1987-predictions-2

Why prediction markets can’t bring value to the fuzzy front end of innovation

Recently our VP of sales turned to me and asked, “In fact, Wim, why don’t we have prediction markets? It seems like a neat feature, and our prospects often inquire about it. Any particular reason why this isn’t on our development roadmap?”

There are several reasons, actually. And different perspectives to look at this interesting topic.

The Business Perspective

Prediction markets are hyped by smart marketers. That might sound like a bold statement, but the fact is that ideation tools with prediction markets are a pretty easy sell to senior executives. Because these people are very familiar with the concept of stock markets, they immediately understand how the same concept applies to ‘idea markets’. Companies become ideas, shares become votes, and currency becomes virtual credits. You get the point. It sounds great, so they dig it!

What those smart marketers won’t tell you is that prediction markets are expensive – they require a lot of people spending a lot of time playing the stock market instead of doing their job. But even worse, prediction markets don’t even deliver. There’s still no solid proof that these systems perform any better than idea ranking algorithms (similar to the ones used to rank search results). Beyond the hype, I doubt that this technology will stand the upcoming maturity test in innovation management systems. Their benefits – if any – just don’t justify the cost.

The Methodology Perspective

The true purpose of prediction markets is to attempt to predict the outcome, i.e. the successful market introduction of ideas and concepts. Now here’s a real problem: the fuzzy front end is a non-linear and complex system. Complex means by definition that its outcome is not predictable. Prediction markets might work well in their economical habitat which is – or used to be (?) – deterministic, reversible, and predictable. But they are not designed for an uncertain, chaotic and organic environment like the fuzzy front end of innovation.

In addition, prediction markets are very poor at predicting anything but very short-term outcomes. However, in most businesses a market introduction is at least a few years away, which is way to far for prediction markets to accurately predict the outcome. They must have sufficient information completeness to accurately predict outcomes with a reasonable degree of certainty. Therefore, they are quite useless in the fuzzy front end of innovation where there’s a high degree of uncertainty and lack of information.

At best prediction markets are disguised idea ranking systems that determine which ideas will be selected at the next stage gate. So if they can’t predict the outcome, why use them at all? We can use smart algorithms for that, at a much lower cost.

The metaphorical perspective

Innovation Managers should foster creativity as gardeners. Just as a gardener doesn’t waste time trying to predict which seeds will grow, an innovation manager shouldn’t waste his or her time (nor the community’s for that matter) trying to predict which ideas will be successful.

Instead, a gardener creates the best conditions for ANY plant to grow – providing fertile soil, water and light – and then fosters the ones that do. So should an innovation manager: create the right climate for innovation, keep a global view on the ecosystem – and the ideas wanting to emerge – and then collaborate to make it happen.