SaaS Metrics: The 20 That Matter, With Formulas and Benchmarks

SaaS metrics are the numbers a subscription software company uses to measure how big it is, how fast it grows, whether customers stay and expand, and what each dollar of growth costs. Twenty of them do most of the work. This page puts all 20 in one table with the formula, a sourced benchmark and ORM's glossary entry.

If I could put one chart in front of every SaaS CFO, it would be the monthly ARR waterfall: beginning ARR, the three lines that take ARR away, the three that add it, and ending ARR, with gross and net revenue retention on the same page. Most of the 20 metrics below are a cut of that one reconciling chart. Get it right and the rest follow. Get it wrong and every ratio built on top of it is wrong too.

The pages that rank for this query today list anywhere from 8 to 28 metrics, usually as cards or headings with no table, and most of their benchmarks carry no source. Several have formula errors, such as a revenue retention formula that adds new-customer revenue. This page sources every benchmark to a named survey or investor, shows which metrics matter at which stage, and works one illustrative company through all of them so you can check the math.

What are SaaS metrics?

SaaS metrics are measures built for recurring revenue: they track contracts that renew, expand or cancel, not one-time sales. That is what separates them from generic business KPIs. A software company's revenue this year is mostly decided by customers it signed in earlier years, so the metrics that matter most watch that base: how much of it stays, how much grows, and what it cost to win.

They fall into five families:

• Growth: how big the recurring base is and how fast it grows (ARR, ARR growth, net new ARR, ARPA).
• Retention: whether existing customers stay and buy more (GRR, NRR, logo retention, expansion share).
• Unit economics: what a customer costs to win and what it returns (CAC, CAC payback, LTV, LTV to CAC).
• Efficiency: how much cash and margin the growth consumes (gross margin, magic number, burn multiple, Rule of 40, quick ratio).
• Pipeline and forecast: whether next quarter's new ARR will show up (pipeline coverage, win rate, forecast accuracy).

The first four families describe the past. The fifth is the only one that looks forward, which is why a board that trusts the first four still asks how confident you are in next quarter.

The 20 SaaS metrics in one table

Here is the whole cheat sheet: each metric, its formula, a benchmark with its source, and the ORM glossary entry that goes deeper. Read the benchmarks as reference points, not targets. Each survey has its own sample, and the right number for a $3M ARR SMB product is not the right number for a $60M ARR enterprise platform. Three notes on the formulas. NRR and GRR must use the same customers at the start and end of the period, with new customers left out; a formula that adds new-logo revenue is measuring growth, not retention. LTV and CAC payback need gross margin in them, or they overstate what a customer returns. And the magic number and burn multiple both put net new recurring revenue on top, so they move together when growth slows.

Which SaaS metrics matter at each stage?

Early companies should track fewer metrics and read them by hand; later companies need ratios because nobody can read every account. The table shows where each family earns its place. The stage bands are ORM's working view, not a survey. Two points from working with B2B SaaS companies at each of these stages. First, small companies still need clean ARR reporting even when they are too small to forecast with a model. I once worked with a company at about $5M ARR with three reps and a founder in every deal. Forecasting there was a conversation with the reps. ARR reporting was still worth building, because it is the base every later metric needs.

Second, growth expectations changed after 2022. In 2021, Bessemer put median ARR growth at 165% for companies between $1M and $10M ARR (Bessemer, 2021). The 2024 median across private SaaS was 26% (Benchmarkit, 2025). A growth plan built on the old tables will look like a miss even when the business is healthy. For the sales-side view of the same question, see which sales metrics to track at each company stage.

How do SaaS metrics connect?

Almost every SaaS metric is built from five inputs: the ARR bridge, customer counts, sales and marketing spend, gross margin and cash burn. Change one input and several metrics move at once. That is why reading one ratio alone gives a wrong picture. Two relationships are worth memorizing. NRR minus GRR equals expansion as a share of beginning ARR, so a company with 88% GRR and 101% NRR is expanding by 13 points to stay just above flat. And a burn multiple is roughly the inverse of efficiency: if net new ARR falls while burn holds, both the burn multiple and the magic number get worse in the same quarter. For how the waterfall itself is built, see how to build an ARR waterfall and the ARR bridge entry.

Worked example: one company, all 20 metrics

Here is one illustrative B2B SaaS company run through the table, so you can see the metrics agree and where they disagree. All figures are illustrative.

The company starts the year at $10.0M ARR with 200 customers. During the year it adds 40 new customers worth $2.0M, expands existing accounts by $1.2M, loses $0.3M to downgrades and $0.6M to 14 cancelled customers. It spends $3.5M on sales and marketing, $2.8M of it on new logos, runs a 78% gross margin and burns $2.5M of cash. It forecast $2.2M of new-logo ARR.

Read together, the story is mixed, and that is the point. GRR of 91% and NRR of 103% sit above the medians. The burn multiple of 1.09x is in Craft's "great" band. LTV to CAC of 6.2 looks excellent. But CAC payback of 21.5 months is past Bessemer's 18-month mid-market target, and the Rule of 40 score is near zero. The high LTV to CAC comes from low churn, not cheap acquisition. A board that saw only LTV to CAC would approve more spend; one that saw payback would ask why each new customer takes almost two years to earn back. Both are true. (Rule of 40 normally uses revenue; average ARR stands in for it here.)

Which SaaS metrics should you report to the board?

Report six to eight: ending ARR with its bridge, NRR and GRR, the forecast as a range, burn multiple with runway, CAC payback, gross margin and next quarter's pipeline. Everything else goes in an appendix, built from the same data the operating team uses.

Credibility with a board is won in the detail. A board trusts a forecast that is consistent, traceable and framed as a range. If a number appears on one slide, it had better appear on the next, or there should be a clear reason it changed. It should roll up the same way from region to total. You should be able to drill all the way down and explain how it came together. And it should not pretend there is one outcome: show a low case, an on-target case and an upper bound.

Much of the work of pulling these numbers together can be automated, and we do that for customers today. But the revenue leader still has to be able to speak to every number on the page. For slide-level detail, see which sales metrics belong in a board deck, the board revenue update template and how to present a forecast to the board.

Which SaaS metrics mislead?

The metrics that mislead most are the ones read without their context: pipeline coverage, blended churn and LTV. Each is a real number that answers a narrower question than people think. Pipeline coverage without context. Most teams still lean on a 3x to 5x pipeline-to-goal rule. Most ORM customers run about 3.5x, and we have seen them as low as 1.4x and as high as 5x. A company can have 4x coverage and still miss badly if the pipeline is stale, sits in the wrong segment or carries inflated amounts. One pattern we see: a pipeline with an $80,000 average deal size where closed-won deals average $40,000. Across ORM customers, only about 20% of the pipeline with a close date in the quarter on day one closes in that quarter. Blended churn. Logo churn and revenue churn answer different questions, and a company can lose many small customers while revenue churn stays low, or the reverse. Track both, and split SMB from enterprise. See revenue churn vs logo churn. LTV built on a tiny churn rate. Divide by a 3% annual churn rate and LTV assumes customers stay for 33 years. As the worked example shows, a high LTV to CAC can sit beside a slow payback. Read LTV to CAC next to CAC payback, never alone. Retention metrics are lagging. By the time GRR drops, the customers have left. The leading signals sit elsewhere. Across ORM customers, renewal risk usually comes into focus about six months before the renewal date. Support tickets behave in a way most health scores miss: a customer with zero cases in a year is a churn risk, so is one with seven or more, and one with three to five resolved cases is usually among the happiest.

SaaS acronyms cheat sheet

These are the SaaS acronyms you will see in board decks, investor updates and benchmark reports.

How ORM uses these metrics

ORM builds the ARR bridge, retention, pipeline and forecast metrics from your CRM and billing data, so they reconcile to what finance reports. The same monthly waterfall feeds GRR, NRR and expansion share, and the forecast sits on top of it as a range. Accuracy on new and expansion revenue built by hand usually reaches about 90%, with a lot of manual work. ORM targets 95% with no manual changes, from day 1 to day 90 of the quarter. See ORM's SaaS analytics for cohorts and health scores, and the net revenue retention entry for the NRR benchmarks by ARR band.