Normalizing (the Right Kind of) Experimentation
Last week, I wrote about the power of resets. At the end of the post, I mentioned that a reset is ideal when you can easily reverse the experiment.
Part of what I had in mind was Harvard professor Amy Edmondson’s definition of “intelligent failures,” which have four key attributes. As she writes in The Right Kind of Wrong, “[An intelligent failure] takes place in new territory; the context presents a credible opportunity to advance toward a desired goal (whether that be scientific discovery or a new friendship); it is informed by available knowledge (one might say “hypothesis driven”); and finally the failure is as small as it can be to still provide valuable insights. [...] A bonus attribute is that the failure’s lessons are learned and used to guide next steps.”
In other words, you must know enough about the space you’re experimenting in to have a “credible opportunity” to learn. In contrast, if you take action without enough domain knowledge to anticipate the potential consequences, that’s not an experiment. That’s [messing] around, which, of course, often leads to finding out.
It’s like the difference between “let’s cancel all of the standing meetings and see whom we need to meet to get things done” and “let’s fire the whole department and see what happens.” [That second case is almost too absurd to write because no reasonable organizational leader would pursue such a policy.]
The intelligent failures framework is also about creating a cultural shift in an organization. A friend recently told me about her colleagues, “People hate the word experiment.” People dislike experiments because they conjure uncertainty and failure—things that worsen their work life and don’t help people get ahead.
Last week, I observed a leadership team outline key performance indicators for its strategic plan. One indicator related to an organization-wide goal to be more innovative. Of course, that is hard to measure. Should 10% of employees’ time be spent on innovation? Should 10% of revenue come from new products? Or maybe aim to spend 10% of expenses on innovative initiatives?
Afterward, I asked whether the innovation goal should be considered a goal for every department, team, or even individual. This would shift innovation from an abstract corporate goal to something everyone must care about.
For example, planning conversations might sound like, “I see you have 10 initiatives for the quarter. Which one is the experiment? What are we trying to learn?” Such a dynamic signals to everyone that they should be trying and intelligently failing all the time—it’s normal. In fact, it’s abnormal if you aren’t smartly pushing the envelope. When everyone experiences “failing” experiments, it reduces the power of “failure” as a cultural label to avoid.
Of course, the direct effect of a culture in which everyone plays a part in innovation is that there’s a larger stock of potential good ideas, which creates many more opportunities to find the ideas that will make the 10% innovation goal—or whatever the right benchmark—easy to achieve.