How crowdsourcing innovations work

As discussed before, innovation is largely about identifying a problem, finding a solution and assessing the fit between them. Exactly when and why crowdsourcing is a suitable method for problem solving is still a matter of debate. However, its usefulness depends on the type of problem, the characteristics of the solution and the crowd (Afuah and Tucci 2012). Dynamics between users, process iterations, different selection mechanisms (Vuculescu and Bergenholz 2014) and award structures (Terwiesch and Xu 2008) can further impact the ability of the crowdsourcing system to find high-quality solutions.

Where an organisation already understands the problem at hand and knows what solutions are likely to work, it is perhaps best to conduct the development project internally. When there is uncertainty about the problem (what the market needs) or solutions, then opening up the innovation process through crowdsourcing can be beneficial (Terwiesch and Xu 2008). The common approach is to use crowdsourcing to gather solution suggestions to more or less well-defined problems, but focusing on problems might be useful too. As research and development organisations are usually good at creating solutions, crowdsourced problems could be used as a starting point for a project pipeline, without challenging the professional status of employees (Cummings et al. 2012). Crowdsourcing can also give organisations access to lead users, that is, users who experience needs ahead of the main market (Marchi et al. 2011). It has been suggested that deploying crowdsourcing in business-to-business environments is more difficult than doing so in consumer contexts as there often is no direct contact with customers or end users (Simula and Vuori 2012). Finally, the nature of the problem and the means to solving it are interrelated. If the required solution is unlikely to come from a single actor, or if it requires building extensive new knowledge, it may be more useful to rely on collaborative communities or consortiums rather than crowdsourcing contests (Boudreau and Lakhani 2009).

Crowdsourcing innovations in contest format are a form of distributed or parallel idea generation. Here, idea refers to solution suggestion. Idea generation has been widely studied: a quick search on the term ‛idea generation’ brings up 1,367 search results on Scopus and 1,063 on the Web of Science Core Collection (as of 10 February 2015).

Brainstorming approaches are a common way to generate ideas, and it is no surprise that the keyword is one of the most popular within the idea generation literature, measured in

terms of the number of mentions and citations (Figure 3.3). Brainstorming has many potential benefits in the organisational context including supporting organisational memory of design solutions, impressing clients and creating income (Sutton and Hargadon 1996). However, standard group brainstorming may not be the best way to go if the quality of ideas is the main concern (Schirr 2012; Mullen et al. 1991). Instead, it has been suggested that a hybrid format, where people start by working individually and then as a group, produces the best results in terms of the number of different ideas generated and the quality of the best ideas generated (Girotra et al. 2010). In theory, crowdsourcing should therefore be a useful way to generate ideas, as it typically combines initial individual work with varying amounts of collaboration at the later stages. Indeed, it has been found, when compared to ideas generated by an internal development team that a crowdsourcing process can generate ideas that score higher in terms of novelty and customer benefit, although not in terms of feasibility (Poetz and Schreier 2012). Finding ways to improve the quality of the ideas generated in the crowdsourcing context is currently being investigated. For example, a crowd could be used to navigate large sets of analogies in search of new solutions (Yu et al. 2014).

Increased diversity among the problem solvers compared to internal sourcing could be one reason for the success of crowdsourcing (Terwiesch and Xu 2008). A large population of problem solvers includes people in the technical and social margins with different perspectives and heuristics. Such people have been shown to play an important role in successful problem solving (Jeppesen and Lakhani 2010).

With regard to the evaluation of solutions, crowdsourcing could offer a fast, cheap and effective way to generate relevance assessments (Alonso and Miszarro 2012).

Crowdsourcing has been used for idea screening in stage-gate innovation processes (Onarheim and Christensen 2012) with promising results. Algorithmic approaches to improve the quality of the judgments have also been developed (Vuurens and de Vries 2012). Despite these developments, a final verdict on the usefulness of crowdsourced evaluations has not yet been made. Although there is a significant correlation between crowd evaluations and executive choices, evaluation bias has also been identified (Onarheim and Christensen 2012). The accuracy of evaluations could perhaps be improved by using multi-criteria scales instead of simple voting schemes or prediction markets, as the scales using multiple attributes have been found to significantly outperform them (Riedl et al. 2010; Blohm et al. 2011).

Figure 3.3: The most popular keywords in idea generation literature (Web of Science 2014).

Previous research demonstrates that crowdsourcing can be used to identify problems, possibly by tapping into lead users, to generate solution suggestions, and to evaluate these ideas. Using crowdsourcing explicitly to identify problems is not currently a common approach. Instead, the problem is usually defined by the crowdsourcing organisation. To be suitable for crowdsourcing, the problem should be relatively uncertain, that is, it should not be known in advance what kind of solutions will be the best. In this way, it is possible to take advantage of the diversity of solutions that crowdsourcing often provides. Limited scope is also important, so that individual actors can provide solutions. The creation of solutions through crowdsourcing is a form of distributed or parallel idea generation. In the light of existing research, crowdsourcing should be well suited for idea generation, as the applications often resemble the so called hybrid format of brainstorming, where participants first work alone, and then as a group. Idea generation in hybrid format has been found to provide good results both in terms of the quality and quantity of ideas. The use of crowdsourcing for evaluating ideas appears to be a promising approach, although the question is still open with regard to optimal evaluation mechanisms and their accuracy

in practical contexts. In a less investigated research avenue, it has been suggested that collective intelligence might contribute to the effectiveness of crowdsourcing (Bonabeau 2009; Malone et al. 2010). Connecting large numbers of people around a shared problem-solving task might evoke phenomena that could be considered collective intelligence, a system that has qualitatively different properties than the individuals forming it (Heylighen 2013). Next we will look in more detail at what is already known about collective intelligence.