Time to fix measures the process, not the product, much like the recovery-time metrics engineering teams track for incidents. A team can find divergences quickly and still leave them open for months; this number shows that.
How to measure it
Record two dates per finding: when it was first seen and when it was resolved. Report the median, not the average, because a few very old findings skew an average. Count open findings at their age so far; otherwise the oldest problems drop out of the median, and the number looks better the worse things get. Track it as a trend over weeks.
Illustrative example.
What it tells you
- Short and stable: findings have owners, and fixes flow.
- Growing: findings are visible but nobody owns them. That's an ownership problem, not a tooling one.
- Short, but the list keeps growing: fixes are fast, but new divergences arrive faster. Look at where they come from.
Why it stays outside the health score
Mixing speed into the health score hides both signals: a fast team with many problems and a slow team with few would look the same. Keep them side by side.