Friday, January 11, 2013

Migrating from Beliefs to Hypotheses

Is there a difference, and does it matter?

A recent set of conversations about how various marketing campaigns were performing all included the phrase: "I believe".
  • I believe the drop in acquisitions is due to the holiday period
  • I believe the lift in sales is incremental
  • I believe the best to time to contact consumers is before or after work
Beliefs are premises that we hold be true based on personal conviction.  Challenging them can get ugly, so we often condition ourselves not to offer alternatives - particularly when dealing with HiPPOs (Highest paid person's opinion.)  In comparison, hypotheses - which can be stated with equal conviction - are meant to be challenged and proven to be correct or rejected. 

Sometimes it is easier to work under a belief system than a hypothesis-driven one for the simple reason: there is a lot more work to do.
  • What are the right questions?
  • How do we go about answering them?
  • How can we be sure?
In a post about 'The Year of Measured Retail', the folks at RSR make the point:  Lame questions lead to lame insights and that "if no one asks the right questions, no one will get the game-changing answers."  And that means being free to get creative rather than shot down. And their research suggests that the best performing retailers have an outspoken visionary leading the change.

And to be creative, we need to be able to see the shopping experience from the consumer perspective.  As evidence of this view, the Media Decoder blog of the NYT highlighted a set of M&A deals that reflect a roll up along the journey in order to provide clients with end-to-end solutions.   No more silos means lots of unknowns.

The surety issue suggests that we need the tools to reduce risk by quantifying uncertainty - and that is the domain of analysis in general and statistics in particular.   These disciplines are grounded on the notion of testing hypotheses.  Thick skins required; it won't be personal.

So, how do we transition from a belief system to a culture of testing?  Some recommendations:
  1. State what assumptions have to be true for a belief (or opinion) to be valid.
  2. Focus on confirming the assumptions, one at a time, where being proven wrong has less impact.
  3. Build to the conclusion slowly, in bite size pieces, to allow it all to sink in since the mind is the hardest thing of all to change.
In the end, try to get everyone in the habit of asking questions rather than stating opinions as facts.

No comments: