Cloud FinOps is no longer just an engineering discipline. For private equity portfolio companies, it is an EBITDA recovery lever that exposes wasted cloud spend, assigns accountability, and turns recurring technology costs into governed financial performance.
Every quarterly business review at a mid-market portfolio company carries some version of the same slide: cloud spend up, revenue up, and a footnote noting that utilization “remains an area of focus.” Nobody on the call asks what that footnote has cost over the last four quarters. If someone did, the number would be hard to defend, and closing that gap is exactly what a real cloud FinOps program does.
Cloud waste has sat at 27 to 32 percent of total spend for five straight years, according to Flexera’s State of the Cloud reports. Not shrinking. Not responding to better tooling or more disciplined cost reviews. Flat. On $5 to $10 million in annual cloud spend, a typical range for a mid-market portfolio company, that works out to $1.3 million to $3.2 million a year with zero return attached to it.
The Gap Between What’s Wasted and What Gets Found
Here’s the part most cost-review decks skip: the percentage a company actually recovers is almost always lower than what it’s wasting. A first-pass review tends to catch the obvious layer, the license nobody’s using, the instance somebody forgot to shut down. It rarely reaches the layer underneath: a reserved-capacity commitment sized for a workload that shrank two renewal cycles ago, or a dev environment still billing every night because decommissioning it was never anyone’s job. That second layer is where the 27 to 32 percent figure actually lives, and it’s why a company can run a “successful” cost cleanup and still be sitting on most of the waste a year later.
Why It Survives the Deal
Acquisition should be the moment this gets fixed. A new ownership group arrives, a 100-day plan gets built, and cost review usually makes the list. In practice, cloud spend gets a glance in diligence and barely any attention during integration. Diligence checks the total against revenue and moves on. Integration is busy stabilizing the org chart, the ERP, and the go-to-market motion. Cloud infrastructure keeps running quietly in the background, mostly because it isn’t causing a visible problem.
That’s the actual issue. Waste doesn’t get flagged because nothing looks broken. Servers stay up. Applications respond. Dashboards read green. The company is simply paying for capacity, storage, and commitments it doesn’t need, and none of that trips the kind of alarm a CFO is trained to watch for.
The Part Most Operators Get Wrong
The common fix is a showback dashboard: give engineering leads visibility into what their team spends, and trust that seeing the number changes behavior. It doesn’t, not on its own. A dashboard with no budget authority attached is information, not accountability. Engineering leads who can see the number but aren’t measured against it will look at it exactly once. Real recovery requires that number to sit inside someone’s actual performance review, not just their inbox.
Where the Waste Actually Lives
Saasrooms data from mid-market portfolio companies shows this pattern recurring across industries and company sizes, and it tracks with the layered explanation above.
Consider a healthcare services company managing roughly $9.5 million in annual contracted technology spend across 53 vendors, a stack running from its core ERP and CRM platforms through its cloud reseller agreement. Saasrooms surfaced 49 separate savings opportunities inside that stack, worth $755,100 in approved, actionable recovery, about 8 percent of contracted spend. That 8 percent is what a first pass caught and got signed off. It isn’t the ceiling. A single bundling and contract-restructuring opportunity accounted for $300,000 of it on its own. Seven more, caught by anomaly detection rather than a manual audit, added another $310,000: duplicate tools, unused licenses, plans priced above what actual usage justified. The remaining 41 opportunities, mostly rightsizing and license-level fixes, contributed roughly $145,000, spread across line items small enough that none had been reviewed individually in years.
A different kind of company shows the same shape at a different scale. An agriculture and animal nutrition business with about £920,000 in annual contracted technology spend had £171,000 identified, close to 19 percent, nearly two and a half times the recovery rate of the healthcare example. Two-thirds of that came from a single license-level and cost-avoidance opportunity worth £114,000. The rest came from six smaller items nobody had ever reviewed together.
Same underlying condition, two different recovery rates. Neither number tells you the size of the total waste still sitting in the stack. Both tell you the same thing: nobody had reconciled what was contracted against what was actually being used, and the gap grew for exactly as long as that stayed true.
What Changes the Number
Fixing this for one quarter is easy. Keeping it fixed requires treating cloud cost the way finance already handles any other recurring liability: reviewed on a schedule, owned by a named person, tied to a reporting rhythm the board trusts. Four things separate the operators who hold onto the savings from the ones who watch the number creep back within a year.
Showback tied to budget, not just visibility. A number an engineering lead can see but isn’t accountable for gets checked once and ignored after that.
Rightsizing decoupled from renewal timing. Waiting for a contract renewal to check usage means up to twelve months of unnecessary spend before anyone looks. Usage-level visibility catches an oversized instance the month it happens, not the month the invoice arrives.
Commitments reviewed quarterly, not purchased and forgotten. Reserved instances and savings plans get bought once while the workloads underneath them keep shifting. A quarterly check against actual usage catches the mismatch before it becomes a wasted renewal.
Accountability that has a name attached to it. Waste without an owner stays flat by default. The portfolio companies that actually move the number assign it like any other line item: to a person, with a target, on a schedule.
Before the Next Board Meeting
None of this requires a platform migration or an architecture overhaul. A practical first pass takes three steps: build a current inventory of every cloud and IT vendor contract measured against actual usage, flag the commitments and licenses untouched for the last two quarters, and put one name toward closing the top five findings before the next board meeting. Firms that build this into the value creation plan, rather than treating it as a one-time integration task, tend to find the number stays down instead of just going down.
The 100-day plan ends. The waste doesn’t, not unless someone keeps checking. The portfolio companies that hold onto the recovered EBITDA are the ones that stopped treating cloud cost as an IT line item and started managing it like the recurring liability it’s always been.
