The $10M nobody could see

Company

Miro

Role

Senior EM · FinOps Lead

When

Sep 2022 – Feb 2025

Result

largest share of $10M+ AWS savings

In 2022 I joined Miro as FinOps Lead. The company had just lived through the fastest growth of its life — and the AWS bill had grown with it, then kept growing after the growth slowed down. That's the part that worries a board.

Here's what I found: nobody was wasting money on purpose. Every service had a reason to exist, every instance had a story. But when I asked "whose cost is this?" the answer, again and again, was "not ours." The bill was everyone's problem, which meant it was no one's.

Nobody wastes cloud money on purpose. They waste it because nobody can see it.

So the first thing I built wasn't a savings plan. It was visibility: cost broken down per team, per service, per feature — delivered to the engineers who shipped the code, the week they shipped it. Suddenly a five-figure monthly line item wasn't an abstraction in a finance deck. It was yours.

Then the levers, in order of ROI: rightsizing driven by real utilization data, not vibes. Commitment coverage — reserved instances and savings plans — managed like an investment portfolio, with targets and rebalancing. And where the unit economics demanded it, architecture redesign: the expensive service isn't a discount problem, it's a design problem.

The twist? The biggest savings didn't come from my team's work. They came from hundreds of engineers making slightly different decisions because they could finally see the number. Cost became a first-class engineering signal, like latency.

It compounded into the largest share of the $10M+ I’ve saved in AWS across roles. The full playbook — five levers, ranked, with the failure modes that kill most FinOps programmes — is in the essay.