What is savings realisation rate (and how to measure it)?
The single FinOps metric that tells you whether your cost program is working, and a simple way to calculate it from numbers you already have.
Savings realisation rate is the percentage of the cloud savings you identified that you actually realised over a period. It is the clearest single answer to the question every cost program eventually has to face: is this working?
How to calculate it
The formula is simple: realised savings divided by identified savings, over the same window. If your tooling and team identified £40,000 a month of opportunities last quarter, and £26,000 of that actually came off the bill and stayed off, your realisation rate is 65%.
The only hard part is being honest about both numbers. Identified savings has to be the full opportunity set, not a flattering subset. Realised savings has to be measured against the actual bill after a stabilisation window, not the estimate attached to the recommendation when it was first raised.
What counts as a good rate?
There is no universal benchmark, because identified savings is partly a function of how aggressively your tooling flags opportunities. But the shape of the number tells you a lot. A realisation rate below 50% means more than half of what you find is leaking away unactioned, a process problem, not a discovery problem. Mature programs tend to push toward 70% and above, and they get there by acting faster, not by finding more.
Why it beats total identified savings
Total identified savings is the metric most dashboards lead with, and it is close to a vanity number. It rewards finding, not fixing. A team can double its identified savings simply by turning up the sensitivity on its recommendations, while the bill does not move an inch. Realisation rate cannot be gamed that way; it only goes up when work actually ships.
How to improve it
- Shorten time-to-action: the longer a recommendation sits, the more likely the workload changes and the estimate goes stale.
- Assign a single owner to every opportunity, with a remediation SLA based on its value.
- Verify every completed change against the bill, so realised savings is measured, not assumed.
- Retire stale recommendations deliberately, so your denominator reflects live opportunities, not a graveyard.
- Feed the result back: track realisation rate by team and by month, and review it where decisions get made.