Self-designing a product with a toolkit means, in essence, solving an ill-defined problem in a creative way (Franke, Keinz, and Schreier 2008). The problem is ill-defined because in the beginning of the self-design process, users typically only have a rough idea of the desired product (Simonson 2005). Hardly ever they will command so clearly defined preferences that they are able to construct the product in all its details in their minds and just “convey” the solution to the producer via the toolkit interface (Huffman and Kahn 1998). The process is creative because the solution is usually novel for them (Jeppesen 2005, Randall, Terwiesch, and Ulrich 2005) and the motive of originating something anew has been found to be a dominant reason for starting self-design processes with toolkits (Franke, Schreier, and Kaiser, 2010). Support for this interpretation can be seen in the trial-and-error processes visible when users interact with toolkits (e.g. Franke and Hader 2014; von Hippel and Katz 2002).
From this perspective, a good toolkit is a toolkit that supports customers in their creative problem-solving efforts. However, only few studies have adopted this notion consequently. They analyzed how specific characteristics of toolkits such as starting solutions (Hildebrand, Häubl, Herrmann 2014), pre-configurations based on preference articulation (Boller, Schlager, Franke, Herrmann 2016), or peer-to-peer support by communities (Franke, Keinz, and Schreier 2008; Jeppesen 2005) lead to improved process and outcome satisfaction. A look at the many hundreds of toolkits on the web suggests that this holds true for practitioners as well. Their architecture suggests the implicit notion that customers know what they want and use toolkit just as a communication interface.
In this research project, we will extend this line of research by investigating the influence of different problem-solving styles on value creation with toolkits. Almost by definition, creative problem-solving processes are heterogeneous and involve many variants. Yet, extant research suggests that two fundamental archetypes of human problem-solving styles can be distinguished: the “holist” and the “serialist” approach (Lee et al. 2009; Ford 2000; Clewley, Chen, and Liu 2011; Ford and Chen 2001; Pask 1979). Both are seen as personality traits and thus relatively time-stable and robust against situational influences (Entwistle and Ramsden 2015).
• Holists are “top-down” problem-solvers who first focus on the big picture and then work on details. An example is architects or fashion designers who start their problem-solving with rough paper-and-pencil sketches they redo and refine, and refrain from details until they are satisfied with the big picture.
• Serialists are “bottom-up” problem-solvers who work on details, while the overall solution is emerging step by step. An illustration is reading an encyclopedia from front to back.
Our first research question is whether this dichotomy enhances our understanding in the area of self-design with toolkits for user innovation and design. This would be the case if the users’ different self-design processes can be grouped into these two archetypes in a meaningful way. In order to analyze this, we plan the following:
• A think-aloud study in which we ask users to comment on their behavior while self-designing a product.
• An in-depth analysis of interaction protocols in which we quantitatively determine in how far the two cluster solution is satisfactory regarding variance explained (Franke, Reisinger, Hoppe 2009), corresponds to the pattern predicted by theory (Lee et al. 2009x), and can be matched with independent tests of the holist vs. serialist dichotomy (Clewley, Chen, and Liu 2011).
The second research question addresses the effects of matching the cognitive style and the toolkit architecture. If our thoughts are correct and indeed the holist vs. serialist dichotomy applies to users self-designing with a toolkit, then toolkits that are geared towards the two problem-solving styles should result in a higher subjective value of the outcome, the self-designed product. Analogous evidence from web-based learning (Bajraktarevic, Hall, and Fullick 2003; Ford 2000; Clewley, Chen, and Liu 2011; Ford and Chen 2001) adds plausibility to our hypotheses.
• We investigate key features of toolkits that serve the two distinct problem-solving styles (e.g. forward/ backward buttons, alphabetical indices, or an interactive picture of the product).
• We plan a series of experiments, in which we vary those toolkits characteristics, measure the individuals’ problem-solving style, and observe the resulting effects, satisfaction with the process (visible e.g. in the tendency to continue or abandon the self-design process), willingness-to-pay for the product, and willingness-to-purchase.
The relevance of this research project is obvious as numerous scholars and practitioners have called for more research on the value created in self-design processes and the determining toolkit features (Franke and Hader 2014; Franke, Keinz, and Schreier 2008; von Hippel and Katz 2002). Our approach is also new, as most extant research on toolkits has not differentiated between different user types (with exceptions such as Dellaert and Stremersch 2005 or Franke, Keinz, and Steger 2009).
By August 2016 we expect to finalize the literature review regarding problem-solving types. Moreover, we envisage to complete a script for the first line of studies, which includes the experimental set-up, specification of the constructs that need to be measured, and the operationalization of the variables.
