SaaS metrics: the 15 numbers that matter, how they connect and how investors weigh them

Every SaaS board meeting is, at some level, a translation exercise. The founder knows the business as a set of customers, products and people; the investor knows it as a set of ratios. SaaS metrics are the shared language in between, and like any language they have a grammar: some words only make sense in combination, some mean different things to different speakers, and some have quietly changed meaning over the past decade.

This guide is the grammar. It explains how fifteen widely used SaaS metrics connect, which ones carry weight at each stage of a company’s life, how investors’ attention has moved from one to another since 2008, and where the metrics travel well beyond classic B2B software and where they break. Each metric has its own guide with a calculator and dated benchmarks, linked as it comes up.

Five questions, fifteen metrics

An investor reading a SaaS company works through roughly five questions, in order.

How big is the recurring business? That is MRR, ARR and, for companies with long contracts, remaining performance obligations. The ARR vs MRR guide covers the conversion and the run-rate trap.

Do customers stay and grow? That is the churn rate, gross revenue retention, net revenue retention and the SaaS quick ratio.

What does a new customer cost? That is CAC payback, the SaaS magic number and the LTV:CAC ratio.

What does growth cost the whole company? That is the burn multiple, the Rule of 40 and revenue per employee.

What is it worth? That is the valuation multiple the market puts on all of the above.

The order matters. Acquisition efficiency means little if customers leave; valuation multiples mean little if growth is bought at any price. Most arguments about which SaaS metric matters most are really arguments about which question is most in doubt for a particular company.

1. 01

   Scale

   How big is the recurring business?

   • MRR
   • ARR
   • ARR vs MRR
   • RPO
2. 02

   Retention

   Do customers stay and grow?

   • SaaS churn rate
   • GRR
   • NRR
   • SaaS quick ratio
3. 03

   Acquisition

   What does a new customer cost, and when is it paid back?

   • CAC payback period
   • SaaS magic number
   • LTV:CAC ratio
4. 04

   Efficiency

   What does growth cost the whole company?

   • Burn multiple
   • Rule of 40
   • Revenue per employee
5. 05

   Value

   What is it all worth?

   • SaaS valuation multiples

The revenue spine: MRR, ARR and RPO

Everything starts with recurring revenue, and recurring revenue starts with a definitional problem. The SaaS Metrics Standards Board keeps it simple: “ARR is the sum of subscription recurring revenue on an annualized basis” (SaaS Metrics Standards Board). Real companies are messier. Samsara’s annual report states the problem outright: “ARR does not have a standardized meaning and is not necessarily comparable to similarly titled measures presented by other companies” (Samsara, 10-K).

The metric’s rise is visible in securities filings. The phrase “annual recurring revenue” appeared in 118 annual reports filed with the SEC in 2025, up from 42 in 2020 and 10 in 2015 (our EDGAR full-text searches). Popularity has brought looseness. Jamin Ball, in March 2024: “ARR (Annual Recurring Revenue) is one of the most popular SaaS (Non-GAAP) metrics. However, it’s also one of the most loosely used metrics, and is frequently misused” (Clouded Judgement, 2024). Jason Lemkin is blunter: “Most B2B companies count everything as “ARR”, right or wrong” (SaaStr, 2025).

The sharpest fight is over run-rate. In April 2026 Y Combinator’s Garry Tan told founders that “Annual Revenue Run Rate is never abbreviated to ARR”, reserving the acronym for contracted recurring revenue (Artificial Lawyer, 2026). A month later TechCrunch described AI startups presenting annualized run-rate figures under the same acronym (TechCrunch, 2026), a report that relies partly on anonymous sources.

MRR is the monthly view of the same thing, and it is the better instrument for companies whose customers pay month to month: Shopify still reports MRR in its filings (Shopify, 10-Q). Its real value is the MRR bridge, the monthly walk from new and expansion revenue to contraction and churn, which feeds every retention metric below.

At the other end of the spectrum sits the only metric in this family with an accounting definition. Remaining performance obligations come from the revenue standard, which requires public companies to disclose “The aggregate amount of the transaction price allocated to the performance obligations that are unsatisfied (or partially unsatisfied)” (FASB, 2014). Salesforce turned the twelve-month slice into a headline metric in 2018, calling it “current remaining performance obligation” (Salesforce, 2018). RPO shows how much revenue is locked in, and how far out: Oracle expects only about 13% of its $664 billion to become revenue within twelve months (Oracle, 10-Q).

Retention: the number underneath every other number

If recurring revenue is the spine, retention is the bone density. David Skok explained why in 2013: “the churn rate, combined with the rate of new ARR adds, not only defines how fast you can grow the business, it also defines the maximum size the business can reach” (Skok, 2013).

Churn compounds in ways intuition misses. ChartMogul’s rule of thumb: “A monthly churn rate of 5% corresponds to an annual churn rate of 46%” (ChartMogul, 2022). That is why investors convert everything to annual retention ratios. Gross revenue retention counts what a company keeps from last year’s customers before any expansion; net revenue retention adds expansion back in.

The benchmarks have softened. The median private B2B SaaS company kept 91% of its revenue gross and 101% net in SaaS Capital’s 2025 survey of more than 1,000 companies (SaaS Capital, 2025). Benchmarkit’s median gross retention fell from 88% to 84% in 2025, its biggest one-year drop (Benchmarkit, 2026). Among public software companies, median net retention was 110% in September 2026 (Clouded Judgement, Sep 2026), down from 120% in June 2022 (Clouded Judgement, Jun 2022).

Investors now read the two together. Lemkin, in May 2026: “And watch the GRR/NRR gap. If your GRR is 92% and NRR is 95%, the gap tells you exactly how thin the magic actually is” (SaaStr, 2026). A high NRR built on a few large upsells can hide a broad base of customers quietly leaving; gross retention exposes it.

The SaaS quick ratio packs the whole monthly bridge into one number: revenue added divided by revenue lost. Mamoon Hamid introduced it at SaaStr Annual in 2015 with a simple instruction, “Maintain a Quick Ratio > 4” (Hamid, 2015). On today’s typical growth and retention, few mature companies get near that; the ratio is most useful early, when the base is small and the bridge moves fast.

What a customer costs: CAC payback, magic number and LTV:CAC

The acquisition metrics are the oldest part of the SaaS toolkit, and they are close cousins.

CAC payback asks how many months of gross profit a new customer needs to repay what it cost to win. Skok set the first rule of thumb in 2010: “My own rule says that startups need to recover their cost of customer acquisition in less than 12 months” (Skok, 2010). Bessemer refined it by segment: “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” (Bessemer, 2021). Reality runs longer: public software’s median payback was 30.5 months in Q2 2026, up from 17.2 in Q3 2021, on Meritech’s measure (Meritech, 2026).

The magic number asks the same question at company level: how much new annual revenue each dollar of last quarter’s sales and marketing produced. Lars Leckie’s 2008 rule still frames it: “if you are below 0.75 then step back and look at your business, if you are above 0.75 then start pouring on the gas for growth” (Leckie, 2008). Across more than 1,000 enterprise SaaS companies, Scale Venture Partners puts the long-term median at 0.7 (Scale, 2021).

The two are linked by arithmetic. At an 80% gross margin, Dave Kellogg notes, CAC payback in months is roughly 15 divided by the magic number (Kellogg, 2023), so a magic number of 1 means a 15-month payback.

LTV:CAC asks about the whole customer lifetime rather than the payback window. Skok’s 2009 bar remains the reference: “It appears that LTV should be about 3 x CAC for a viable SaaS or other form of recurring revenue model” (Skok, 2009). It is also the most contested of the three, because lifetime is a forecast. Tomasz Tunguz warned in 2017: “But this figure is often meaningless for early stage startups” (Tunguz, 2017). Stage 2 Capital’s Jay Po went further in 2025: “LTV:CAC has been around forever. It’s simple to calculate. It’s easy to talk about. But it’s obsolete” (Stage 2 Capital, 2025).

Payback has largely won that argument for cash planning, because it needs no lifetime forecast. Kyle Poyar’s reading of the High Alpha data puts it alongside retention as a predictor: “The two strongest predictors of long term and profitable growth are CAC payback period and net revenue retention (NRR)” (Poyar, 2025).

What growth costs the company: burn multiple, Rule of 40, revenue per employee

Acquisition metrics look at sales and marketing. The efficiency metrics look at everything.

The burn multiple is the cash a company burns for each dollar of net new ARR. David Sacks introduced it as the COVID downturn hit in April 2020: “Burn Multiple = Net Burn / Net New ARR”, adding that “The beauty of the Burn Multiple is that it’s a catch-all metric” (Sacks, 2020). It is the right lens for a cash-burning company; once a company is profitable, it stops meaning much.

The Rule of 40 is the lens for scaled companies: growth rate plus profit margin should reach 40%. Brad Feld popularized it in 2015 (Feld, 2015) and revisited it in 2026: “I still like it. It’s a clean way to compress two things that usually fight each other - growth and profitability - into one number” (Feld, 2026). It is a high bar: across more than 200 software companies from 2011 to 2021, businesses cleared it only 16% of the time, McKinsey found (McKinsey, 2021).

Revenue per employee is the bluntest efficiency measure and the one AI has disrupted most. The median private SaaS company reached $141,125 per employee in 2026 (SaaS Capital, 2026). Lemkin’s thresholds: “The old good number was somewhere around $200K-$250K per employee. Great companies hit $300K” (SaaStr, 2026). Cursor reported crossing $1 billion of annualized revenue with “a team of over 300” in November 2025 (Cursor, 2025).

What it’s all worth: valuation multiples

Every metric above ultimately feeds one number: what the market pays for a dollar of revenue. That number has been on a wild ride. The median public cloud company traded at 20.8x next-12-months revenue in February 2021 (Clouded Judgement, Feb 2021) and at 4.2x in September 2026, when companies growing over 22% commanded a median 18.0x against 3.3x for those growing under 15%.

Retention earns a premium too. Software Equity Group found that public companies with net retention above 110% traded at “a median 8.0x EV/TTM revenue”, “a 77.8% valuation premium relative to the 90%-110% cohort” (SEG, 2026). SaaS Capital’s Randall Lucas summed up the new climate: “It’s a “rich get richer” valuation environment, and merely being a SaaS company is no longer a ticket to premium ARR multiples” (SaaS Capital, 2025).

How investors’ weight has shifted

The SaaS metric set was assembled remarkably quickly. Between 2008 and 2013, Leckie’s magic number, Skok’s LTV:CAC and payback rules, and Bessemer’s cloud metrics gave founders and investors a shared operating vocabulary. In 2015 Feld’s Rule of 40, Battery Ventures’ “triple, triple, double, double, double” growth path (TechCrunch, 2015) and Hamid’s quick ratio added a growth scorecard.

The next phase belonged to retention. Eloqua’s 2012 prospectus is the earliest SEC filing we found to use “net dollar retention” (Eloqua, 2012), and by 2016 Twilio was reporting a dollar-based net expansion rate of 155% in its IPO filing (Twilio, 2016). Through the low-rate years net retention became the signature of a great SaaS company, and in 2021 growth was priced at almost any cost.

Then the reset. When money stopped being free, the metrics that measure cost came forward. The a16z growth team reported that after 2022 “magic numbers fell below 0.75” (a16z, 2023). KeyBanc and Sapphire wrote that “The Rule of 40 is more important to valuation than it was in the free money era” (KeyBanc & Sapphire, 2023). Bessemer proposed the Rule of X, weighting growth two to three times more than profit (TechCrunch, 2023), a reminder that markets still pay most for growth even when they demand discipline.

Since 2024 the weight has moved again, towards the quality of revenue. Ray Rike found growth and the Rule of 40 explaining valuation multiples better than NRR (Rike, 2024). ICONIQ called the Rule of 40 “the strongest predictor of valuation” (ICONIQ, 2025). Lemkin relayed in 2025 that pre-IPO investors “are now more interested in GRR” than in net retention (SaaStr, 2025). And in 2026 AI began rewriting the denominators, from revenue per employee to what counts as recurring.

1. 2008–2013

   The operating playbook

   • LTV:CAC ratio
   • CAC payback period
   • SaaS magic number
   • MRR

   David Skok: "It appears that LTV should be about 3 x CAC for a viable SaaS or other form of recurring revenue model." (Skok, 2009)

2. 2015

   Growth gets a scorecard

   • Rule of 40
   • ARR
   • SaaS quick ratio

   Brad Feld popularizes the Rule of 40: "The 40% rule is that your growth rate + your profit should add up to 40%." (Feld, 2015)

3. 2016–2021

   The retention era

   • NRR
   • ARR

   Net retention moves into IPO prospectuses; Twilio's reports a dollar-based net expansion rate of 155% for 2015. (Twilio, 2016)

4. 2020–2021

   Growth at almost any price

   • ARR
   • NRR

   In February 2021 the median public cloud company traded at 20.8x next-12-months revenue and the top five at 61.5x. (Clouded Judgement, Feb 2021)

5. 2022–2023

   The efficiency reset

   • Burn multiple
   • Rule of 40
   • SaaS magic number

   KeyBanc and Sapphire: "The Rule of 40 is more important to valuation than it was in the free money era". (KeyBanc & Sapphire, 2023)

6. 2024–2025

   Profit, payback and gross retention

   • Rule of 40
   • GRR
   • CAC payback period

   ICONIQ calls the Rule of 40 "the strongest predictor of valuation" among the metrics it tracks for public software companies. (ICONIQ, 2025)

7. 2026

   AI rewrites the denominators

   • Revenue per employee
   • ARR
   • SaaS churn rate

   Tomasz Tunguz: "Five years ago, $100K ARR per employee was standard for SaaS startups." AI-native companies now report millions per head. (Tunguz, Feb 2026)

Which metrics matter at which stage

Seed. There is not enough history for ratios to mean much. Lemkin’s guidance is to watch growth and keep churn falling: NRR and churn “aren’t necessarily statistical significant before Year 2-3 and before $1m-$2m in ARR” (SaaStr, 2024). MRR growth, logo churn, the quick ratio and runway carry the weight.

Series A and B. The company now has cohorts and a sales motion, so efficiency questions arrive: burn multiple, CAC payback and the magic number. Craft Ventures built its Series A metric set around them (Craft Ventures, 2021). Net retention becomes measurable and starts to drive valuation.

Growth stage. Gross and net retention, payback by segment and the Rule of 40 dominate. Investors want to see efficiency improving with scale.

Public. The Rule of 40, net retention, RPO and forward revenue multiples lead, with revenue per employee as a check on discipline.

How much each metric matters, by stage

Metric	Seed	Series A–B	Growth stage	Public
MRR	Core	Core	Useful	Useful
ARR	Useful	Core	Core	Core
RPO	Rarely used	Rarely used	Useful	Core
SaaS churn rate	Core	Core	Useful	Useful
GRR	Rarely used	Useful	Core	Core
NRR	Rarely used	Core	Core	Core
SaaS quick ratio	Core	Useful	Rarely used	Rarely used
CAC payback period	Useful	Core	Core	Useful
SaaS magic number	Rarely used	Core	Core	Useful
LTV:CAC ratio	Rarely used	Useful	Useful	Rarely used
Burn multiple	Useful	Core	Core	Rarely used
Rule of 40	Rarely used	Rarely used	Core	Core
Revenue per employee	Rarely used	Useful	Useful	Core
SaaS valuation multiples	Rarely used	Useful	Core	Core

Beyond B2B SaaS: where the metrics travel and where they break

The SaaS metric set was built for subscription software sold to businesses, but it has spread to almost every recurring-revenue model. Some metrics travel well; others need translation.

Business model	What travels	What breaks or needs care
SMB and self-serve software	MRR, logo churn, quick ratio	Churn runs high: ChartMogul found median monthly churn of 6.1% for accounts paying under $25 a month
Enterprise software	ARR, GRR, NRR, RPO	Long payback is normal: Bessemer’s enterprise target is under 24 months
Usage-based and infrastructure	NRR, revenue multiples	ARR and NRR swing with usage, and RPO misses usage above commitments
AI-native products	Revenue per employee, burn multiple	Retention can be far lower and ARR definitions are contested
Vertical SaaS	LTV:CAC, payback	Smaller markets raise acquisition costs
Marketplaces and payments	MRR on subscription fees	Gross revenue makes ratios like revenue per employee incomparable

The details matter. Low-priced subscriptions churn fastest: ChartMogul’s data showed median monthly customer churn falling from 6.1% for accounts paying under $25 a month to 2.2% above $500. Enterprise contracts retain best: companies with contracts above $250,000 a year kept 95% of revenue gross in SaaS Capital’s survey.

Usage-based companies are the hardest fit. Snowflake warns that “RPO is not necessarily indicative of future product revenue growth because it does not account for the timing of customers’ consumption or their consumption of more than their contracted capacity” (Snowflake, 2026). Confluent warns that its method “may result in increased volatility in NRR” (Confluent, 10-K). Datadog counts usage in ARR (Datadog, 10-K); others count only committed subscriptions.

AI-native products stretch the set further. Kyle Poyar’s analysis of ChartMogul data found AI-native companies’ median gross revenue retention rising from 27% in January to 40% in September 2025, and noted that “AI-native products that sell for <$50 per month see just 23% GRR and 32% NRR. This is 20 points worse than either B2B or B2C SaaS” (Poyar, 2025). At the other extreme, AI-native leaders post revenue per employee that classic SaaS never approached. Vertical software sits in between: Benchmarkit’s vertical SaaS companies had a median CAC payback of 18 months against 14 for horizontal, but a higher median LTV:CAC, 5.6x against 4.1x.

And payments-heavy businesses need the most care of all. When Block cut 40% of its workforce, Tunguz noted that “Revenue per employee jumped 67%” (Tunguz, 2026), but Block reports gross revenue, so its level isn’t comparable with software ARR per head.

The identities that hold it together

A handful of identities connect the metrics, and they are the fastest way to sanity-check a deck.

• ARR = 12 × MRR. If a company’s ARR and MRR don’t reconcile, one of them is defined differently.
• Gross revenue churn ≈ 1 − GRR, and net revenue churn ≈ 1 − NRR, so NRR above 100% means negative net churn.
• Quick ratio = 1 + net new MRR ÷ lost MRR, so it falls as the base grows and losses mount.
• Quick ratio = 1 + growth ÷ (1 − GRR), with growth as net ARR growth, when all three cover the same period and the ratio’s losses are the churn and contraction GRR counts.
• CAC payback ≈ 15 ÷ magic number at an 80% gross margin.
• LTV:CAC = customer lifetime ÷ payback period, when LTV is measured on gross margin.
• Burn multiple = 1 ÷ Bessemer’s efficiency score, the same ratio upside down.
• Valuation = multiple × revenue, so moving from low to high growth changes value far more than adding revenue does.

When a set of reported metrics breaks one of these, something is being measured differently, and the difference is usually where the story is.

Fireside: life in the metric realm

It is worth stepping back from the formulas to notice what kind of thing SaaS metrics are. They are not accounting. Christoph Janz said it plainly in 2017: “Since there is no official US-GAAP definition of churn or retention, different companies use different ways to measure and report these metrics” (Janz, 2017). The SEC’s 2020 guidance on key performance indicators asks public companies to define metrics like these and explain how they are calculated (SEC, 2020); private companies face no such discipline.

That makes SaaS metrics something closer to a negotiated language between founders and investors. Each era’s vocabulary reflects what investors were most worried about: whether a new business model could work (LTV:CAC, payback), whether customers would stay (NRR), whether growth was being bought too dearly (burn multiple, Rule of 40), and now whether a dollar of AI revenue is really recurring at all. Founders adopt the vocabulary that flatters them; investors respond by inventing a stricter one. Ball’s “ERR”, experimental run-rate revenue, is the latest example of a term coined to separate durable revenue from the rest.

The benchmarks deserve the same scepticism. Every median in this guide comes from a particular sample, defined a particular way, at a particular moment. Benchmarkit’s medians differ from SaaS Capital’s, which differ from public-company figures built on revenue × 4. Use benchmarks to locate yourself, not to set targets; a company aiming at someone else’s median is optimizing for the wrong audience.

What has proved durable is the habit of reading metrics in pairs: growth with retention, growth with burn, NRR with GRR, payback with lifetime value. The single-number summaries, the Rule of 40 and the burn multiple, are themselves pairs compressed into one. The founders who handle investor scrutiny best tend to be the ones who show both halves before anyone asks.

A page-one dashboard

If we had to put five numbers on the first page of a board pack for a growing SaaS company, they would be these, each with its definition written underneath (our view):

1. ARR, with how it is calculated and what share is usage-based.
2. Gross and net revenue retention, side by side, by cohort.
3. CAC payback, by segment, on gross margin.
4. Burn multiple, over the last four quarters.
5. Rule of 40, once the company is large enough for profit to mean something.

Everything else in this guide supports, explains or stress-tests those five.