NPS - Statistics & Smiles

I was reading through some linked in posts on a few customer insights groups and it got me thinking  about times I have sat in meeting discussing a variety of customer measurement metrics or insights, only to be interrupted by some  amateur  statistician to ask me about the statistical significance of the results.  In my experience this tends to be asked when the results are, ahem, not terribly favorable for the individual or Company.*

Indeed in one business a senior manager was instructing his teams to call up customers to ask them to make the scores higher, therefore make him look better.  I guess the best way of describing this is a sort of corporate boiler room where operators were pushing their questionable service performance.

In some quarters there seems to be a need to try and legitimize NPS through statistics which is really sad as the measure is and was never set out to be statistically significant, only significant in closing down customer issues that impact loyalty. The closest it gets to statistical significance is asking one to state what N is (how many responses).  The mistake often is to question method and ignore what one will do to address the issue - Take a leaf from Apples book and call the customers to close & resolve the customers perceived issue that drives dis tractor behavior.

So next time you have someone focusing on method rather than outcome, just remember the cartoon above and smile....


*As I majored in statistics, it is pretty easy to confuse these folks, so keeping it simple is a must