SaaS metrics that matter: the short list, and when each one starts to count

Founders tend to track either far too much or the wrong thing for their stage. This guide is the map rather than the manual: what each metric tells you, when it starts to mean something and where it misleads. The formulas get their own deep dives on MRR vs ARR, SaaS churn rate and the LTV:CAC ratio.

Which SaaS metrics matter at each stage?

The right metric is the one that tests your riskiest assumption today. Early on, that is whether anyone sticks around; then whether you can acquire customers profitably; later still, whether growth is efficient.

The last row needs a footnote. Brad Feld's Rule of 40, growth rate plus profit margin adding up to at least 40%, was written for SaaS companies at scale, which he defines as at least $50 million in revenue. Applied to a seed-stage company it tells you very little.

What does each SaaS metric actually tell you?

• MRR / ARR: recurring revenue normalised to a month or a year. Stripe's Billing analytics documentation counts an annual plan as one twelfth a month and leaves out tax, free plans and metered usage.
• MRR growth: new plus expansion, minus contraction and churn. It shows whether the engine is speeding up or stalling.
• Churn: the share of customers, or of revenue, lost in a period. Customer churn and revenue churn can tell very different stories.
• Net revenue retention (NRR): revenue from a group of customers after 12 months, upsells included, against where it started. Above 100% means expansion outruns losses.
• CAC: sales and marketing cost per new customer. Andreessen Horowitz asks for the full cost, stated per customer, and warns that blended figures flatter paid channels.
• LTV: the margin a customer brings over their life with you. a16z insists on profit, not revenue.
• CAC payback: the months of gross margin needed to earn the CAC back.
• Gross margin and burn: what is left after the cost of serving customers, and how fast cash leaves the bank.

How do SaaS metrics fit together? A worked example

Picture a bucket: acquisition pours water in, churn leaks it out of the bottom, expansion makes the bucket bigger. The figures below are hypothetical, chosen to keep the arithmetic easy.

A startup begins the month with 80 customers on $60 a month: $4,800 MRR. During the month 6 new customers sign up (+$360), 2 upgrade (+$40 in total), 3 cancel (−$180) and 1 downgrades (−$20). Net new MRR is $200 and the month closes at $5,000. Customer churn is 3.75%, which sounds modest until it compounds: hold that rate for a year and you lose roughly 37% of the customers you started with.

That is why no metric works alone. Cheap acquisition with heavy churn is a leaky bucket you fill quickly; a high LTV built on three months of churn data is a guess dressed up as a number.

Which SaaS benchmarks are worth trusting?

Treat benchmarks as a thermometer, not a target. Most come from US B2B software companies much further along than a seed-stage startup, and they are rules of thumb rather than datasets.

• LTV:CAC above 3 and CAC paid back within 12 months (5–7 months for the best): David Skok's rules of thumb in SaaS Metrics 2.0.
• Net revenue churn above 2% a month is, for Skok, a sign something is wrong; it adds up to about 22% of revenue a year.
• Burn multiple (net burn divided by net new ARR): David Sacks calls 2 reasonable for an early-stage startup and reads 3 or more as a hint that product-market fit is weaker than it looks.
• Retention rises with price point. In ChartMogul's 2023 retention report, covering more than 2,100 SaaS businesses, 41.1% of companies with average revenue per account above $500 a month kept net retention over 100%, against 2.7% of those below $10.

What makes SaaS metrics misleading?

Definitions far more often than arithmetic. These traps make a number look better than the business behind it, and they hurt most once the figures land in a pitch deck.

• Cumulative charts. A cumulative line only goes up, even in a bad month; a16z lists it among the chart tricks to avoid.
• Annualising one month of cash. Multiply a month that includes set-up fees or annual prepayments by 12 and you get a run rate, not ARR.
• Blended CAC only. Organic sign-ups dilute the cost of paid ones and can hide campaigns that lose money.
• Averages instead of cohorts. Average churn mixes last year's customers with last week's; cohorts by sign-up month show whether the product is improving.
• Shifting definitions. Stripe divides subscribers churned in the last 30 days by subscribers 30 days ago plus new ones in the period; another tool may leave the new ones out. Pick one formula, write it down and keep it.

What is the minimum metrics setup for an early-stage SaaS?

A spreadsheet, one source of truth per number and a fixed monthly routine. If you bill through a subscription platform, most of the data already exists.

1. Write your definitions down: what an active customer is, what counts as MRR, which churn formula you use, which costs go into CAC.
2. Name one source per metric: billing for MRR and churn, your accounts for costs and margin, product analytics for activation.
3. Update monthly on the same day, and annotate each number with what happened: a price change, a campaign, an outage.
4. Review cohorts every quarter; that is where product progress shows up first.

Once you have hundreds of customers and more than a couple of plans, wire the data into a proper dashboard, possibly inside the product itself if you are building a SaaS for customers who need the same view. Until then, discipline beats tooling.