It sounds great. You’ve got a tough, labour-intensive piece of work to do. “Let’s get the crowd to do it for us”, you say. Knock up an app and the job gets done. You save cash and look trendy all at once.

There are some high profile examples that show it works: Galaxy Zoo, Foldit and Wikipedia. What have you got to lose?

Well, there are also plenty of low profile examples where crowdsourcing has gone nowhere. So before you jump into that crowdsourcing strategy, ask yourself some questions:

Why should all those people work for you?

People are only going to engage with your problems if they feel good about it. As a minimum, that means having fun. Better still, you want to help them feel that they’re making a contribution to some bigger purpose. Can you offer that?

Will they enjoy working for you?

Even a bigger purpose wears off over time. For most people, that takes maybe a minute or two. After that, they have to enjoy what they’re doing for it’s own sake. So you need to invest in building an app that’s easy and fun to use. Knocking together something in someone’s spare time won’t cut it.

Will they really enjoy working for you?

If you can’t offer purpose, then you’d better offer a lot of fun. There are some pretty compelling games out there. They’re not trying to be anything other than games, so they have to be good to survive. Can you really compete?

How much work will you need to do to support them?

Recruiting and supporting a community of people is hard work. Labelling it as crowdsourcing doesn’t make it any easier. Look at Wikipedia – there’s an awful lot of work goes into resolving conflicts, arbitrating disputes, and suchlike.

How much work will you need to do to use their outputs?

Can you rely on the crowd’s outputs, or do you need to check and quality assure them before you can use them? I’ve seen situations where it took someone longer to check the crowd’s analysis than it would have taken to do the analysis themselves.

And remember that the crowd isn’t all good – some of the people in it will try to subvert the process, either because they have a problem with you, or simply for the sheer fun of it. Of course, you can build cross-checking into the app, but that complicates the development. (And you now need to recruit and support a much bigger crowd.)

Crowdsourcing works when you have the right problem and the right relationship with your audience (or the right image from which to build that relationship). Even then, you need to think carefully about the app and the support framework. The potential rewards are high, but you have to be prepared to invest time, credibility and thought to achieve them.

Are you prepared to do that work? And, if you are, might it not make more sense to simply knuckle down and do the analysis yourself?