How to measure continuous improvementWe can count retrospectives, but that feels like a vanity metric. How do we know if we're actually improving?
I was recently part of a conversation discussing how to measure continuous improvement. Scrum’s retrospectives, blameless post-mortems, 20% time, training courses, hack days, and many other activities are all designed to foster continuous improvement. But how do know if they’re working?
One of the conversationalists said “We can count our retrospectives, but that feels like a vanity metric.”
I agree, it is a vanity metric.
So what isn’t?
If we take a step back, I think it becomes a bit more clear: What are we aiming to continuously improve? This is the thing we should measure to see if it improves.
This probably isn’t a single thing, and it likely changes over time. But here are some key metrics I can see measuring to see if continuous improvement is happening within a team or organization:
- Revenue or profit
- Customer satisfaction scores
- Customer retention
- Lead time (from product/feature idea to deployment)
- System performance metrics (web page load time, database query time, CPU utilization, etc)
This list could go on for pages. This should be enough to get you started.
What CI metrics have you found to be valuable?
What is your fitness function?
How do you know if your actions are leading to better or worse results?
Normally when we think of up-skilling, we think of taking a class. But what if you never use that knowledge? A knowledge option is a tool to reduce this risk.