The 7 questions every CFO should ask before approving the cloud line item. Sample dashboards, red-flag patterns, and the metrics that matter in a board pack.
Lakshmi Kiranmai Guduru
Founder, CARTIEAI
If you're a CFO at a software company, the cloud line item is now the second or third biggest cost on your P&L. And it's the one you probably understand least.
This isn't your fault — cloud bills are designed by engineers for engineers. But there's a small set of questions you can ask that will tell you whether the cloud line is healthy, growing well, or quietly destroying your margins.
The single most important number: monthly cloud spend ÷ monthly recurring revenue.
Best-in-class SaaS: 5–10%. Average: 15–20%. Concerning: 25%+ (you might have an architecture problem disguised as a cost problem).
Track the ratio, not the absolute number. Revenue grows 30%/quarter and cloud grows 30%/quarter? Healthy. Revenue grows 30% and cloud grows 50%? Find out why this quarter, not next.
This is the unit-economics view. Three flavors:
Pick the one that maps to your product. Track it monthly. Trend matters more than absolute number. If your $/MAU is climbing, you have a cost problem the engineering team probably hasn't named yet.
Cloud bills are extremely top-heavy. In our audits, the #1 service is usually 35–55% of the entire bill. The top 3 services are usually 75–85%.
Knowing the top 3 isn't optional. If your engineering team can't tell you the answer in 30 seconds, they don't have cost visibility — and you don't have leverage.
Examples (from real audits):
How many times this month did spend jump >20% in a single day vs. the prior 7-day rolling average?
In a healthy account, 0–2 anomalies/month that all have explanations.
In a problem account, 5–15 anomalies/month that nobody investigated.
If you're seeing 5+, your engineering team is flying blind and you need to fix that before doing anything else cost-related.
For AWS (Reserved Instances + Savings Plans), Azure (Reservations), GCP (Committed Use Discounts):
Ask the engineering team for their RI/CUD/Savings Plan coverage report. Should take them 5 minutes to send you.
Most engineering teams will tell you "next month will be ~$X." Push for an interval: "what's the 80% confidence range?"
If they can't answer, they're guessing. A good FinOps team gives a tight interval — typically ±5%. If your team gives you ±25%, that's a process gap worth investing in.
If you've told the board you'll spend $X on cloud this year, and engineering's monthly run-rate × 12 = $1.3X, you have a conversation to have this week — not next quarter. Cloud creeps quietly.
We recommend a 5-row cloud table in every board pack:
| Metric | This month | Last month | Trend | Target |
|---|---|---|---|---|
| Cloud spend ($) | $48,200 | $46,100 | +4.6% | $52K |
| Spend / ARR ($) | 11.2% | 11.9% | -70 bps | 10% |
| Spend / MAU ($) | $0.42 | $0.45 | -6.7% | $0.40 |
| Anomalies (#) | 1 | 4 | ↓ | <2 |
| RI coverage (%) | 68% | 65% | +3 pts | 75% |
That's it. Five rows. Every board, every month.
These aren't financial questions — they're cultural ones. If the answers are unclear, the cost problem is downstream of an organizational problem:
We built CARTIE because every CFO we talked to said the same thing: "I don't want another dashboard. I want a Slack message that tells me what I should do."
So we built that. The CFO digest is a weekly Slack briefing that includes:
If you want to see what your bill looks like through a CFO lens, run the free audit — it'll surface the top 5 cost drivers and the unit-economics math in under 60 seconds.
Companion read: Anatomy of a $100K AWS Bill: What Most CFOs Miss.
Amazon Web Services is the largest cloud platform in the world — and the source of more than half of the cloud waste we audit. This guide gives you the 14 prove…
THE FINOPS BRIEF
Built for finance & engineering teams who are tired of paying for cloud they don't use. No fluff. Just what works.
Unsubscribe anytime. We never sell your data.
ABOUT THE AUTHOR
Founder, CARTIEAI · Building in public
I'm building CARTIEAI to fix the cloud-cost problem I saw drain millions at companies I worked for — where engineering and finance kept talking past each other. If you liked this post, here's where I share unfiltered notes on building this in public:
