Yesterday I made the case for evidence-based management in general. Today I’d like to talk about what this means for managing innovation.

The case in favour of evidence-based management is made in Hard Facts, Dangerous Half-Truths And Total Nonsense: Profiting From Evidence-Based Managementby Jeffrey Pfeffer and Bob Sutton. They talk about a few innovation examples, and there are a few things that we know about managing innovation.

There’s not space to summarise all of the research evidence here, but these are some of the things that know:

• Innovation is risky, but it drives growth: this is what Pfeffer and Sutton say:
  

  We’ve also seen that there is no innovation without failure. Most organizational change efforts have a high failure rate—from mergers, to new product introductions, to technological change efforts. But the rub is that the only thing more dangerous than changing an organization is never changing it at all.

• Innovation works best if you manage it as a process: I think about innovation as the process of idea management:

  

  The biggest mistake I see firms make is to think that innovation is only about having great ideas. That’s part of it, but you also have to select the best ideas to pursue, make those ideas work, and then get the executed ideas to spread.

• Great new ideas are most often combinations of old ideas: here are Pfeffer and Sutton again:
  

  It sounds ironic, but even creativity is mostly sparked by old ideas. Both major creative leaps and incremental improvements come from fiddling with ideas from other places and blending them in new ways. Better ideas result when people act like “nothing is invented here” and seek new uses for others’ ideas. This holds for even the most creative companies like Apple, 3M, IDEO, Genentech, Google, Capital One, and Cirque du Soleil. Unfortunately, too many companies are plagued by the not invented here syndrome, where people insist on using homegrown ideas, especially ideas that can be ballyhooed as new and different. There are, after all, substantial rewards for pretending that the same old ideas are brand-new. Managers can impress bosses with cutting edge ideas. Consultants can sell clients unique services. Gurus can land lucrative book contracts and speaking fees by peddling the next big thing. And journalists can sell newspapers and magazines by giving readers the latest scoop.

  It’s usually smart to be cautious if someone is pushing an idea that they claim is completely novel.

There is a great example of evidence-based innovation management in Game-Changing Strategiesby Constantinos Markides.

It’s a very good book. Markides looks at business model innovation, particularly how to manage it in larger firms. One of the crucial issues in that case is this: how can you effectively manage multiple business models within one firm? This is something that larger firms usually have to come to grips with when trying business model innovation.

You have a couple of choices to make here. The first is: should the new business model be run in a separate business unit, or should it be integrated into the main business? The second is whether or not you should enter one of these configurations in a phased manner.

The strong recommendation is that you should always set up a separate unit – think about Michael Porter talking about being stuck in the middle if you try to achieve multiple objectives.

However, Markides has found that the answer isn’t so straightforward. He studied 68 firms that tried to manage dual business models. Of these, 42 created a separate business unit, while 26 did not.

And of the forty-two that created a separate unit, only ten were successful. This implied that separation on its own is not enough to ensure success – thirty-two firms (out of forty-two) created a separate unit but still failed!

The findings were similarly mixed for those following an integrated strategy. Even though conventional wisdom says this shouldn’t work, in a number of cases it was a successful approach.

What should we make of this?

In comparing the successful firms to the unsuccessful ones, this is what Markides found:

1. Separate units were more successful if they had a high degree of autonomy to make financial and operational decisions, but not when they had autonomy on strategic decisions.
2. Firms were less successful in managing separate business models if the units used different evaluation and incentive systems.
3. Running separate business units works better if the CEO of the new unit is an insider rather than an outsider.
4. Firms that allowed the new unit to develop its own culture were more successful than those that expected them to adopt the dominant culture of the parent.

This is a tricky mix. It shows that running separate business units requires a combination of autonomy and control, and the choices of what you keep control over are crucial to success. These results are not immediately obvious, which is a great benefit to this research – it provides useful insight.

This is typical of evidence-based management. It is harder to do, because there are no clear-cut answers that are always right. While it’s convenient to think that there are, one-size-fits-all solutions usually come from charlatans.

Knowing the boundaries of a great idea is very important – when will the idea work, and when won’t it? Good research can show you the circumstances in which a particular idea is more likely to be successful. We don’t know everything about managing innovation, but we do know a fair bit now.

Practicing evidence-based innovation management doesn’t guarantee success. But it improves your chances. And in a complex environment, that’s a great outcome.