Bad KPIs, bad behavior

I received the following email about my Love/Hate Boss post:

Your recent comments about relationship between boss styles and KPIs really hit home but in a different context than you described.  My boss is currently developing KPIs for me and my team to judge our 2008 performance and he asked me for input.

When workers don’t trust the boss/company, they are more likely to develop and embrace easy measures.  Suspicious employees believe indicators are going to be used against them, so it becomes the old labor vs management model of “kick-back” measures or at least something you “game”. This is especially true when the employees don’t understand the strategy and therefore see the KPIs as “standing alone in a field”. Because if they don’t understand the strategy, how can they understand the connections? How do they understand outcomes and impact?

I am seeing this with my group and better understand, now, why communicating strategy over and over in as many ways and forms as possible is so important. If the boss is bad at communicating strategy, it doesn’t matter if the boss trained with Machiavelli or Leo Buscaglia. You should write a follow-up blog about how even a love-cat boss can be a poor leader if the strategy isn’t articulated and understood.

I agree 100%. We should be managing by objectives, not by KPIs. Objectives tell us the strategy; KPIs only tell us the score.

I observed a related phenomenon in another department here at work.  The manager was in charge of deployments of a new product line and wanted to promote the sharing of best practices between implementation. Following the old adage that compensation dictates behavior, he tied a large fraction of his employees’ salaries to the number of best practices they posted to a company-wide portal. Furthermore, writing these best practices could not interfere with ‘their day job’ so they were expected to squeeze it in between engagements or after hours.

So, what do you think happened? The employees focused on quantity instead of quality. Even if the topic would have been better explained with a little more detail, they tended to write short best practices so they could move on to the next topic. In some cases, enterprising employees broke up topics into multiple best practices so there total count would go up.

It didn’t matter to the employee whether anyone ever read the best practice document, only that it was posted on the portal. After all, this is what they were incented to do. Talk about the proverbial tree in the forest!

Instead of an output measure which tracked activities, the manager would have been better served with an outcome measure focused on impact. My recommendations would be two KPIs to balance breadth (quantity) and depth (quality):

• Number of that individual’s best practices downloaded by more than N people outside the group (breadth)
• Average number of people outside the group downloading that individual’s best practices (depth)

Compensation can indeed dictate behavior. And bad KPIs cause bad behavior.