Beware The Confirmation Bias

We like to understand "why" and we like to use the word "because".

When something is spectacularly successfully, there are no end of commentators that talk about how the success came about. In particular, the owner of the success (CEO, politician, entertainer, investor) likes to attribute that success to something they did, some skill they have or some talent they possess. So, they explain their success with stories that validate the good fortune they came to have.

Cognitive scientists call this a confirmation bias. It's vulnerability to what's known as the "corroboration error". Our tendency is to look for corroboration by finding evidence that supports our viewpoint. If we find that evidence we are satisfied that the viewpoint is upheld.

Our weakness is that we consistently fail to test for the counter argument and find evidence to support it.

One problem is that stories of succcess are more readily available than stories of failure. So we lean towards the success stories.

For every "talented" rock star out there selling a million records there are millions of equally talented musicians who did not make it. Their talent did not neccessarily cause their success, as much as they would like to think so.

Thinking about this from a startup perspective, there are many unanswered questions at the beginning of an entrepreneur's journey. She may have a vision of what the future could look like once she's solved the problem that she believes needs solving. Along the way there is much testing to be done to validate her business model.

The key thing to remember is to continuously define and running experiments using falsifiable hypotheses. Importantly though, those hypotheses should test not only for corroboration but also for the counter position.

For example, if we find a consumer segment that our product appeals to using a specific message (yay! success!), let's not just settle on that proposition, message and segment. Let's test other segments with the same message, let's test with the same segment with a different message, let's test the other segments with a different message. Just because we found something that works doesn't mean it's the best route to follow.

This does not come naturally to us. We are hardwired to corroborate. We have to force ourselves to test for alternatives.

Subscribe to

Don’t miss out on the latest issues. Sign up now to get access to the library of members-only issues.