SaaS Metrics: The KPIs to Track, With Formulas and Benchmarks
SaaS metrics are the numbers that show whether a subscription software business is growing, keeping the customers it wins and earning back what it spends to win them. This guide covers 15 of them, the SaaS KPIs from MRR and ARR to the rule of 40, each with its formula, a worked line for one made-up company and a benchmark from a named source read on 16 November 2026. Two of those benchmarks set the scale: the median private B2B SaaS company in SaaS Capital's 2026 survey grew 22% a year, and where a customer pays less than $12,000 a year, median net revenue retention in its 2025 survey was 98%.
It is written for a founder between the first paying customer and early growth who wants three answers for every metric: what it tells you, how to work it out and what a good value looks like. Acme, a B2B invoicing tool invented for this guide, carries every formula, so each number can be checked against the others. The marketing side of the same numbers, channel by channel, is in our SaaS marketing guide, and the MRR calculator runs the MRR, ARR and retention arithmetic on your own figures.
Track MRR, customer churn and CAC from the first customer, add CAC payback and LTV:CAC at about a hundred customers and leave the magic number and the rule of 40 for later. Read every benchmark next to its sample: Acme sits in the top quarter on ChartMogul's churn figures and 5 points under the median on SaaS Capital's gross retention. Listings are an acquisition cost you pay once: Go Publicly sells one Premium launch for $19, while for $29 Submitator files 40 you choose from 202 directories and launch platforms.
What are SaaS metrics? Four groups of SaaS KPIs
A SaaS metric is any number computed from subscriptions, customers and the money spent winning them. A KPI is the short list of those numbers a company decides to run on this quarter, so every SaaS KPI is a SaaS metric and most SaaS metrics are never anyone's KPI. The 15 in this guide fall into four groups:
- Revenue metrics (MRR, ARR and ARPA): how big the recurring business is and how fast it grows.
- Retention metrics (customer churn rate, revenue churn, gross revenue retention and net revenue retention): whether the business keeps the revenue it has already won.
- Unit economics (CAC, LTV, LTV:CAC and CAC payback): whether winning a customer pays, and how soon.
- Efficiency metrics (quick ratio, magic number, burn multiple and rule of 40): how much cash or spend each dollar of growth costs.
SaaS growth metrics cut across the groups: net new ARR sits with revenue, net revenue retention with retention and the burn multiple with efficiency. The SaaS financial metrics an investor asks for first (gross margin, EBITDA, net burn) are inputs here rather than rows of their own: gross margin feeds LTV and CAC payback, net burn the burn multiple and EBITDA the rule of 40. Together, CAC, LTV and payback are what people mean by SaaS unit economics.
Key SaaS metrics: a cheat sheet with formulas and benchmarks
Acme, the made-up company behind every number
Acme does not exist. The invoicing tool has one plan at $100 a month and the round numbers below for its last twelve months. Nothing in it is a benchmark, and none of it describes Submitator or a customer of ours.
- A year ago: 500 paying customers and $50,000 of MRR.
- Over the year: 60 of those customers cancelled, downgrades took $1,500 off MRR and upgrades added $7,500.
- New business: 300 customers won at $100 a month, all still paying at the end of the year.
- Costs: a gross margin of 80%, $480,000 on sales and marketing and $612,000 on product, engineering and administration. No debt and nothing capitalised, so its cash burn equals its operating loss.
- Growth: MRR rose by about the same amount each month, from $50,000 to $80,000.
SaaS metrics cheat sheet: 15 KPIs, their formulas, Acme's numbers and SaaS benchmarks
| Metric | Formula | What it answers | Acme (made up) | Benchmark and source |
|---|---|---|---|---|
| MRR (monthly recurring revenue)Revenue | Paying customers × ARPA, annual plans at a twelfth of their price | How much recurring revenue comes in each month | $50,000 at the start, $80,000 at the end | No target. 53% of 1,043 companies reached $10k MRR at $30 to $299 a month per account (ChartMogul, 2026) |
| ARR (annual recurring revenue)Revenue | MRR × 12 (the annualized run rate) | What the subscriptions running now are worth over a year | $80,000 × 12 = $960,000 | Growth: median 22% a year (SaaS Capital, 2026 survey of 1,000+ companies) |
| ARPA or ARPU (average revenue per account or user)Revenue | MRR ÷ paying accounts (ARPU: ÷ users) | What the average account pays a month | $80,000 ÷ 740 = about $108; new customers pay $100 | No target; it decides which benchmarks apply (SaaS Capital sorts retention by contract value) |
| Customer churn rateRetention | Cancellations in the period ÷ customers on its first day | How many accounts leave | 60 ÷ 500 = 12% a year, about 1% a month | Median 3% to 4% a month, top quarter 1% to 2% (ChartMogul, 2,100+ companies, 2023) |
| Revenue churn (gross MRR churn)Retention | (Churned MRR + contraction MRR) ÷ MRR at the start | How much revenue leaves, downgrades included | ($6,000 + $1,500) ÷ $50,000 = 15% a year | 100% minus GRR: about 10% a year at SaaS Capital's 90% median under $12k (2025) |
| GRR (gross revenue retention)Retention | MRR last year's customers pay now, upgrades left out ÷ their MRR a year ago | How much of last year's revenue base is still there | $42,500 ÷ $50,000 = 85% | Median 90% under $12k a year per customer, 95% over $250k (SaaS Capital, 2025) |
| NRR (net revenue retention, or net dollar retention)Retention | The same with upgrades counted | Whether existing customers pay more or less than a year ago | $50,000 ÷ $50,000 = 100% | Median 98% under $12k, 106% over $250k (SaaS Capital, 2025); 82% for B2B in ChartMogul's data (2025). Work out yours |
| CAC (customer acquisition cost)Unit economics | Sales and marketing spend ÷ new customers won, same period | What sales and marketing pay for each new account | $480,000 ÷ 300 = $1,600 | No dollar target. Median spend: 15% of ARR on selling, 8% on marketing (SaaS Capital, 2026) |
| LTV (customer lifetime value)Unit economics | ARPA × gross margin ÷ monthly customer churn | Expected gross profit from one account over its life | $100 × 80% ÷ 1% = $8,000 | No target. 28.3% of 35,512 cohorts missed the formula by more than half (ChartMogul, 2026) |
| LTV:CAC ratioUnit economics | LTV ÷ CAC | Whether a customer is worth what it cost | $8,000 ÷ $1,600 = 5.0 | 3 or more (David Skok, 2013; Bessemer, 2021) |
| CAC payback periodUnit economics | CAC ÷ (a new customer's ARPA × gross margin) | Months until a customer repays its CAC | $1,600 ÷ $80 = 20 months | Under 12 months for SMB, 18 mid-market, 24 enterprise; average 15 at $1M to $10M ARR (Bessemer, 2021) |
| Quick ratioEfficiency | (New MRR + expansion MRR) ÷ (churned MRR + contraction MRR) | MRR gained for each dollar of MRR lost | ($30,000 + $7,500) ÷ $7,500 = 5.0 | Above 1, MRR grew over the period |
| Magic numberEfficiency | (This quarter's revenue - last quarter's) × 4 ÷ last quarter's sales and marketing spend | New annualized revenue per $1 of sales and marketing | $22,500 × 4 ÷ $120,000 = 0.75 | Median 0.7 across 1,000+ growth-stage companies (Scale Venture Partners, 2020) |
| Burn multipleEfficiency | Net burn ÷ net new ARR | Cash burned for each $1 of new ARR | $468,000 ÷ $360,000 = 1.3 | Under 1x amazing, 1x to 1.5x great, over 3x bad (David Sacks, 2020) |
| Rule of 40Efficiency | Revenue growth rate (%) + profit margin (%) | Whether growth and profit balance at scale | 60% + (-60%) = 0 | 40 or more at $50M+ revenue (Brad Feld, 2015); public SaaS median 41% (Scale, 2020) |
Acme is made up: one $100 plan, 80% gross margin, the twelve months set out above the table; figures are rounded to the dollar, the ratios to one or two decimals. Benchmarks, each read on its source's own page on 16 November 2026. SaaS Capital: its 2025 survey of private B2B SaaS companies, medians by annual contract value from the chart on its retention page of 18 Sep 2025 (the page gives no respondent count), and its 15th annual survey, completed in March 2026 with more than 1,000 respondents, for growth (page of 22 Jul 2026) and spending (10 Jun 2026). ChartMogul: anonymized subscription data from the accounts on its platform; churn from more than 2,100 SaaS businesses (report of 24 May 2023), B2B NRR from a dataset of about 2,700 B2B SaaS companies, counting those with $250,000 of ARR or more through 2025 (10 Dec 2025), LTV from 35,512 cohorts of 3,331 accounts (15 Jul 2026), $10k MRR paths from 1,043 companies (3 Jun 2026). Bessemer: "Scaling to $100 Million" (21 Sep 2021), its own cloud portfolio from 2010 to the first half of 2021, which it says is not a random sample. Scale Venture Partners: the Scale Studio dataset of 1,000+ growth-stage companies (8 Sep 2020) and 68 public SaaS companies (9 Mar 2020). David Sacks (23 Apr 2020), David Skok (16 Jan 2013) and Brad Feld (3 Feb 2015): rules of thumb from investors' experience, with no sample. The burn multiple bands and SaaS Capital's medians by contract value are copied from images on those pages.
Read down the last two columns and Acme looks like three different companies. By ChartMogul's churn figures it sits in the top quarter: 1% of customers lost a month against a median of 3% to 4%. By SaaS Capital's retention figures it is below the median for its price band: gross revenue retention of 85% against 90%. Its LTV:CAC of 5.0 clears the 3 that David Skok and Bessemer set, while its CAC payback of 20 months runs past every target in Bessemer's report but the enterprise one.
None of that is a contradiction. ChartMogul reads anonymized subscription data from businesses on its platform, B2C among them; SaaS Capital asks private B2B companies to report their own retention; Bessemer reads its own cloud portfolio. Each benchmark describes the companies behind it, so set yourself against the one whose companies look like yours: the same price band, the same kind of buyer, a similar size. SaaS benchmarks without a sample are a number with nothing to compare it to.
ARR vs MRR: what each one measures
MRR is the recurring revenue your subscriptions bring in a month, with annual and quarterly plans counted at their monthly value. ARR is the same revenue over a year. ChartMogul's glossary gives the usual meaning in one line: “In modern SaaS, ARR usually means Annualized Run Rate: your Monthly Recurring Revenue (MRR) multiplied by 12.” Acme ends its year on $80,000 of MRR, which makes its ARR $80,000 × 12 = $960,000.
So ARR vs MRR is a choice between two views of one number. MRR moves month to month and shows a price change or a bad month of churn within weeks; ARR is the figure investors, lenders and yearly contracts speak in. A company selling mostly monthly plans steers by MRR, and one selling annual contracts plans and reports in ARR. The movements behind both (new, expansion, contraction and churned MRR, plus the net new MRR they add up to) are what our MRR and ARR calculator works through for your own numbers.
What ARR means: annual recurring revenue or annualized run rate
ARR has two readings. As annualized run rate it is MRR × 12, whatever the billing period. As annual recurring revenue in the strict sense it counts contracts of twelve months or more and leaves monthly plans out. If 100 of Acme's customers pay yearly, its strict annual recurring revenue is $120,000, against a run rate of $960,000. Say which one you mean whenever you print the number. Neither reading takes in one-time fees, and ChartMogul names the slip: “Common mistake: folding one-time fees or services revenue into ARR overstates predictable, recurring revenue.”
Why ARR is higher than revenue
Because ARR is a snapshot and revenue is a sum. Acme's revenue for the year came to about $780,000: twelve months of MRR that started at $50,000 and ended at $80,000. Its ARR at the year's end is $960,000, the next year priced at the closing run rate. In a growing company ARR runs ahead of the revenue already booked; in a shrinking one it falls behind. And ARR is worth what the retention behind it is worth: with gross revenue retention at 85%, about $144,000 of Acme's run rate leaves within a year if the next year repeats the last.
ARR growth is the headline growth rate. Acme's went from $600,000 to $960,000, 60% in a year. The median in SaaS Capital's 2026 survey of more than 1,000 private B2B SaaS companies was 22%, and ChartMogul, looking at 6,525 software companies, found that “for startups that reach $1 million ARR, it usually takes between two and five years from initial monetization.” Four movements decide ARR: win more, expand more, lose less and downgrade less.
SaaS churn rate: customer churn, revenue churn and retention
Churn in SaaS is the share of customers, or of revenue, that leaves in a period. The two versions answer different questions, and Acme shows why a company needs both. Customer churn counts accounts: 60 of its 500 customers cancelled, a customer churn rate of 12% a year, about 1% a month. Revenue churn counts dollars and adds the downgrades: ($6,000 + $1,500) ÷ $50,000 = 15% a year. When the customers who leave pay less than average, revenue churn runs below customer churn; when large accounts leave or step down, it runs above.
How to calculate churn rate in SaaS
- Fix the period and its base: the customers, or the MRR, you had on its first day.
- Count what left that base. Customers who joined after the first day are not in it, so their cancellations go in a separate line rather than into the rate.
- Divide. Customer churn is cancellations ÷ customers in the base; revenue churn is (churned MRR + contraction MRR) ÷ MRR in the base.
- Compare like with like. A yearly rate divided by twelve, as this guide does for Acme, is a round approximation of the monthly one: 1% a month compounds to about 11.4% a year, not 12%.
Gross revenue retention (GRR), and what NRR adds
Retention turns revenue churn around and follows one group of customers over a year. Gross revenue retention (GRR) is the MRR that last year's customers still pay, with upgrades left out, divided by what they paid a year ago: for Acme, $42,500 ÷ $50,000 = 85%. SaaS Capital's definition sets the ceiling: “For this reason, gross retention cannot exceed 100%.” Net revenue retention (NRR, also called net dollar retention) adds the upgrades back: $50,000 ÷ $50,000 = 100%. Acme's existing customers paid as much at the end of the year as at its start, even after 60 of them left. Above 100%, a company would grow with no new customers at all.
Average SaaS churn rate: why the benchmarks disagree
The average SaaS churn rate depends on whose companies are averaged. ChartMogul, on anonymized data from more than 2,100 SaaS businesses in 2023: “As companies find product market fit and grow, the median customer churn rate initially declines and then stabilizes at 3-4% per month.” and “If you have managed to bring down your customer churn rate to 1-2% per month, you are already among the top 25% of all SaaS companies.” SaaS Capital's 2025 survey of private B2B companies put median GRR at 90% where customers pay under $12,000 a year and 95% above $250,000: 10% and 5% of revenue lost a year. Its 2026 survey found a median of 91% among bootstrapped companies with $3 million to $20 million of ARR. Bessemer, reading its own portfolio in 2021: “Looking at gross retention by ARR range, we find that it is relatively consistent at 85-90%.”
Net retention splits the sources further. ChartMogul's report on retention through 2025 drew on about 2,700 B2B SaaS companies and counted those with at least $250,000 of ARR: “B2B SaaS is relatively sticky. The median NRR was 82%.” SaaS Capital's medians for the same metric run from 98% to 106% by contract value: two sets of SaaS churn rate benchmarks for B2B companies, 16 points or more apart.
How to reduce churn in SaaS
- Find out who leaves and why. One question in the cancellation flow, and a call with the first twenty customers who cancel.
- Shorten the road to a first result. An account that never got value from the product has nothing to lose by leaving.
- Recover failed payments. Card retries and reminder emails bring back customers who never meant to go.
- Offer a step down. A cheaper plan or a pause in the cancel flow keeps the account, and with it the chance of an upgrade later.
- Sell yearly plans. A customer on an annual plan decides once a year rather than twelve times.
How to calculate CAC (customer acquisition cost)
Customer acquisition cost divides what a company spent on sales and marketing over a period by the paying customers that period brought in. Acme spent $480,000 and won 300, so its CAC is $1,600. Three choices decide whether the number means anything.
- What goes in. Fully loaded CAC counts every sales and marketing cost: salaries and commissions, ads, tools, agencies, launches and listings. Paid CAC counts media spend alone and flatters any company that sells through people. Payback and LTV:CAC use the fully loaded figure.
- Which customers count. New paying customers: not signups, not trials and not upgrades from existing accounts, which belong to expansion.
- Which period. Spend in one month often wins customers in the next month or the next quarter, so a quarter or a year gives a steadier CAC than a month.
How much CAC is good has no answer in dollars: $1,600 is cheap for a product at $1,000 a month and ruinous for one at $10, which is why the payback and LTV:CAC sections below measure it against what a customer pays. Spend as a share of revenue is the other check: Acme puts 50% of its closing ARR into sales and marketing, more than twice the 23% (15% selling, 8% marketing) of the median company in SaaS Capital's 2026 survey.
CAC by channel: one-time costs and monthly costs
A blended CAC hides which channel does the work, so split it: each channel's cost over the customers you can trace to it. The split is easy for costs that recur and awkward for costs that do not.
Acme's $480,000 breaks into $336,000 of salaries and commissions, an ad budget of $8,000 a month ($96,000 over the year) and $48,000 of one-off items: a launch, a conference and directory listings. A monthly cost such as ads or an agency retainer enters every month's numerator in equal parts and stops when the spending stops. A one-time cost enters once. Counted in the month it was paid, it makes that month's CAC jump and every later month look free, even while the launch page or the listings are still up and still being found.
There are two fixes, and either works as long as you write down which one you used. Measure CAC over a window long enough to hold the asset's working life, a quarter or a year. Or spread the cost: a $29 directory package written off over twelve months adds $2.42 a month to the numerator. Content and programmatic SEO pages behave the same way, with the cost in building them and whatever they bring arriving later. What each channel costs to start is set out stage by stage in the SaaS marketing guide, and what directory listings cost, free and paid, in our directory submission cost breakdown.
Go Publicly: $19 for one Premium launch. Submitator: $29 for 40 directories.
Go Publicly is one of the 81 launch platforms and startup directories we submit to, filed on its free tier. Pick it and 39 more; each gets a description written for its form, and every badge one of them asks for stays live on our side. 70 for $49, 110 for $79, one payment.
LTV:CAC ratio: formula, a good ratio and what moves it
The LTV:CAC ratio compares what a customer is worth with what it cost to win: customer lifetime value divided by customer acquisition cost. LTV estimates the gross profit a customer leaves behind over its life: ARPA × gross margin ÷ monthly customer churn.
LTV = $100 × 80% ÷ 1% = $8,000; LTV:CAC = $8,000 ÷ $1,600 = 5.0
Put in words, LTV vs CAC for Acme: a customer it pays $1,600 to win returns $8,000 of gross profit over its life, 5.0 times the cost.
A good LTV:CAC ratio: 3 or more
Three is the line two investors drew eight years apart. David Skok, in SaaS Metrics 2.0 in 2013: “The best SaaS businesses have a LTV to CAC ratio that is higher than 3, sometimes as high as 7 or 8.” Bessemer, in 2021: “We recommend investing in customer acquisition when CLTV / CACs are 3x+”. Under 1, each new customer loses money over its life; between 1 and 3 the business works on paper with a thin margin for error.
Customer lifetime value: the formula and its blind spot
LTV is the weaker half of the ratio, because it is a forecast. ChartMogul tested the formula against 35,512 customer cohorts and reported in July 2026: “28.3% of all cohorts have a prediction error larger than 50% in either direction.” Its advice: “LTV is best used as a directional metric, not a precise forecast.” Two habits keep the error down. Use what new customers pay rather than the blended average: Acme's blended ARPA is about $108 after a year of upgrades, while a new customer starts at $100. And use customer churn rather than net revenue churn: Acme's net revenue churn is zero, and a lifetime value divided by zero never ends.
What moves LTV:CAC
Churn moves it most. At 2% a month Acme's LTV halves to $4,000 and the ratio falls to 2.5; at 3%, the low end of ChartMogul's median, LTV is about $2,667 and the ratio 1.7. Price and gross margin move LTV in proportion, and CAC moves the denominator. Because both halves are estimates, work them out from the same period.
CAC payback period: formula, benchmarks and what moves it
The CAC payback period is the number of months a new customer takes to repay its acquisition cost out of gross profit. The CAC payback formula divides CAC by what a new customer earns you each month after the cost of serving it:
CAC payback = $1,600 ÷ ($100 × 80%) = $1,600 ÷ $80 = 20 months
Some versions leave gross margin out, which would make Acme's payback 16 months and flatter any company that pays to serve its customers. Bessemer measures it the way the formula above does, against gross margin-adjusted ARR.
What is a good CAC payback period?
It depends on the buyer. Bessemer's targets, from its cloud portfolio in 2021: “For cloud companies selling into SMB-focused accounts, you should target CAC payback <12 months; for mid-market-focused accounts, target CAC payback <18 months; and for enterprise-focused accounts, target <24 months.” The same report found the average at the small end: “What we find is that the average CAC payback in the $1-10MM ARR bucket is 15 months”. David Skok, earlier and stricter: “And many of the best SaaS businesses are able to recover their CAC in 5-7 months.” and “It shows how the profitability is anemic if the time to recover CAC extends beyond 12 months.” Acme sells to small businesses, so its 20 months run 8 past the SMB target.
LTV:CAC says yes, payback says wait
Acme passes the LTV:CAC test at 5.0 and fails the payback one at 20 months, and both readings are right. LTV:CAC asks whether a customer pays back eventually; payback asks how long the cash is out first. A company that grows fast on a long payback needs money to fund the wait, which is why investors ask for both.
What moves CAC payback
- Price. At $200 a month, double the price, the same CAC pays back in 10 months.
- CAC. For a 12-month payback at $100 a month, CAC has to come down to $960.
- Gross margin. Hosting, support and payment fees come off every month's revenue before it repays anything; the Stripe fee calculator shows the payment part.
- Billing. A yearly plan paid up front brings twelve months of revenue in on the first day, so the cash returns sooner than the formula's months suggest.
Rule of 40: what it is and when it applies
The rule of 40 says a SaaS company's growth rate and its profit margin should add up to at least 40%. Brad Feld heard it from a late-stage investor at a board meeting and wrote it up in February 2015: “The 40% rule is that your growth rate + your profit should add up to 40%.” Growing at 60% while losing 20% passes; growing at 20% while earning 20% passes too.
Rule of 40 formula and how to calculate it
Rule of 40 score = revenue growth rate (%) + profit margin (%)
Feld measured growth simply: “Just do year-over-year growth rate of monthly MRR.” For profit he took EBITDA margin: “I prefer to use EBITDA here as the baseline and then back test with the other percentages.” Acme grew MRR by 60%. Its revenue of $780,000 met $156,000 of cost of revenue, $480,000 of sales and marketing and $612,000 of everything else, an EBITDA margin of -60%. Its score: 60% + (-60%) = 0.
Rule of 40 in SaaS: the scale it was written for
Not at Acme's size, by the rule's own sources. Feld set the floor in the same post, for SaaS companies at scale: “assume at least $50 million in revenue”. Scale Venture Partners, which invests in early-stage enterprise software, writes in its primer: “Scale doesn’t use the Rule of 40 to evaluate prospective investment opportunities.” Bessemer applies its own version, the efficiency score (free cash flow margin plus ARR growth), from $25 million of ARR. Where the rule does apply, it holds up: among 68 public SaaS companies in Scale's database, with median annual revenue of $185 million, “The median across the last ten years has been 41%”. Below that scale, the burn multiple weighs the same trade-off between growth and cash.
SaaS efficiency metrics: magic number, burn multiple and quick ratio
SaaS magic number
The magic number measures sales efficiency: new annualized revenue for each dollar of sales and marketing. Scale Venture Partners, which says it developed the metric, builds it from GAAP revenue: “MN replaces Net New ARR in the numerator with the difference in GAAP revenue between the most recent periods.” In practice:
Magic number = (this quarter's revenue - last quarter's) × 4 ÷ last quarter's sales and marketing spend
Acme's revenue rose from $206,250 in its third quarter to $228,750 in the fourth, and it spent $120,000 on sales and marketing in the third: $22,500 × 4 ÷ $120,000 = 0.75. Scale's benchmark, across the 1,000+ growth-stage companies in its dataset in 2020: “the 0.7 figure is the median Magic Number (MN) for the dataset.”
Burn multiple
The burn multiple asks what growth costs in cash. David Sacks defined it in April 2020 as “Burn Multiple = Net Burn / Net New ARR”: in his words, “how much is the startup burning in order to generate each incremental dollar of ARR?” Acme burned $468,000 over the year and added $360,000 of net new ARR, a burn multiple of 1.3. On Sacks' scale for venture-stage startups that is great: under 1x amazing, 1x to 1.5x great, 1.5x to 2x good, 2x to 3x suspect and over 3x bad. Before the first revenue it does not apply: “Obviously if the startup is pre-revenue, net new ARR will be zero so the ratio will not even compute.”
SaaS quick ratio
The SaaS quick ratio sets the MRR gained against the MRR lost: (new MRR + expansion MRR) ÷ (churned MRR + contraction MRR). Acme gained $30,000 + $7,500 and lost $6,000 + $1,500, a quick ratio of 5.0: five dollars in for every dollar out. Above 1, MRR grows; below 1 it shrinks however many customers are won. Expansion revenue sits in the numerator, so upgrades lift the ratio as surely as new customers do.
SaaS metrics by stage: what to measure at 10, 100 and 1,000 customers
Most of the metrics above need customers before they mean anything. At Acme's price, the three stages look like this:
SaaS growth metrics at 10, 100 and 1,000 customers: what one cancellation moves and what to measure
| Customers | MRR at $100 | One cancellation moves monthly churn by | Measure | Leave for later |
|---|---|---|---|---|
| 10 | $1,000 | 10 points | MRR; every cancellation by name and reason; what you spent and where each customer came from | GRR, NRR, LTV, LTV:CAC, the magic number and the rule of 40: one customer swings each of them |
| 100 | $10,000 | 1 point | Customer churn and revenue churn by month; CAC, CAC payback and LTV:CAC by quarter; what new customers pay | GRR and NRR until a monthly cohort is a year old; the rule of 40 |
| 1,000 | $100,000 | 0.1 points | GRR and NRR by cohort; LTV:CAC by segment; the burn multiple and the magic number by quarter | The rule of 40, until revenue nears $25 million to $50 million |
MRR at Acme's $100 a month. One cancellation is counted against the customers at the start of a month. The $25 million is the ARR from which Bessemer applies its efficiency score, the $50 million the revenue Brad Feld set for the rule of 40.
With ten customers a single cancellation is a tenth of the base, so a retention figure says more about one account than about the business. ChartMogul left very young companies out of its 2025 retention report for that reason, counting those with $250,000 of ARR or more, since “historical retention data is less reliable for very early-stage startups with few customers coming up for renewal.” At Acme's price, a hundred customers is $10,000 of MRR, and that price band is the common road there: among 1,043 companies that reached $10k MRR, ChartMogul found “53% of companies hunted rabbits on the way to $10k MRR”, its word for accounts paying $30 to $299 a month. Finding those first customers has a guide of its own.
A SaaS metrics dashboard for a founder
A SaaS metrics dashboard worth opening every Monday fits on one screen: MRR and its four movements, customer churn and revenue churn for the month, CAC and CAC payback for the quarter, cash in the bank and the burn multiple; the stage table says when the rest earns a place. A billing system or a metrics tool computes most of it, and the formulas above are how to check the tool.
B2B SaaS metrics: contract value decides the benchmark
B2B SaaS metrics add what a sales team runs on: annual contract value (ACV), the length of the sales cycle, the pipeline and the win rate. ACV also decides which benchmarks apply, which is why SaaS Capital sorts its retention figures by it, and a median hides a wide spread inside each band. Its page on 2025 retention: “companies with ACVs between $25,000 and $50,000 show a median net revenue retention of 102% with the top quartile reporting NRR of 111% and the lowest quartile reporting 97% NRR.”
Where Submitator fits in your CAC
In the channel split above, Submitator is a one-time line. It takes the directory and launch-platform part of a launch off your hands. Your dashboard lists the catalog with a Domain Rating beside every site, and you choose from 202 sites, split into 81 launch platforms and startup directories, 62 on the SaaS and software list and 59 on the AI tools list. No site is added to the catalog before a submission to it has gone through from the first field to the last. Every directory gets a description we write for that site alone, cut to fit its form, and every form is filled in by us; the CAPTCHAs on them are handled on our side.
Badge directories want their badge shown on your site. One component, installed a single time, shows whichever badges your picks call for, and keeping each of them live afterwards is our job, with no commit per directory. Every listing we file then goes through IndexMachine, which Starter, Growth and Scale include as free indexing ($25 value). You have nothing to set up, and whether a page gets indexed remains Google's call.
Three plans, each a single payment and so a single line in your CAC, however you spread it: $29 buys Starter's 40 directories, $49 buys Growth's 70 (at $0.70 a directory, the best rate of the three) and $79 buys Scale's 110. All Directories costs $149 and covers the full catalog plus whatever joins it later. On timing, most directories are delivered in 24 to 48 hours, every directory within 21 days at most. Should a directory reject your product, you choose another in its place by default; All Directories has no replacement to give, since it holds every directory already. And a refund can be requested within 30 days of purchase, on the conditions in our terms. The startups we have submitted number 580+, and our submission statistics count how the filings from July to November turned out.
What it leaves out: no ads, no outreach. Go Publicly, a launch platform in the catalog, charges $19 a launch for Premium: a featured tag and highlighted card, priority placement in its Top Results and any launch day from tomorrow on. Our customers go there through the free launch, which counts as one pick of the 40 in the $29 plan. For the rest of a launch, week by week, see the product launch checklist.
Sources. Every benchmark: the note under the cheat sheet; each page was read on 16 November 2026, quotes are character-exact and each date is the page's own. Also read that day: SaaS Capital's bootstrapped benchmarks (24 Apr 2026, from the same 2026 survey of more than 1,000 companies), ChartMogul's ARR glossary and its report on the odds of reaching $1 million ARR (15 Oct 2025, 6,525 software companies), and Go Publicly's pricing page, which lists Premium at "$19 /launch" with "Featured tag & highlighted card", "Priority placement in Top Results" and "Launch any day from tomorrow", and Premium Plus at "$39 /launch". On our side: the 202 directories and launch platforms are the category-hub rows of the catalog exported on 17 November 2026, intercepts excluded; Go Publicly's free tier as our route is from its guide on this site; per-directory descriptions, form filling, CAPTCHAs, the badge component with its upkeep, pushing listings through IndexMachine and the rule that a site joins the catalog after one complete submission are as our founder describes them; prices, the indexing bonus and the 580+ count come from the pricing data behind our home page; the 21-day ceiling, the swap when a directory rejects you and the 30-day refund window are in our terms, and the founder gives the 24-to-48-hour figure for most directories. This guide promises no ranking, Domain Rating, traffic, signup or indexing result, and it claims no effect of a purchase on any metric above.
FAQ
What are the top 10 SaaS metrics?
For a young SaaS company, these ten: MRR, ARR, customer churn rate, revenue churn, gross revenue retention, net revenue retention, CAC, LTV:CAC, CAC payback and burn multiple. The other five in this guide (ARPA, LTV, quick ratio, magic number and rule of 40) feed into those or matter later. Each one has its formula, a worked example and a sourced benchmark in the cheat sheet above.
What is a good GRR for SaaS?
Around 90% or better: that is the median for private B2B SaaS. By SaaS Capital's 2025 figures it runs from 90% where customers pay under $12,000 a year to 95% above $250,000; bootstrapped companies of $3 million to $20 million ARR reported 91% in its 2026 survey, and Bessemer's cloud portfolio held 85% to 90% across revenue ranges (2021). Because GRR leaves upgrades out, 100% is its ceiling.
What is a good LTV to CAC ratio?
Three or higher. David Skok wrote in 2013 that the best SaaS businesses have an LTV to CAC ratio above 3, sometimes as high as 7 or 8, and Bessemer recommended in 2021 investing in customer acquisition at 3x or more. Check CAC payback beside it: the made-up company in this guide has a ratio of 5.0 and still takes 20 months to earn back each customer's cost, because a long customer lifetime lifts LTV without bringing the cash back any sooner.
How much CAC is good?
A good CAC is measured against what a customer pays, not in dollars. Divide CAC by a new customer's monthly gross profit to get the payback: Bessemer's 2021 targets were under 12 months for small-business buyers, 18 for mid-market and 24 for enterprise. At $100 a month and 80% gross margin, a 12-month payback allows a CAC of up to $960. LTV should come to at least 3 times CAC. As a share of revenue, the median private B2B SaaS company in SaaS Capital's 2026 survey spent 15% of ARR on selling and 8% on marketing.
Are ARR and run rate the same?
Usually. ChartMogul's glossary says that in modern SaaS, ARR usually means annualized run rate: MRR × 12. The stricter annual recurring revenue counts contracts of a year or longer and leaves monthly plans out, so a company with mostly monthly customers has a run rate far above its strict ARR. A run rate built from total revenue, one-time fees included, is neither: one-time fees stay out of ARR in both senses.
What is a good MRR for a startup?
One that grows month after month; no single figure is good at every age. ChartMogul's data gives two markers. Of 1,043 companies that reached $10,000 MRR, 53% got there with customers paying $30 to $299 a month, roughly 100 customers at $100 (published June 2026). And among startups that reached $1 million ARR, about $83,000 of MRR, it usually took two to five years from the first revenue; 3.3% got there within a year (October 2025, 6,525 companies).
Is the rule of 40 still relevant?
For SaaS companies at scale, yes: across 68 public SaaS companies in Scale Venture Partners' database, the median rule of 40 score over the ten years to March 2020 was 41%. Below the $50 million of revenue Brad Feld wrote it for in 2015, it says little, and Scale does not use it to judge early-stage investments. Early on, the burn multiple and CAC payback weigh the same trade-off more usefully.
Keep reading
Their $19: one launch. Our $29: 40 directories.
Go Publicly charges $19 for each Premium launch: a featured card and the day you choose. Submitator files 40 directories you pick for $29, or 110 for $79, paid once, with a description written for each one and every badge kept live.
