Tuesday, August 2, 2011

Just as I thought, the perils of confirmation bias

Imagine you are a business leader who has just experienced the workplace equivalent of Melbourne Football Club’s humiliating failure to perform to expectation. In the AFL club’s case, they were thrashed by 186 points by Geelong and the coach has been sacked.   In your case, you may have screwed up a key contract, botched a product release, missed the deal of a lifetime, or blown your operational budget.  Whilst many factors may have contributed to your failure, chances are that confirmation bias may have been part of the problem. 

The perils of confirmation bias
Once we have formed a view, we are prone to look for evidence that supports rather than contradicts that position.  It's a bit like looking in the mirror rather than out of the window.  You’ll find that data will be construed to support your side of the argument whilst new, unsupportive information is ignored. It is the ultimate “yes man” situation where you surround yourself with “yes” data. 

You can see this playing out with the climate change debate, where even the same source of data (for instance Professor Garnaut’s report ) has been construed by both the Federal Government and Opposition to apply to their climate change policy positions.

Weak spots for confirmation bias
Look out for confirmation bias when commissioning, structuring and then applying market research to your business idea because a key reason for product business model failures lies in this tendency to use data to support whatever position we held initially. 

Likewise be careful of confirmation bias when interpreting customer feedback as you may tend to agree with or dispute their claims based on your own views. 

Recruitment and staff appraisals are another fertile ground for confirmation bias.  If you’ve heard of the halo effect, you’ll know that we tend to ascribe a positive bias towards people who have previously been successful to the extent that we can overlook subsequent failures.  Same goes for negative biases where we may look for confirmation of incompetence, seeing every problem that person and ignore evidence of their successes.

How to counter confirmation bias? 
Two things. First, independence.  Seek out parties that are not directly invested in the outcome and use them as a sense check.  If you are within a larger organisation, this will mean involving business units from other market areas. If you have a market research team, they should be filling this role for you. Likewise HR.  If you are in a smaller organisation, use your peer network to vet your ideas.

Second, objective criteria.  If you cannot engage independent parties, or even if you do, establishing objective criteria before you launch into analysing the worth of the project is essential to counteract confirmation bias.  This is why having a pre-determined interview guide for recruitment is a good move.  But remember, it is the substantiation of the criteria that most prone to bias. What do I mean?  Say one of your criteria for a product launch is achievement of 2% growth of the mobile handset market.  How you define the market and what industry sources you choose to supply the data are both points at which confirmation bias can rear its head, so don’t trick yourself by choosing the most supportive measure.

Confirmation bias is an inherent behaviour, so understanding your own tendency to censor contradictory information is an important part of making better decisions. Mitigating this bias takes courage, discipline and awareness, but the payoff is broader thinking and through that, reduced risk of failure.  Happy bias beating.   

 (This post also appeared in http://www.smartcompany.com.au/)

No comments:

Post a Comment