The Measurement Audit
A simple exercise to take stock of what you’re tracking and focus everyone’s attention on the metrics that deserve it.
The idea with a measurement plan is that before you start tracking stuff you identify specific goals and map them to actions and metrics. And then you decide what behavior and events you need to measure.
It’s a great idea. I wonder if anyone has tried it.
In real life what tends to happen is that you start out tracking some basic things, and then you add some new features, and then you add some more tracking, and then you add some more features, and so on, and you end up somewhere—most likely without a measurement plan.
No big deal but how do we get things back on track? It may be too late for a measurement plan but it’s not too late for a measurement audit.
A measurement audit is a simple exercise that documents everything you’re currently tracking, asking important questions about why you’re tracking what you’re tracking. By casting light on what is often a dark pit of data, this exercise can identify waste and opportunities. It can also help to align focus and put metrics on a path to action.
Here’s the format I’ve been experimenting with.
- What are we tracking?
- How are we tracking it?
- Why are we tracking it?
- What are we going to do with this information?
- When are we going to do it?
- Who will do it?
The first two — what are we tracking and how — are easy. With the question of why you should be as thorough as possible. Err on the side of lots of words and try to break through to the root objective. Avoid vague or circular statements like, “We want to measure how many people are clicking this button.”
“What are we going to do with this information?” is the critical question. If you can’t define the use value of the data, if you can’t describe in plain terms how the data will be used to inform business decisions, you’re probably collecting data you don’t need.
We could stop here, having done the important work of defining what will be done with the data. But if we want to make sure that something actually gets done, we should specify when it’ll be done and who is responsible. As with the previous what question, if you can’t define when it will happen and who will do it, there’s a good chance it’s never going to get done — which means there’s a good chance the data is wasting your time.
Want to try it out? Here’s a Google Spreadsheet template to get you started.