A Series C pitch deck once boasted about "80% gross margin." In diligence the FinOps dashboard told a different story: margins were falling a full point every quarter, masked by surging ARR. The buyer added a claw-back clause, and the founders spent the closing week chasing some surprise S3 bills. Cloud spend isn't evil - unless the bill keeps growing faster than the revenue.

Why cloud cost sits on the Tech Due Diligence critical path

  • Margin is not optional. Every extra euro spent on compute, storage, or egress etc. comes straight out of the gross margin - a line item investors can't ignore (no matter in which stage).
  • Spend trends predict long-term scalability. A platform that doubles cost for every 50% ARR jump will choke at scale.
  • FinOps maturity signals governance. Tagged resources, budget alerts, and show-back reports are a good start. These things demonstrate that the team measures what it should later fully control.

The FinOps "vital signs" investors are looking for

Cost per Euro of ARR: In my opinion <0.20 Euros per ARR Euro is healthy for growth-stage SaaS. Climbing ratios might trigger investment model adjustments (discount).
Cloud spend growth vs. revenue growth: When cloud spend grows more slowly than ARR, it signals healthy economies of scale. If the two curves rise at the same pace, or, worse, cloud cost climbs faster than ARR, efficiency is leaking and margins will compress over time.
Tagged-resource coverage: If fewer than 75% of cloud assets carry owner, environment, and cost-center tags, finance can't trace spend to the teams/features/products that generated it. Allocation and margin models break, show-back reports go dark, and investors might see a red flag for hidden waste and weak governance.
Unnecessary idle resource burn: Overnight dev clusters, orphaned snapshots, unused old docker images and "zombie" volumes running 24/7 signal weak cost hygiene. When idle hours start eating a noticeable slice of the monthly bill, investors assume margin erosion and price in the cost of a FinOps cleanup. Of course some fixed resources are non-negotiable - production databases, customer-facing web nodes, load balancers, and monitoring agents must stay up around the clock. Scrutiny falls on everything outside that core footprint.

Stage sets the resolution

  • Series A - snapshots: total monthly cloud bill, top five services, manual tag heat-map.
  • Series B - automated FinOps board: cost per service, per team, trend lines, budget alerts.
  • Series C / Buy-out - hour-level granularity, per-customer/per-project COGS, commitment-discount plan, and policy-driven shutdown rules.

Red-flag patterns investors look at

  • Spend curve steeper than ARR curve: growth is expensive, margin compression imminent.
  • Show-back dashboards missing: no one can tell which team/process/project/product burned last night's 8k spike.
  • Commitment discounts unused: indicates ad-hoc purchasing and poor forecasting.
  • Tag entropy: three tag schemas fighting each other, none enforced. Show-back boards are based on "owner" tag, but half of the resources carry "Owner" or "team" tags instead.

If two or more of these pop up, founders might have to sprint on it in the 100-day plan right after closing.

Habits founders should adopt

  • Run a Monthly Spend Retro. 1 hour, cross-functional, led by the Infra/DevOps team. Review the four vital signs, flag anomalies, assign owners, and log actions in a FinOps Kanban. Even if you don't have a full-blown FinOps platform, start by combing through the detailed cloud bill. You'll often spot hidden costs and an occasional security red flag.
  • Create separate cloud accounts for each product and each environment from Day-1. Finance can then allocate costs and calculate margins with minimal effort. Trying to untangle mixed products and environments later, after everything lives in one giant account, is painful and expensive.
  • Enforce a tagging policy at resource creation. It proves you can trace every euro to a team, feature, or customer - no hidden waste.
  • Set automated budget alerts. It shows real-time governance and prevents surprise overruns.
  • Auto-terminate underutilized instances. Low CPU/memory utilization at 2am signals that it was forgotten.
  • Right-size review every sprint. It demonstrates continuous optimization, not one-off cleanups.
  • Maintain a commitment-discount calendar. It signals disciplined forecasting, reserved instances and savings plans bought on time.
  • Publish cost-per-customer (or per-tenant) dashboards. This links cloud spend directly to unit economics (exactly what investors model). Finance will love it.

Adopting even only two of these will show that your cost controls are ongoing, quantifiable, and owner-driven rather than ad-hoc.

Common traps

  • Cutting staging environments before benchmarking: bugs shift to production and MTTR spikes.
  • Chasing tiny right-sizing wins while ignoring petabyte-scale storage left on standard class.
  • Mixing dev/test spend with production; ratios look fine until an outage bill hits the wrong cost center.

Mini-Glossary

  • FinOps: Discipline of aligning cloud spend with business value.
  • Show-back: Report that surfaces cost to internal teams or products without charging them.
  • Idle burn: Spend on running resources that deliver no direct customer value. These cloud costs are fixed and they don't scale down with usage. For example dev clusters left overnight, orphaned volumes, forgotten sandboxes. Unlike core production spend, this cost can drop to (almost) zero once auto-shutdown or deletion policies kick in.
  • Commitment discount: Savings plans or reserved instances bought up-front for reduced hourly rates.
  • Margin drift: the gradual erosion of gross margin over time. In cloud-based businesses the classic pattern is: revenue keeps rising but variable costs (mostly cloud spend) rise faster, so the gross-margin percentage slides quarter by quarter.
  • ARR: Annual Recurring Revenue. It's a quick way to gauge the steady, predictable income the SaaS business can expect from its customer subscriptions going forward.

Your turn

When did cloud cost last ambush your forecast? Was it idle GPUs, runaway egress, or forgotten test clusters? Drop the story, let's compare notes.

Founders: Need a second set of eyes on your spend curve? Need a migration plan to a multi-account setup? Let's talk.

Investors: Need clarity on FinOps practices of an investment target company? Let's talk.

Next in the Playbook

Edition 6 will tackle incident response - why MTTR, MTTC, and post-mortem culture now sway term-sheet terms. Stay tuned! Subscribe and it'll land in your inbox.

Originally published on the Tech Due Diligence Playbook newsletter on LinkedIn.