Gone are the days when firms could afford to think of innovation as an external process, in which participation was an option. As discussed in my forthcoming book Making Innovation Last (co-authored with David Gotteland and Christophe Haon of Grenoble Ecole de Management, France), companies that are to survive over the long term need to keep growing continually, which means their innovation engines need to be firing on all cylinders, at all times.

The innovation engine is fuelled by great ideas, but a fully-formed, actionable, and innovative idea rarely emerges on its own. In most cases, hundreds of ideas must be generated and vetted to get to the gold. This is a game that favours organisations with plentiful and eclectic sources to draw from - provided they are also equipped with the search and selection skills that will help them choose the best ideas faster. This is especially true when it comes to crowdsourcing, a practice more and more organisations - including Dell, IBM, Procter & Gamble, and Starbucks - are adopting.

Customers as idea providers

Crowdsourcing is based on the assumption that customers are capable of putting forth ideas to match even the experienced designers at a company renowned for its innovation such as P&G. A growing body of research backs this supposition.

A 2012 study found that user-initiated products launched through the furniture division of Japanese consumer goods company Muji, were three times more successful than designer-generated products in terms of first-year sales revenue, and four times more in terms of gross margins. Moreover, the researchers concluded that this performance gap increased over time.

However, Muji’s results carry a sizeable caveat. Although customers came up with blockbuster ideas, their hit-to-miss ratio was a lot lower. Ideas submitted by company designers numbered only about ten per theme on average, while the lowest number of crowdsourced ideas per theme exceeded 400. This presents a possible danger, as prior INSEAD research shows that companies seeking innovative ideas through crowdsourcing can get paralysed by the sheer volume of ideas that come in. Because not all customers are equally creative, companies will likely have to build selection into the basic structure of their open innovation efforts if they are to be successful.

The “innovation contest”

To select its crowdsourced furniture designs, Muji used an “innovation contest”, a crowdsourcing technique that has been growing more popular with companies recently. A 2008 study provides a handy summary of how this works: “A firm (the seeker) facing an innovation-related problem … posts this problem to a population of independent agents (the solvers) and then provides an award to the agent that provides the best solution.” Despite being associated with several major breakthroughs, innovation contests are far from uncontroversial, with some observers arguing that they serve as a crutch for risk-averse companies.

Designing the contest

How, then, should innovation contests be designed so as to achieve the most innovative result? The extant research around open innovation focuses on three principal questions companies should consider.

1) How should the problem be formulated? – When possible, a firm should formulate the problem somewhat ambiguously. One study demonstrated in an analytical model that when the firm poses a highly specific problem, participants tend to focus on the most promising solutions and not to consider all possible options. Conversely, underspecifying the problem results in a more diverse array of possible solutions to the problem.

2) How many customers should participate in the contest? – Drawing on various studies in the existing literature, we can hypothesise that as the number of participants increases, the diversity of the solutions they propose increases because of the diversity of approaches they use to solve the problem. With a higher diversity of solutions, the likelihood of finding a good solution increases. However, when there are too many participants in an innovation contest, each participant is less likely to win, and participants may not make much effort to find a good solution to the problem. The firm can reduce the risk of participants’ underinvestment by (1) giving performance-contingent awards (such as royalties or profit-sharing) instead of fixed-price awards to the winners and (2) providing participants with a greater variety of activities, demanding different skills and talents, to solve the innovation contest problem.

3) How should the winner(s) be rewarded? – Again, there are several studies that have investigated different aspects of this question. On the whole, we can take away that, as you would expect, giving high awards stimulates participants’ performance—but each award has a threshold beyond which increasing the prize value ceases to have an effect. Offering multiple awards rather than one big payoff leads to more diverse solutions when (1) the problem is not precisely specified and/or (2) the problem allows for a variety of possible approaches to finding a solution and/or (3) the firm has a significant degree of uncertainty as to whether the problem will ultimately be solved by any of the participants.

The prize for companies

In addition to generating profitable ideas, innovation contests have the potential to help companies identify and form collaborative relationships with highly valuable lead users, as these contests are well-established ways of nurturing innovation. Research shows that lead users participate in innovation contests. Further, in an innovation contest, a firm can ask the participants to supply not only new ideas or new concepts but also new prototypes, which can allow the firm to determine how effectively these participants can develop new versions of existing products.

The evidence on open innovation is not uniformly positive. We lack a clear understanding on whether and how innovation contests can replace or supplement in-house innovation processes. But the benefits of this approach may emerge more clearly once we have clarified the desired end result: not only innovative ideas but also the chance to get better acquainted with knowledgeable and innovative individuals.

Follow INSEAD Knowledge on Twitter and Facebook