There is so much heartburn when picking and implementing some metrics. So in this article, I want to give a pretty quick overview of my approach and some ideas to start with. But before I get too far into it, I want to start with a message:
No metric is perfect.
Leading and Lagging
We need to establish what leading and lagging metrics are so that the rest of this will make sense. Leading metrics are visible before the desired change, and a lagging metric occurs afterward.
I want to bring this up because metrics with leading/lagging pairs help round out what we see and do an excellent job of preventing misinterpretation.
I’ll give a straightforward example from the development industry.
Lines of Code (LOC or kLOC) is a leading indicator of productivity. The number of features finished is a lagging indicator of LOC but a leading indicator of product completeness. Product completion is a leading indicator of revenue, but a lagging indicator of cost.
I hope that makes sense.
The point is that you’ll want leading and lagging pairs to help balance out what you see. Most groups only focus on one half of the equation and struggle with it.
I’m looking at you, velocity.
Measure Only What Informs Decisions Or Action
There are so many things we can instrument and measure, but the truth is we only should start with or measure things that we would take action on.
I’ll pick on velocity again. Velocity is visible in every tool, people talk about it all the time, it gets thrown around as a measure of productivity, and yet there is no real action to take based on what it does.
Most managers will beg teams to take action to improve their velocity, but the measure itself prompts nothing.
That’s because it’s an incomplete measure and fails the action test.
Let me give a more visceral example and talk about your bank account. You’d immediately investigate if you saw a sudden drop in the funds in your bank account.
That measure of funds in your account prompts action.
So, when picking metrics, we want to pick things that prompt us to make decisions or take action. Don’t measure things that are there just because it’s easy.
Money almost always prompts action, so if you’re looking for something to spark action, make your measures connect to money.
Start With You
Another issue that pops up almost universally regarding measures is that everyone wants to measure what other people are doing. That doesn’t usually get a lot of traction.
For example, management wants measures to hold the development teams accountable, and development teams want measures to hold product people accountable.
Nobody says, “I want measures to hold myself accountable.”
Yet, that is the easiest way to start, and it just so happens you are also an expert in the world around you.
So place measures on yourself first, use that as a catalyst for improvement, and show the results to others. This approach almost always gains traction with others who were shy about measuring things or didn’t know where to start.
The last thing I want to tackle is what to measure first. There are a ton of things that will come to mind as ideal measures. You might want to do the DORA metrics, for example. The problem will be that you don’t have enough instrumentation to pull it off, and the manual data collection is significant.
Those are bad choices to start. Pick measures that are easy to get even if they’re imperfect.
Remember, they still have to pass the action test. I bring this up because some of you might be thinking, “Oh, so velocity then.”
If you’re a developer, you can quickly get ahold of cycle time, lead time, defect rates, cost of ownership, innovation ratios, and run rates. Those measures are all things within your world.
Once you’ve started creating metrics and measures, you’ll see a lot more opportunities to add one more bit of data to expose some key decision points.
So don’t let perfect be the enemy of good here.
Start small, create leading and lagging pairs, choose metrics that prompt action, and start with you. You’d be amazed how far you can go in just a few weeks of applying these guidelines.