You Can't Trust Your Workers Metrics Unless Your Workers Trust You

It's common to place a lot of faith in the power of what metrics can tell us. I've met more than one executive who romanticized about some kind of super dashboard that could pinpoint all the risks and issues, allowing them to base key decisions largely on the data they are seeing.

This could be a desirable target state, if you are doing something simple, like delivering pizzas.

In today's customer experience economy many of us are involved in something a little bit more complex. We are more likely to engage in tacit activities, activities that require analysis, learning, collaboration and on-the-fly synchronization. Chances are, we are engaged in knowledge work.

A profound truth that many fail to appreciate is that knowledge work is inherently variable. Each request for a service or product is always a little bit (or a lot) different from the last one. And you can't remove this variability from the equation. Doing so will remove the work of its value, the work will no longer be something new. Robots could do it.

This makes getting meaningful metrics somewhat problematic, comparing one unit of work to another will always require some creativity.

What this means is you can't measure knowledge work effectively unless your knowledge workers want you to.

Measuring knowledge work means measuring abstractions of the work your knowledge workers do. A trial and error exercise is required to come up with something meaningful. You need commitment to get to a stable performance baseline.

Even more so you'll need courage. Metrics for knowledge work are extremely easy to game, and being transparent about bad data is something very few organizations do well.

And it's not a one-time thing, the approach you take to measuring whatever it is you want to measure will always be little bit wrong. The context that knowledge work takes place in is always changing.


If your workers don't trust why these metrics are being used, you won't have accurate data. You won't have the insight required to know what to do with that data. And you will make bad decisions in the name of that data.

Metrics can be a good thing. But trust comes first.