In the literature, solutions to overcome the mass confusion phenomenon can be found within two areas: (1) the development of appropriate toolkits for customer co-design and (2) the introduction of strong customization brands. Many authors comment on the need to develop and implement dedicated toolkits for mass customization (Berger et al., 2005; Dellaert & Stremersch, 2005; Franke & Piller, 2003; Khalid & Helander, 2003; Liechty, Ramaswamy & Cohen, 2001; Novak, Hoffmann, & Yung, 2000). The idea is to prevent mass confusion by appropriate interface design and usability, representation and visualization techniques, and the restriction of choice presented to the customer. Other authors argue that most mass customizers lack a strong branding and demand the development of dedicated customization brands for signaling trust (Gummesson, 2002; Rieck, 2003). In the following, we propose a third way to overcome mass confusion: collaborative customer co-design in online communities.
Communities and customer-to-customer interaction are often mentioned in the literature as a promising way to abolish some of the hurdles of integrating customers into company activities (Schubert, 1999, 2000). One example, already used a lot in e-commerce sites, is supporting social navigation by harvesting user profile information and making this information available to other customers (Höök, Benyon, & Munro, 2003; Munro, Höök, & Benyon, 1999). Further, there is a growing body of research discussing how interaction in communities supports creative activities (Franke & Shaw, 2003; Gascó-Hernández & Torres-Coronas, 2004; Nemiro, 2001; Sawhney & Prandelli, 2000; von Hippel & Tyre, 1995). Finally, discussions with managers from Adidas and Lego indicated the potential benefits of using interactions among customers as a means to reduce mass confusion. The companies observed self-organized customer communities around the companies' products, where customers interacted with each others in the course of the elicitation process.
From Communities of Transaction to Communities for Co-Design
Today, communities are often seen in the context of virtual (online), Internet-enabled communities. Research on communities, however, can be traced back to a long time before the rise of the Internet or even the existence of any electronic communication. Communities have been a concern of many social theorists, scientists, and philosophers in the nineteenth and twentieth centuries (Hillery, 1955). In general, a community is made up of its member entities and the relationships among them. Communities tend to be identified on the basis of commonality or identification among their members. This can be a neighborhood, an occupation, a leisure pursuit, or the devotion to a brand (Mc Alexander, Schouten, & Koenig, 2002). Accordingly, Dyson (1997) defines a community “as the unit in which people live, work and play” (for a discussion of the definition see, Hagel & Armstrong, 1997; Mathwick, 2002; Porter, 2004; Preece, 2000; Schubert, 1999). Today, the old idea of communities is reincarnated in the form of virtual communities as the result of increasingly cheaper communication and interaction in a networked world. The Internet serves as an enabling technology for human interaction.
Despite the fact that there is a vast body of literature on virtual communities, there is still no consensus among researchers regarding the appropriate definition for the term (Hillery, 1955; Preece, 2000). There have been propositions for classification schemes (Armstong & Hagel, 1995; Mathwick, 2002; Schubert, 1999), but none of them has really been accepted and adopted by the scientific community. Two fundamentally different kinds of communities have been discussed in the literature: business communities and socially-oriented communities. We will focus on the first kind. Hagel and Armstrong were the most prominent authors to discuss the value of business communities. Authors generally acknowledge the potential benefits of virtual communities for business purposes (Armstrong & Hagel, 1996; Barnatt, 1998; Brown, Tilton, & Woodside, 2002; Bughin & Hagel, 2000; Hagel & Armstrong, 1997; Horrigan, 2001; Jones & Rafaeli, 1998; Rothaermel & Sugiyama, 2001; Schubert & Ginsburg, 2000; Williams & Cothrel, 2000). In the context of this article, we are dealing with virtual communities as groups of customers who are drawn to the Internet in order to perform online purchasing transactions and collaborate in the process of product purchases. We refer to them as “virtual communities of transaction” (Schubert, 1999).
These communities are supported by electronic commerce platforms that offer special community features, such as feedback, discussion, voting, ratings, etc. Electronic product catalogues often form the core of such electronic shopping environments (e-shop). Virtual communities of buyers and sellers can be merged in a single locus. This combination of an e-shop (based on an electronic product catalog) and a community platform has been termed “Participatory Electronic Product Catalogue (PEP)” (Schubert, 2000). The PEP provides a link between the product description (e.g., a book on amazon.com) and contributions from customers (e.g., a rating, review, recommendation of a particular book). Based on the aggregated customer profiles, special community features become feasible, e.g., recommendation services, personalized newsletters and alerts, chat rooms, etc. The coordination mechanisms are a necessary instrument to leverage subgroup preferences (e.g., using collaborative filtering) and to exploit the intelligence embedded in prior transaction histories and experiences. To provide such vital pathways, systems must support the notion of these virtual communities of buyers as they cultivate the process of a collective awareness.
A “community for customer co-design” extends the conception of a community of transaction even beyond a PEP by adding features related to the configuration of customized offerings. Special community features are used to support the individual or collaborative co-design process. Recall that we have defined co-design earlier as a cooperation between a firm and its individual customers during the configuration process of a customized product. The idea of communities for co-design extends this understanding. The co-design process is now conducted either jointly in a collaborative mode among several individual customers (and the firm), or is based on the collaborative input of several customers, even if the co-design process itself is conducted just between the firm and the customer. Community platforms, which support communication among people, thus can be used for collecting information about these people (to be used in automated personalization), for collecting (trusted) comments from users, and for establishing direct relationships and joint learning-processes among customers. By breaking down the barriers among users of a co-design toolkit and involving different customers in a joint interaction process, the customization process can be improved, leading to less mass confusion and, thus, a higher value of the customization offering.
Communities for co-design are similar to “user developer communities” or “communities of innovation” in new product development (e.g., the Linux developer communities; see Franke & Shah, 2003; Jeppesen & Molin, 2003; Lüthje, 2003; Prandelli, Verona, & Sawhney, 2004; Sawhney & Prandelli, 2000; Wikström, 1996), but differ from those in two aspects. First, in communities for co-design, almost all customers can be members of the community instead of just some lead users as in the case of innovation communities. Second, communities of innovation address the creation of a new solution space, and not the utilization of an existing solution for the purpose of configuration (of a customized product). As a result, in communities for co-design the scope of the collaborative design tasks is geared to the creation of trust, sharing experiences, and is often fostering aesthetic creativity instead of the joint solving of technical problems.
The idea of communities for co-design challenges an implicit assumption of many authors on personalization and customization: (Mass) customization and personalization is about offering each individual customer a customized product or service according to his or her personal needs, resulting from an interaction between the firm and the customer (e.g., Pine, 1993; Rieck, 2003; Squire, Brown, & Cousins, 2003; Wind & Rangaswamy, 2001; Zipkin, 2001). Individual needs of a customer can relate to one or more of the three generic dimensions of customization, (aesthetic) design/taste, functionality, and fit/size (Piller, 2003). In our case studies we could observe that a customization dimension is often influenced by the requirements or constraints of a group rather than that of a single person. Customization with regard to (aesthetic) design is often influenced by peers and the taste of a group rather than by the individual taste of a single person. Customers do not just follow their own “individual taste” when selecting a customized offer, but are guided by a special design which is likely to appeal to their peers. Often, consumers (especially the younger ones) are trying to copy the look of a role model. This notion was very strong in the cases of American Eagle and Swatch Via Della Spiga. Also, the original idea of customizing Lego sets (in the factory for a single consumer) proved difficult as toys are often used in role play, getting their value from the interaction possibilities between one child and her friends in a neighborhood. When playing with each other, children demand toys with matching themes.
Also, customization with regard to functionality is often defined by the needs of a group of users. Interface requirements, network effects, security standards, etc. ask for a customized solution that exactly matches the solutions of others, and not just that of a single person. This is the situation at Usertool.com, where the self-created online game has to meet specific technical requirements, especially when users play it in a shared environment. In these cases, groups of customers—communities—define or even restrict the range of customization. But communities may also provide support for users during their own customization process, as we will further explore in the following sections. Table 4 provides an overview on the community concepts we found in the six case studies of this research.
Table 4. Evidence of communities for co-design as observed by the authors in the case studies
|Case||Setting of collaborative co-design||Participants at collaborative co-design process||Mediation and facilitation of collaborative process|
|Adidas||Today: self-organized online sports communities discuss customization options of miAdidas||Customers of miAdidas products and people interested in buying them||Communities evolve around existing Internet communities provided by sports clubs and specialized sites|
| ||Planned (pilot): company-driven online platform for exchange and collaboration between users / potential buyers of miAdidas products||Customers of miAdidas products and registered users of the site; Adidas representatives (specialized product trainers)||Company initiated community hosted on miAdidas website|
|Lego||Past: communities of practice where users exchange ideas and parts; communities of co-design where users exchange new models with self created software||Adult users of Lego toys (fans)||Lego fan clubs, single users|
| ||Today (pilot): virtual design environments where users can exchange models and ideas how to use (standard) building blocks for individual models||Core user group (4-12 year old children)||Lego company (global marketing)|
|My Virtual Model||Use (export) of virtual models in online communities to discuss personal styles and ideas||Leading edge-users to user interaction||Single users; company supported, but not initiated|
|Usertool.com||Assessment of user developments by other users; online chat room to exchange design ideas||User-to-user interaction||Company initiated and facilitated|
|American Eagle||Co-design and co-production of product in workshop (offline) supplied by the company, store layout is build to foster exchange in-between customers||User-to-user interaction with support of company representatives (specialized sales clerks)||Company initiated and facilitated|
|Swatch Via Della Spiga||Co-design and co-production of product in store-based workshop (offline), store layout is built to foster exchange in-between customers||User-to-user interaction with support of company representatives (specialized sales clerks)||Company initiated and facilitated|
An Early Pilot of a Community for Collaborative Co-Design at Adidas
The third step of our research framework builds upon what we learned from the case studies and literature research, and aims at the implementation and evaluation of a pilot application for a community for customer co-design at Adidas. Due to its pioneering status, the project should cover the whole value creation process and allow for several kinds of co-design: Users shall be able to both discuss their individual footwear designs and to provide feedback to each other about possible customization possibilities, but also be able to interact in a more innovative mode on the level of idea creation and evaluation, technology assessment, and concept testing. Building on research on the design of such customer-firm interaction systems (Dahan & Hauser, 2002; Franke & Piller, 2004; Franke & Schreier, 2002; Füller et al., 2004; Nambisan, 2002; Thomke, 2003; Thomke & von Hippel, 2002; von Hippel & Katz, 2002), the research team transferred this idea into a design brief and a professional web agency programmed it. Today, the Internet-based lab allows customers of Adidas to come up with their own ideas (in a structured manner) as part of an idea generation competition and to evaluate both the ideas of other users and ideas that were generated in-house. All processes are based on typical forum applications where users can comment on an existing entry or start a new one. Several techniques are applied to enrich the users' imagination, for instance presenting different future scenarios or some vague, drafted solutions as catalysts that lower the barrier to start brainstorming new ideas. In addition, customers have the possibility to return to a previous design stage in order to modify their inputs. Thus, the virtual customer lab incorporates a “trial and error” process.
During the ongoing piloting phase (until the end of 2005), the virtual customer lab is only accessible to customers of miAdidas products in selected markets. For legal reasons, customers in a retail unit are asked to participate in the project. Customers willing to participate receive a personal access code to the project's website (http://www.miadidas-und-ich.de). On the welcome page, users are addressed personally, and, if available, a clickable picture of their last purchased shoe is displayed. Since the tool was launched at the end of 2004, 272 customers participated actively in this co-design community (from about 550 invitations). Of those, 127 customers actively participated in idea creation (46.7% of responding customers, the rest contributed to the other campaigns only). Overall, the team at Adidas was impressed by the number of contributions and the innovativeness of many co-created new product ideas.
At the moment, we are surveying the community members about their experiences with the system. Early data from these surveys show that users really liked the collaborative co-creation process and also appreciated the company's initiative to create such a community. There were almost no reports on feeling “abused” by the company. This pilot proved early assumptions that customers of mass customized products are often highly involved in the purchasing process and are very eager to provide feedback or continue an interaction after the product has been delivered (Berger & Piller, 2003). Consider that this pilot has just covered some of the potential benefits of combining the concept of virtual communities and the mass customization co-design process. Many of the applications discussed in this article have not been piloted yet (like the contribution of a community to support automatic (collaborative) filtering to get a better starting configuration).