SaaS Metrics for CEOs, CTOs and COOs: The Numbers to Track in 2026 and How to Build Them In

Market-wide SaaS statistics go stale fast, and none of them tells you whether your product works. Your own numbers do. A SaaS team needs a short list: monthly recurring revenue, how many customers leave, how much revenue each group of customers keeps paying, how many trials turn into payers, how often the product is up, and what each customer costs to serve. Each number needs one owner and one written definition, and most of them need code that exists before launch. Below they are grouped by who usually owns them, with definitions from the billing and analytics tools you’re likely to use.

Why this list replaced the 2021 statistics

This post used to list SaaS market figures from 2021. They’re out of date now, and they never answered the question a founder asks on a Monday morning: are more people paying us this month than last month, and why? The numbers below answer it. They also fit on one page, so a CEO, a CTO and a COO can review them together in one meeting.

Revenue numbers: the CEO’s page

Use the definitions your billing system uses, so nobody argues about the math. Stripe publishes its own in the Stripe Billing analytics docs, and they are a good default even if you bill elsewhere.

Monthly recurring revenue (MRR) comes first. Stripe counts the monthly value of every active and past-due subscription. It leaves out taxes, free plans and usage-based products. Its example: 100 subscribers on a $100 monthly plan plus 50 on a $600 yearly plan give an MRR of $12,500.

Churn is the second number. Stripe’s subscriber churn rate divides the subscribers who left in the past 30 days by the subscribers you had 30 days ago plus the new ones since. Start the month with 1,000, add 100 and lose 100, and your churn rate is 100 divided by 1,100, or 9.1%.

Revenue retention follows one group of customers over time. Stripe groups subscribers by the month they started paying and tracks how much of that group’s revenue is left each month. Upgrades count, so a good group can keep more than 100% of its starting revenue.

If you want an outside reference point, SaaS Capital’s 2025 retention benchmarks for private B2B companies, published in September 2025, put median net revenue retention at 102% for companies whose annual contracts are worth $25,000 to $50,000. They also found that retention runs higher when contracts are larger. Their figure compares December 2024 revenue from customers who were already paying in December 2023 with all of December 2023’s revenue. Compare your number only if you count it the same way.

Trial conversion is the share of trials that turn into paid plans. Stripe divides the trials that converted in the last 30 days by the trials that ended in the same 30 days.

Two smaller numbers come from these. Average revenue per user is MRR divided by active subscribers. Lifetime value is that average divided by the churn rate, and Stripe itself calls it an estimate. Treat it as a direction, never as a forecast.

One decision to write down early: do discounts reduce your MRR? Stripe lets you choose, and calls subtracting them the more conservative way.

Operations numbers: the COO’s page

These show whether the business can serve more customers without breaking.

• Time to first value: how long a new customer takes from sign-up to the first job your product does for them.
• Failed payments: cards that bounced and accounts that are past due. Stripe keeps past-due subscriptions in MRR until they’re canceled or marked unpaid, so watch them as their own number. Much of that money can come back. Stripe’s Billing page says businesses using Stripe recover 55% of failed payments on average, and that its tools recovered US$8.2 billion in failed payments in 2025.
• Support load: questions per customer per month, and the top three reasons.
• Cost to serve: servers, storage, AI calls, payment fees and support tools, per customer. Count the billing tool too. Stripe Billing’s pricing page lists 0.7% of billing volume on its pay-as-you-go plan.

Cost to serve can fall as a product learns. On IPOPBOOKS, a bookkeeping platform we built for US small businesses, every merchant a bookkeeper approves once gets sorted from memory the next time, with no AI call. Over time each client needs fewer AI calls, and that only shows up if you measure cost per client.

Product and engineering numbers: the CTO’s page

• Uptime and error rate, measured against a target you’ve written down. Our guide to SaaS uptime and what to promise customers covers how to set that target.
• Activation: the share of new accounts that complete the actions which mean “this product works for me”.
• Releases: how often you ship, and how many releases you had to roll back.
• Cloud and AI cost per customer, shared with the COO.

Activation needs events, and events need code. The Google Analytics help page on event types explains that recommended events have names and parameters set in advance, while custom events are ones you define. You have to implement both kinds yourself. Nothing records “created first invoice” unless a developer adds it.

The whole page in one table

Number	Owner	How it’s counted	Data source
MRR	CEO	Monthly value of active and past-due subscriptions, without taxes and free plans	Billing system
Subscriber churn rate	CEO	Subscribers lost in 30 days divided by subscribers at the start plus new ones	Billing system
Revenue retention	CEO	Revenue left from one starting group, month by month	Billing system
Trial conversion	CEO	Trials converted divided by trials ended, in the same 30 days	Billing system
Time to first value	COO	Time from sign-up to the first completed job	Product events
Failed payments	COO	Past-due accounts and declined charges	Billing system
Cost to serve	COO and CTO	Cloud, AI, payment and support costs divided by customers	Cloud bills, product logs
Uptime and errors	CTO	Share of time and requests that worked, against the target	Monitoring
Activation	CTO	Share of new accounts that complete the key actions	Product events

Build the numbers in before launch

A first release that can’t show these numbers will cost you a second release just to add them. Four things to build in:

1. Billing as the source of truth: read revenue from the billing system, not from a copy in your own database that drifts.
2. A short event plan: pick the handful of actions that mean a customer got value, name them, and record them from day one.
3. An admin panel with numbers per customer: the IPOPBOOKS admin panel shows clients, staff, which bookkeeper serves whom, and a log of every posted entry.
4. Cost tags: mark cloud and AI usage per customer where you can, so cost to serve isn’t a guess.

QuizWithIt, a quiz platform we built for a science YouTuber, launched with a free tier and paid subscriptions. It got 5,000+ users in its first 3 months, about 100 paid subscriptions right after launch, and reached break-even in 4 months. Numbers like these are only visible when the plans, payments and sign-ups are counted from the first day.

How we would estimate your SaaS

The same numbers decide whether a build pays for itself. You can see what your SaaS costs to build with our project estimate. It works in four steps:

1. Your brief becomes user stories, including the plumbing for these numbers: billing, the event plan, the admin view and cost tags. You confirm what is in scope, and every hour in the budget traces back to a story you kept.
2. We check the hours the model suggests against what our team actually books on finished projects, then multiply by our published rate of $39-55 an hour. It’s a planning range, not a fixed quote.
3. For market size, the estimate suggests customer groups and gives each a TAM, SAM and SOM with a note on what the number covers. Treat them as hypotheses to test, not verified demand. Keep only the groups you would really sell to, and the market and unit economics are worked out only for those.
4. The unit economics show how many customers pay for the build and in which month it is paid back. You can change any starting number and the math updates.

Here is an example run of our estimate’s math with sample inputs, for a B2B web app with teams, roles and subscription billing. The inputs are assumptions, not a forecast: a customer pays $99 a month at an 80% margin, stays 24 months and costs $600 to win, and 15 new customers sign up each month. With those guesses, the build comes out at $6,600 to $13,600 over 5 weeks. Each customer brings $79 of profit a month and $1,896 over their lifetime, which leaves $1,296 after the cost to win them. 8 customers pay for the build. The product earns more than it spends from month 8 and pays the build back in month 16. Once your churn is real, change how long a customer stays to match it, and every number after it moves.

What we would do

1. Write one definition per number on a single page, starting from your billing provider’s own definitions.
2. Give each number one owner: the CEO, the COO or the CTO.
3. List the handful of product events that show value, and add them to the first release.
4. Build an admin view with the numbers per customer before launch.
5. Review the page monthly, and change one thing in the product because of it.

If you’re building a SaaS, our SaaS development services page shows two ways to pay for the build and the SaaS products we’ve shipped. If you’re still choosing who builds it, our comparison of SaaS development companies lists rates, minimum projects and Clutch ratings with a source for every figure. The QuizWithIt numbers are in the QuizWithIt case study.

FAQ

Which SaaS metrics should an early-stage founder track first?

MRR, subscriber churn rate, trial conversion and activation. Add cost to serve as soon as you pay for AI calls or heavy cloud usage, and uptime as soon as customers depend on the product for daily work.

How do you calculate MRR?

Add up the monthly value of every active and past-due subscription, without taxes, free plans and usage-based products. In Stripe’s example, 100 subscribers at $100 a month plus 50 at $600 a year give an MRR of $12,500.

How do you calculate SaaS churn rate?

Stripe divides the subscribers who left in the past 30 days by the subscribers you had 30 days ago plus the new ones since. If you started with 1,000, added 100 and lost 100, churn is 100 divided by 1,100, or 9.1%.

What is a good churn rate for SaaS?

We don’t quote a single benchmark, because public figures mix products with very different customers and prices. SaaS Capital’s 2025 data, for example, shows revenue retention running higher when contracts are larger. Track your own rate every month, by customer group, and look for the reason behind every change.

When should I set up analytics in a new SaaS?

Before launch. Recommended and custom events in Google Analytics only exist once a developer implements them, so plan the events with the first release and check that they record during testing.