Don’t blame the numbers for the way people use them
I’ve just been reading about the recent $185 million fine levied on Wells Fargo bank by various regulators in the US, after around 1.5 million bank accounts were opened and 565,000 credit cards applied for without the authorisation of the actual account holders. It’s a pretty shocking story - I know I’d be less than happy if money started disappearing from my account even faster than normal!
Now, is $185m actually of real consequence to the most highly capitalised bank in America? Well, that’s a different question. What really got me thinking was a comment from Jez Humble that I came across on Twitter. It went like this: ‘You get what you measure. This is entirely predictable. Senior management should have been fired, not the employees.’
Now, I’m not sure I entirely agree with Jez on this one. I’m pretty sure the people involved all knew that it was wrong to move clients’ money to fake new accounts without telling them, regardless of their motivation. I do feel there’s some truth in his point, though.
Metrics, measurement and the associated penalties or rewards clearly can drive behaviour, so there’s always the danger of setting up perverse incentives that both encourage undesirable behaviour and drive apparently ‘desirable’ numerical outcomes (at least compared with the goals that were set). The Wells Fargo farrago (sorry) seems a classic case in point.
However, this kind of outcome doesn’t necessarily mean that performance measurement is a bad thing. Without some kind of measurement it’s going to be very hard to drive continuous improvement. You need to have some way to understand whether you’re doing things ‘better’.
So in that case, what does the Wells Fargo example tell us? In my view it tells us that not only is serious thought needed when deciding what metrics should be used to drive growth, but the same - or even more - consideration should be given to the actions that will be based on the outcomes that are measured. It tells us that consideration needs to be given to how people might go about driving ‘desired’ outcomes, and that the people responsible for creating a flawed strategy using inappropriate metrics should also be held accountable (I do agree with Jez on this one, at least to a point).
It also means that sometimes - not always, but sometimes - a healthy dose of cynicism about human nature is probably a good thing, at least to help you consider all sides of the story. My guess would be that there was assumption no honest, ‘normal’ sales people would behave the way some Wells Fargo staff did…
But so what? Where’s the news? People can do strange things, and setting of good targets is a bit more complicated than all that...
Well, what it got me thinking about was whether the risk of driving bad behaviour through the use of performance measurement is worse than one of the alternatives - not looking at the data at all. And I can’t see that it is.
Having not been in the world of SAP very long, I still find it strange that so many SAP users have been content to run their SAP estates without any practical way to understand exactly where money is being spent, where effort is being wasted and where things could most easily be improved.
Luckily, it’s now possible to gain all kinds of insight into SAP development and delivery thanks to DevAnalytics, the latest addition to Basis Technologies’ DevOps toolset. It won’t guarantee that the right targets are used in future - it can’t prevent Wells Fargo-like perverse incentives - but it does provide a powerful new way to analyse current processes, helping you to see how money can be saved, time to value reduced and how those expensive SAP dev resources might best be utilised.
There’s more about DevAnalytics on our website. You might also want to check out this from my colleague Darren Thorpe on KPIs, incentives and mis-interpretation of data.