regularly and exclusively designed by professional designers who work for Company A.” In the user-design condition, they were told that “new products of Company A are regularly and exclusively designed by the members of its user community.”
Measures
We measured purchase intention, identification, trust, empowerment, and quality using multi-item scales presented in randomized orders to each participant. We used five items from Schreier, Fuchs, and Dahl (2012) to measure purchase intention (e.g., “I would seriously consider purchasing products from this company”; 1 = “strongly disagree,” and 7 = “strongly agree”). We averaged the standardized item scores to form a purchase intention index (α = .96). Identification (α = .96) and trust (α = .89) were measured in the same manner as in Study 1. The empowerment (α = .93) and quality items were taken from Paharia and Swaminathan (2019).
Participants also answered a manipulation check and other measures. Design philosophy was measured using two items (e.g., “What type of design philosophy does this company adopt?” 1 = “designed by designers,” and 7 = “designed by the user-community”; r = .94). PDB was measured using three items (e.g., “For the time being, I mainly think that” 1 = “social hierarchy is important,” and 7 = “social equality is important”; α = .94). We used five items from Yoo, Donthu, and Lenartowicz (2011) to measure uncertainty avoidance (e.g., “It is important to have instructions spelled out in detail so that I always know what I’m expected to do”; 1 = “strongly disagree,” and 7 = “strongly agree”; α = .88). Six items from Dahl, Fuchs, and Schreier (2015) were used to measure customer orientation (e.g., “This company has the customers’ best interest in mind”; 1 = “strongly disagree,” and 7 = “strongly agree”; α = .93). The detailed stimuli and measures can be found in Web Appendix E.
Results
Manipulation checks
A 2 × 2 ANOVA on the design philosophy manipulation check items revealed only a significant main effect of design philosophy (Mdesigner = 3.23, Muser = 6.16; F(1, 449) = 346.36, p < .001, ηp 2 = .44). Another 2 × 2 ANOVA on PDB index revealed only a significant main effect of PDB (Mlow_PDB = 5.48, Mhigh_PDB = 5.08; F(1, 449) = 7.34, p = .007, ηp 2 = .02). There were no other significant effects (ps > .20). Thus, the design philosophy and PDB manipulations were successful.
Alternative explanations
A 2 × 2 ANOVA on uncertainty avoidance items revealed no significant effects (ps > .10). This enabled us to rule out uncertainty avoidance as an alternative explanation. Another 2 × 2 ANOVA on customer orientation revealed a significant main effect of design philosophy (Mdesigner = 4.90, Muser = 5.15; F(1, 449) = 4.52, p = .034, ηp 2 = .01), which is consistent with Fuchs and Schreier (2011). Further moderated mediation analysis indicated that customer orientation did not mediate the relationship between design philosophy and purchase intention in either high-PDB or low-PDB consumers (see Web Appendix F). Accordingly, we excluded customer orientation from further analyses.
Purchase intention
We ran a 2 × 2 ANOVA on purchase intention with design philosophy and PDB as two factors. The main effects of design philosophy (F(1, 449) = .01, p > .90) and PDB (F(1, 449) = .16, p > .60) were not significant, but the interaction between design philosophy and PDB (F(1, 449) = 15.69, p < .001, ηp 2 = .03) was statistically significant. In the low-PDB condition, purchase intention for the user-designed cereals was higher than the designer-designed cereals (Mdesigner = 4.36, Muser = 4.89; F(1, 449) = 8.05, p = .005, ηp 2 = .03). In contrast, those in the high-PDB condition had higher purchase intention for the designer-designed cereals than the user-designed cereals (Mdesigner = 4.95, Muser = 4.42; F(1, 449) = 7.65, p = .006, ηp 2 = .03). Thus, H1 is supported.
Moderated mediation
A moderated mediation analysis (Model 8 in PROCESS) using identification and trust as mediators revealed results similar to those obtained in Study 1. Thus, H2a and H2b are supported again. Another moderated mediation analysis using empowerment and quality as mediators also revealed significant effects in agreement with Paharia and Swaminathan (2019) (see Web Appendix F). We conducted a third moderated mediation analysis using identification, trust, empowerment, and quality as simultaneous mediators. An index of moderated mediation remained significant for both identification (95% CI: [−.67, −.12]) and trust (95% CI [−.48, −.10]), but empowerment (95% CI: [−.11, .04]) and quality (95% CI: [−.12, .05]) were no longer significant as moderated mediators. This suggests that the mediating effect of empowerment (quality) for low- (high-) PDB participants is mediated by identification (trust).
To test the complete mechanism, we conducted a moderated serial mediation analysis using Hayes’ (2015) PROCESS macro (Model 85; 5,000 bootstrapped samples). An index of moderated serial mediation was significant (95% CI: [−.60, −.11]) when empowerment and identification served as serial mediators (i.e., design philosophy → empowerment → identification → purchase intention; see Web Appendix F). When we reversed the serial order of empowerment and identification (i.e., design philosophy → identification → empowerment → purchase intention), the index of moderated serial mediation was not significant (95% CI: [−.10, .01]). Another moderated serial mediation analysis using quality and trust as serial mediators revealed a significant index of moderated serial mediation (95% CI: [−.59, −.16]). The index of moderated serial mediation was not significant when reversing the order of quality and trust (95% CI: [−.09, .02]) (see detailed results in Web Appendix F). The results support that, compared with empowerment and quality, identification and trust are more immediate mediators of purchase intention in low-PDB and high-PDB participants, respectively.
Discussion
Study 3 examines possible mediators and supports our theorizing that identification (trust) is a more immediate mediator accounting for low- (high-) PDB participants’ preference for user- (designer-) designed products. Next, we seek to strengthen the theoretical account and managerial relevance of our research by identifying a boundary condition of the observed effect. Specifically, we test the role of product complexity as a boundary condition on the interaction effect between design philosophy and PDB (H3). Furthermore, we increase generalizability of our findings by using two other products.
Study 4: Product Complexity
Method
Participants and design
This study adopted a 2 (design philosophy: designer design vs. user design) × 2 (PDB: low vs. high) × 2 (product complexity: low vs. high) between-subjects design. Participants were 624 U.S. consumers recruited from MTurk for a cash incentive (age range: 18 to 73 years; Mage = 33.56 years, SD = 10.04; 49.2% were female).
Procedure
Participants were randomly assigned to the eight cells and completed tasks that were similar to those in Study 3. First, they read an advertisement that manipulated their PDB. Second, they were presented with descriptions about a company specializing in making umbrellas (low complexity) or Bluetooth earphones (high complexity), which included a manipulation of design philosophy (discussed subsequently). Then, they completed items related to the dependent variables and other measures (see measure section). Last, they provided demographic information.
PDB manipulation
We used the English version of the same advertisement used in Study 2 to manipulate PDB. As in Study 2, participants read an advertisement and signed their initials on it to show their support for the campaign.
Design philosophy manipulation
We used the same manipulation from Study 3. In the designer-design (vs. user-design) condition, participants were told that “new products of Company A are regularly and exclusively designed by professional designers who work for Company A (vs. members of its user community).”
Product complexity
Following Schreier, Fuchs, and Dahl (2012), we used umbrellas (household products) to represent low-complexity and Bluetooth earphones (electronics) to represent high-complexity products. Both are functional products and their prices are similar. More importantly, firms have been relying on users for product development in both categories (Von Hippel, De Jong, and Flowers 2012). However, as shown in the pretest in Schreier, Fuchs, and Dahl (2012), these product categories have different levels of perceived complexity.
Measures
Items for purchase intention (α = .95), identification (α = .96), trust (α = .89), design philosophy (r = .93), and PDB (α = .96) were identical to those used in Study 3. We measured product complexity (r = .89) using the same items as in Study 1. To examine alternative explanations suggested in the literature (Schreier, Fuchs, and Dahl 2012), we measured additional variables regarding the two products: product involvement, product knowledge, and product symbolic function. Detailed measures for each variable are available in Web Appendix G.
Results
Manipulation check
2 × 2 × 2 ANOVAs on design philosophy, PDB, and complexity revealed only significant main effects of design philosophy (Mdesigner = 3.18, Muser = 6.03; F(1, 616) = 418.15, p < .001, ηp 2 = .40), PDB (Mlow_PDB = 5.91, Mhigh_PDB = 4.77; F(1, 616) = 83.54, p < .001, ηp 2 = .12), and complexity (Mumbrella_(low_complexity) = 3.97, MBluetooth_earphone_(high_complexity) = 4.93; F(1, 616) = 60.56, p < .001, ηp 2 = .09). There were no other significant effects. Thus, the manipulation of all three factors was successful.
Furthermore, we ran a 2 × 2 × 2 ANOVA on each measure of product involvement, product knowledge, and product symbolic function and found no significant differences across conditions (ps > .10). This helps rule out these factors as alternative explanations.
Purchase intention
A 2 × 2 × 2 ANOVA on purchase intention revealed a significant two-way interaction between design philosophy and PDB (F(1, 616) = 18.18, p < .001, ηp 2 = .03), replicating the results of earlier studies. The interaction between design philosophy and product complexity was also significant (F(1, 616) = 8.07, p = .005, ηp 2 = .01). Most importantly, a design philosophy × PDB × product complexity interaction was statically significant (F(1, 616) = 4.68, p = .031, ηp 2 = .01) (see Figure 2).
Effect of design philosophy, PDB, and product complexity on purchase intention (Study 4).
In the low-complexity condition (i.e., umbrellas), the interaction between design philosophy and PDB was significant (F(1, 616) = 21.14, p < .001, ηp 2 = .06), replicating the results of previous studies. Participants in the low-PDB condition preferred the user-designed umbrella to the designer-designed umbrella (Mdesigner = 4.63, Muser = 5.37; F(1, 616) = 13.23, p = .001, ηp 2 = .08). In contrast, participants in the high-PDB condition preferred the designer-designed umbrella to the user-designed umbrella (Mdesigner = 5.34, Muser = 4.71; F(1, 616) = 8.38, p = .004, ηp 2 = .05).
In the high-complexity condition (i.e., Bluetooth earphones), the interaction between design philosophy and PDB was not significant (F(1, 616) = 2.16, p > .10). Participants in the low-PDB condition did not differ in their purchase intention for the user-designed and designer-designed Bluetooth earphones (Mdesigner = 5.16, Muser = 4.83; F(1, 616) = 1.98, p > .10). Participants in the high-PDB condition still preferred the designer-designed earphones to the user-designed earphones (Mdesigner = 5.48, Muser = 4.71; F(1, 616) = 15.31, p < .001, ηp 2 = .09). These results support H3. We also conducted mediation analyses in which the results supported H2a and H2b (please see Web Appendix H for details).
Discussion
In Study 4, we tested the role of product complexity as the boundary condition for the interaction between design philosophy and PDB, and it replicates our earlier findings in the low-complexity condition: PDB is no longer a significant moderator when product complexity increases. The attenuated interactive effect between design philosophy and PDB in the high-complexity condition is due to the nonsignificant difference between low-PDB consumers’ purchase intention for user-designed and designer-designed products. As product complexity increases, low-PDB consumers demonstrate stronger trust toward the designer-driven company. At the same time, low-PDB consumers still identify more with the user-driven company. Identification (positive indirect effect) and trust (negative indirect effect) offset each other and lead to a nonsignificant difference between user-designed and designer-designed products for low-PDB consumers, thereby attenuating the interaction effect between design philosophy and PDB. Taken together, these results indicate that the moderating effect of PDB on the relationship between design philosophy and consumers’ preference is more robust when product complexity is low rather than high. In the next two studies, we test the moderating effect of PDB on preference for user- versus designer-designed products using country-level data.
Study 5a: Country-Level Study: Software
Study 5a tests H1 with regard to the association between country-level PD scores and the market shares of user-designed and designer-designed software products. Hofstede (2001) suggested that countries with high PD scores tend to have higher proportions of high-PDB consumers than countries with low PD scores. In accordance with H1, we expect that high-PDB consumers are more likely to purchase designer-designed products. As such, the market share of designer-designed products would be greater in higher-PD countries, whereas the market share of user-designed products would be greater in lower-PD countries.
Measures
Product preferences
Apache, an open-source web server software developed by its user community, is a frequently cited example of a user-designed product (Pitt et al. 2006). Microsoft IIS is another web server software developed by professional Microsoft developers. For this study, we used Apache’s country-level market share to operationalize consumers’ relative preference for user-designed products in different countries. We used Microsoft’s market shares in different countries to represent consumers’ preference for designer-designed products. We obtained the market shares of these products for the final week of 2015 from Datanyze (www.datanyze.com), which tracks over 40 million websites worldwide to provide weekly cross-sectional market share numbers.
PDB and other cultural dimensions
Consistent with Gao, Zhang, and Mittal (2017) and Winterich and Zhang (2014), we used Hofstede’s PD scores (geert-hofstede.com) as the independent variable. Country-level scores for individualism, masculinity, uncertainty avoidance, long-term orientation, and indulgence served as controls.
Control variables
We measured how long Apache and Microsoft have been available in each country, as this can influence the market shares of the two products. We got the entry time data in each country from Netcraft (www.netcraft.com) to calculate the length of market availability until the end of 2015. We used per capita gross domestic product (GDP) and average years of school education for each country (data.worldbank.org) to control for economic development and educational attainment within each country. Both factors could potentially affect consumer preferences for software. In addition, we captured the relative development of the internet in each country using three measures: (1) the percentage of internet users within the country to measure internet adoption, (2) international internet bandwidth to represent the openness of the internet, and (3) fixed broadband internet tariffs to represent the cost of surfing the internet. We collected these three measures—the percentage of internet users, international internet bandwidth, and fixed broadband internet tariffs—from the International Telecommunication Union (ITU World Telecommunication/ICT Indicators Database 2015, December 2015 edition). We also included the internet and telephony sectors’ competition index to measure the degree of liberalization of Internet services, international long-distance services, and mobile telephone services. This measure uses a 0-to-2 (best) scale. We acquired these measures of the internet sectors’ competition index from the Global Information Technology Report (Baller, Dutta, and Lanvin 2016).
Results
After matching the data from the different sources, we had 75 countries with market shares of the two software brands, all six of Hofstede’s cultural orientation scores, and the control variables. The means, standard deviations, and correlations are reported in Table 3.
Descriptive Statistics and Correlations (Study 5a).
*p < .05.
**p < .01.
We conducted a multivariate regression analysis for the two dependent variables—market share of Apache and market share of Microsoft—as shown in Table 4. In both models, all variance inflation factors (VIFs) were less than 5, indicating multicollinearity is not a concern (Hair et al. 2006).
Association of Country PD on Market Share for User- and Designer-Designed Brands (Study 5a).
*p < .05.
**p < .01.
Note: Standard errors are reported in parentheses.
Model A in Table 4 shows the results for Apache, a user-designed product. PD has a statistically significant negative effect on the market share of Apache (b = −.40, t (61) = −2.05, p = .046). Thus, Apache’s market share was smaller in countries with higher PD than in countries with lower PD. All other cultural dimensions and the control variables were statistically nonsignificant (ps > .098).
Model B in Table 4 shows the results for Microsoft, a designer-designed product. PD has a statistically significant and positive association with market share of designer-designed products (b = .42, t (61) = 2.17, p = .034). All other control variables are statistically nonsignificant (ps > .067). Taken together, the results support H1 such that higher PD scores are associated with lower market share for the user-designed product, and vice versa.
Study 5b: Country-Level Study: Hotel
One might argue that the market shares of the web server software brands in Study 5a primarily rely on business-to-business sales (although market share may be based on factors other than PD). To address this concern, Study 5b tests H1 using hotel booking data, which is more reliant on end-user customers and is relatively less utilitarian than software products. Replicating the previous results with this new product category will further enhance the external validity of our research.
Measures
Product preferences
Airbnb is one of the most successful crowdsourcing companies in the hospitality industry. Its online platform enables consumers to find, book, and pay for accommodations owned and designed by end users. Airbnb accommodations designed by individual hosts provide a unique product mix in terms of amenities and user experiences. Moreover, Airbnb users are aware of its design philosophy because the site shows photos and personalized profiles of the host for each property. In contrast, Booking.com, one of the largest e-commerce travel companies, connects consumers with professionally run hotels. Thus, Airbnb represents a brand with a user-design approach, whereas Booking.com is a brand with a designer-design approach. The preference for each brand can be measured using its relative market share in each country. Despite its limitations as a measure of consumer preference, the market share measure has the benefit of being consistent across countries. We obtained the market share of each brand in October 2018 from SimilarWeb (www.similarweb.com). This source captures all the traffic on Airbnb and Booking.com (as well as other hotel booking websites) from millions of IP addresses around the world and it calculates the market shares of these websites in a certain country according to the IP addresses. That is, the market shares of Airbnb and Booking.com are based on traffic from domestic IP addresses. According to Tourism Highlights (UNWTO 2018), most (95%) international travelers booked their hotels in advance in their homelands during this year. Thus, the market share mostly reflects reservations made by domestic individuals.
PDB and control variables
As in Study 5a, we used Hofstede’s PD scores as the independent variable. We have several control variables. First, we included Hofstede’s country scores for individualism, masculinity, uncertainty avoidance, long-term orientation, and indulgence (geert-hofstede.com). Furthermore, we measured how long Airbnb and Booking.com have been available in each country’s market from their market entry to October 2018 (www.statista.com). We also controlled for per capita GDP, average years of school education, and the percentage of internet users in each country. We included other industry-specific factors as well. We controlled the direct contribution of travel and tourism to GDP (data.worldbank.org), which represents the scale and prevalence of the tourism industry in each country. Because price is one of the factors for consumers when choosing between Airbnb and Booking.com, we included a price comparison index (average hotel price divided by average apartment rent price) in the regression model. Also, consumers may avoid booking an accommodation through Airbnb because of safety concerns (since it represents individual residences). To control for this, we used the safety index of each country as another control variable. We got the data about average hotel price, average apartment rent price, and safety index of each country from HikersBay (hikersbay.com).
Results
We obtained the market shares of Airbnb and Booking.com in 59 countries using SimilarWeb. After excluding observations with missing values, we retained a final sample of 50 countries. The means, standard deviations, and correlations are reported in Table 5 and the main results are shown in Table 6.
Descriptive Statistics and Correlations (Study 5b).
*p < .05.
**p < .01.
Association of Country PD on Market Share for User- and Designer-Designed Brands (Study 5b).
*p < .05.
**p < .01.
Note: Standard errors are reported in parentheses.
We ran two multivariate regressions using the market shares of Airbnb and Booking.com as the two dependent variables (see Table 6). In both models, all VIFs were less than 5, indicating multicollinearity is not a concern (Hair et al. 2006). Model A in Table 6 shows the results for the market share of Airbnb. Consistent with our hypothesis, the association between higher PD and the market share of Airbnb, a user-designed product, is negative and statistically significant (b = −.52, t (36) = −2.30, p = .028). All other cultural dimensions and the control variables were statistically nonsignificant (ps > .07).
Model B in Table 6 shows the results for the market share of Booking.com, a designer-designed brand. Higher PD has a statistically significant and positive association with the market share of Booking.com (b = .46, t (36) = 2.83, p = .008). We also find a positive and statistically significant association between uncertainty avoidance and the market share of Booking.com (b = .41, t (36) = 3.34, p = .002). We believe that people in countries with high uncertainty avoidance scores prefer hotels that provide more standardized accommodation and services than private accommodations. All the other control variables are statistically nonsignificant (ps > .066). Taken together, the results again support H1.
Discussion of Software and Hotel Results
Studies 5a and 5b present field evidence using a country-level dataset assembled from several secondary sources that measure PD, market share, and a variety of control variables. Acknowledging that unmeasured differences between the two brands and the different countries could drive the results, we believe they support and are consistent with the experimental findings supporting H1.
General Discussion
A burgeoning research base supports the benefits of a user-design approach in product development, but the actual market performance shows a user-driven approach is not always a silver bullet. As seen in the bankruptcy of Quirky, a pioneering crowdsourcing platform, and the disappointing performance of Threadless in China (Li 2014), a designer-driven or user-driven approach can be either successful or unsuccessful. Working with the idea that one design philosophy is not inherently superior to another for enhancing consumer preferences, this research suggests that consumers’ preference for user-designed versus designer-designed products may depend on their PDB. In six different studies we show that low-PDB consumers prefer user-designed products to designer-designed products because they identify more with user-driven companies. In contrast, high-PDB consumers prefer designer-designed products to user-designed products due to their stronger trust in designer-driven companies. Furthermore, the interaction between design philosophy and PDB on consumers’ preferences is stronger for low-complexity than for high-complexity products.
Theoretical Contributions
This research extends previous findings with respect to user-design approaches. First, we find that PDB moderates the effect of design philosophy on product preference. For consumers with low PDB, we replicate the positive effect of user design, which is consistent with previous research conducted in Western cultures (Dahl, Fuchs, and Schreier 2015). However, user design may be ineffective among consumers with high PDB, as these consumers indicate a stronger preference for designer-designed products. Prior research discusses the moderating roles of status relevance of the products and brand familiarity to explain these mixed findings (Fuchs et al. 2013; Liljedal 2016). Advancing their work, we propose and test another important theoretical boundary condition: PDB.
Second, we identify different underlying processes for consumers with low and high PDB. A key contribution of the current research is to show two different psychological routes—identification and trust—and articulate the conditions under which each is a mediator evaluating user-designed and designer-designed products. We show that, depending on their PDB, consumers are likely to prefer (1) user-designed products because they have a stronger identification with user-driven companies or (2) designer-designed products because they have more trust in designer-driven companies. This finding extends the existing literature on user design by demonstrating that trust is essential in explaining high-PDB consumers’ preference for user- versus designer-designed products. In addition, we supplement Paharia and Swaminathan’s (2019) findings by demonstrating the complete mediation mechanism of empowerment → identification → purchase intention in the low PDB condition and quality → trust → purchase intention in the high PDB condition. Thus, at least theoretically, the user-designed philosophy may not positively affect low-PDB consumers’ preferences if simply empowering consumers does not help them identify with the product (e.g., product categories in which consumer involvement is low). Similarly, the designer-designed philosophy may not influence high-PDB consumers’ purchase decisions greatly if quality fails to affect consumer trust (e.g., cases in which products could have a potential production issue or supply instability).
Managerial Insights and Practical Implications
First, a firm dealing with multicultural markets can use knowledge about consumers’ cultural background related to PDB to develop marketing strategies that advertise the firm’s design philosophy. In low-PD cultures, labeling products as “designed by users” and enhancing customer identification may represent a more efficient strategy for attracting potential customers. Consider Threadless as a practical example. With every t-shirt shipped to the United States, the company sends consumers a greeting card that reads, “You are Threadless. You make the ideas. Make great together.” In high-PD cultures, exploiting “designed by professional designers” to position and build customer trust represents a better option for marketing managers. For example, K-Boxing, a popular suit brand in China, uses the slogan “focusing on jackets for 30 years” to emphasize the expertise of its designers.
Second, even within a single country, marketing executives can segment their customers according to PDB using a survey or by priming it in advertisements. For instance, Xiaomi Smartphones, which highlights consumer participation in product development (user-designed), achieved remarkable sales among young college students (low-PDB consumers) in China (high-PD culture). Moreover, designer-designed products may be preferred even among people living in a chronically low-PD culture if their high PDB is situationally induced. Messages designed to activate high PDB (e.g., the advertisement developed in our Studies 2 and 4) should shed light on the competence and expertise of a company and subsequently lead to increased consumption of the company’s designer-designed products.
Limitations and Future Research
Our results are consistent but not without several limitations. First, in Study 1 and our empirical studies, we used country-level PD as a proxy to gauge individual-level behavior, which may not capture individual differences and situational factors within a single country, as mentioned previously. Second, besides product complexity and PDB, there might be other factors such as uncertainty avoidance and novelty seeking that influence the effectiveness of a user-design approach. Previous literature identified several moderating factors such as brand familiarity (Liljedal 2016) and status relevance of the product category (Fuchs et al. 2013). In addition to exploring additional factors, future research needs to incorporate multiple factors and examine their interactive effects. Lastly, in the experimental studies, we didn’t test the hypotheses using radical innovation or new-to-the-world products. Because the extent of technological leap in new products is likely to increase product uncertainty, we recognize the need for a more nuanced examination of mediators. For example, consumers may prefer designer-designed products due to other motivations like uncertainty avoidance rather than trust, as shown in this research. We should also explore whether this will attenuate differences between high- and low-PDB consumers. Future research should include considerations of these issues and may benefit from testing with radical innovation products.
Supplemental Material
Supplemental Material, Web_Appendix_(JMR.17.0574)_Updated - Consumers’ Preference for User-Designed Versus Designer-Designed Products: The Moderating Role of Power Distance Belief
Supplemental Material, Web_Appendix_(JMR.17.0574)_Updated for Consumers’ Preference for User-Designed Versus Designer-Designed Products: The Moderating Role of Power Distance Belief by Xiaobing Song, Jihye Jung and Yinlong Zhang in Journal of Marketing Research
Footnotes
Acknowledgments
The authors sincerely thank Dengfeng Yan and Huachao Gao for their valuable comments and suggestions. The authors also thank the JMR review team for their constructive suggestions.
Associate Editor
James Bettman
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was conducted while the first author was a visiting scholar at UTSA and supervised by the third author, and it was supported by the National Natural Science Foundation of China (Grant 71972024 and Grant 71472020), awarded to the first author.
Notes
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
