Friday, 21 October 2011
The dangers of spurious data
I was reading an article the other day about how Pfizer now use computer tablets (as opposed to the usual ones they manufacture) to keep track of the conversations their representatives have with physicians. Not only is the partly due to the need to comply with all the legislation around what influence pharmaceutical companies can exert upon doctors, but is also about observing the patterns of interest among their clients.
So all well and good: it is not surprising that Pfizer is investing in clever analytics to carry out its business better.
But the thing that struck me was the senior manager talking about their strategy said (refreshingly I thought) that they were not overly concerned about precision. His view was that since the data is all about helping them manage the future, and the future is fairly fuzzy place, spending endless resources on getting numbers to three decimal points was... pointless.
Also this week, I saw a fascinating graph in the Financial Times which showed just how wrong the Monetary Policy Committee has been about its predictions for the consumer prices index (which stands at 5.2% - near a 20 year high).
All this got me to thinking about precision and spurious data. For me one of the places this often crops up is with 350 feedback tools - where people are told they have scored 3.6 on some competency against an average of 4.1. Naturally people want to know if this is significant or not - and as I know a little about stats - I have to say that I have no idea - as the full data is not there.
So, as a leader, how much spurious data are you forced to read - or indeed how much do you create?
(and yes I have left in the 350 degree feedback just to annoy you...!)