Your Dashboard Probably Sucks.
Most leaders monitor subjective vanity metrics that aren’t exposing areas that need their help. If this sounds familiar, then great opportunities are being lost.
Companies have a few (more likely 50+) OKRs or a so-called BHAG. Product teams have some sort of operational metric like velocity, but those metrics don’t much tell where to help.
What’s missing here is a speedometer: how fast is the product moving toward that big goal, and when will get there? What’s missing is Feature Cycle Time.
Cycle time is how long one thing takes to move through the process from idea to customer value. Crucially, it’s not just one improvement or interface change, but a whole feature or service that’s ready to use.
It’s hard to track this, but it’s worth it. I’ve got three good reasons why.
I’ll Give You Three Reasons.
1. You’ll know the economic value of the team’s work.
Feature cycle time helps calculate how much any given feature either costs or profits the business by simply dividing the feature’s earnings or cost savings by the number of days it took to create.
So a feature that nets $2 million and takes a team 6 months to ship is only half as valuable than one that makes $1 million from 8 weeks’ effort. That kind of economic calculation can really inform future investments.
2. People will try to make it shorter
“What gets measured gets managed”, right? Pretty much, yeah.
So Captain Obvious here, reporting for duty. When people measure feature cycle time, they will try to improve that metric.
3. People will game the metric (and for once that’s a good thing)
Believe it or not, you really want people to game this one.
Think this through from a team’s perspective. “If all that matters is getting features out quickly, we’ll show them. We’ll just break up the work but still take just as long to get the whole thing out.” Oh no, anything but that!
It really is a neat trick to incentivize more iterative planning and get away from risky big bang releases. The team naturally tries cut scope, split the feature, and slim down the solution.
They’re disincentivized to gold-plate, over-engineer, and succumb to analysis paralysis. They’ll learn faster, pivot sooner, and they’ll be stronger in the end.
Now get in there.
If this isn’t reason enough to measure feature cycle time, I don’t know what is. Sure beats velocity.
Think I’m wrong? Great, speak up and let’s talk about it. I’d love to get your two cents.
