The median SaaS company now recovers its customer-acquisition cost in 16 months. The top quartile does it in 6 or less — and that gap is quietly becoming a competitive disadvantage for everyone sitting at the median.
What are SaaS metrics?
SaaS metrics are the recurring-revenue measurements — MRR, ARR, CAC, LTV, churn, NRR, and efficiency ratios like the Rule of 40 — that subscription businesses track because they bill customers monthly or annually instead of once. Many of them don’t appear in a traditional profit-and-loss statement at all.
A one-time software sale shows up on a P&L the month it happens and is finished. A subscription doesn’t work that way. SaaS metrics connect financial performance with customer behavior, revealing how product and pricing decisions affect long-term results — which is exactly the information a standard income statement was never built to show.
Take two companies with identical monthly revenue. One is gaining customers and losing almost none. The other is losing as many customers as it gains, replacing them month after month just to stay flat. That difference only shows up in metrics like churn and Net Revenue Retention — never in revenue alone.
| Category | What it shows | Core metrics |
|---|---|---|
| Revenue | What’s coming in right now | MRR, ARR, ARPA |
| Acquisition | What it costs to bring a customer in | CAC, CAC payback |
| Value | Whether that customer was worth the cost | LTV, LTV:CAC |
| Retention | Whether they stay | Churn, GRR, NRR |
| Efficiency | Whether the engine is sustainable | Magic Number, Burn Multiple, Rule of 40 |
Each category answers a question the others can’t. A healthy MRR with a collapsing NRR is a company refilling a leaking bucket. A strong LTV:CAC with a 30-month payback period is profitable on paper and out of cash in practice.
The rest of this guide goes through each category — what it measures, how to calculate it correctly, and where competing guides get the formula wrong, starting with the revenue metrics every other number in this article builds on.
How do MRR and ARR work?
MRR (Monthly Recurring Revenue) is the normalized, predictable subscription income a SaaS business collects each month, excluding one-time fees. ARR is MRR multiplied by 12 — the same revenue expressed as a yearly figure for board reporting and investor conversations.
MRR is used to measure a business’s growth rate by comparing one period to the next — a business with $100,000 MRR this month and $120,000 next month has a 20% growth rate.
Worked example — the company we’ll track through this entire article: 200 customers, $100,000 MRR, $500 average revenue per account (ARPA), 80% gross margin. ARR = MRR × 12 = $1,200,000.
What’s the difference between ARPA and ACV?
ARPA and ACV measure the same relationship at different time scales, but neither is MRR, and neither is ARR. ARPA is how much one customer pays per month; ACV is the yearly value of a single contract.
| Metric | Formula | For our example company |
|---|---|---|
| ARPA | MRR ÷ active accounts | $100,000 ÷ 200 = $500/month |
| ACV | Contract value ÷ contract years | A 1-year, $6,000 contract → $6,000/year |
A widely read SaaS metrics guide by Maxio defines its MRR formula as “sum of current monthly fees ÷ number of active customers.” That formula doesn’t produce MRR — dividing total monthly fees by customer count is the ARPA calculation. Its own worked example (200 customers × $50/month = $10,000) confirms the error.
Why does MRR move each month?
MRR moves because four components change it every month: New Business MRR, Expansion MRR, Contraction MRR, and Churned MRR. ChartMogul’s own benchmarking tooling tracks a fifth movement, Reactivation MRR, separately.
Want the full retention picture behind these expansion numbers, including the NRR/GRR split by ARR tier?
Read the NRR vs ARR Guide →How do you calculate CAC and CAC payback correctly?
CAC is the total sales and marketing cost to acquire one customer, not marketing spend alone. CAC payback is the number of months of gross-margin-adjusted revenue it takes to recover that cost. The median B2B SaaS company recovers CAC in 16 months; the top quartile does it in under 6.
Why does CAC get calculated wrong so often?
CAC gets calculated wrong because several widely used guides leave sales costs out of the formula entirely, counting only ad spend. The full, correct formula must include:
- Sales team salaries and commissions
- Marketing team salaries and ad spend
- Agency, freelance, and sales tooling costs
- A proportional share of marketing infrastructure
ChartMogul’s own SaaS metrics cheat sheet defines CAC as marketing spend divided by new customers — no sales cost included. A marketing-only CAC understates the real cost of growth, often significantly, in any company with a sales team.
Worked example, same company: $300,000 in combined sales and marketing spend this month, 60 new customers. CAC = $300,000 ÷ 60 = $5,000.
How do you calculate CAC payback, and what’s a good number?
CAC payback is calculated by dividing CAC by the gross-margin-adjusted monthly revenue a new customer generates, and a good 2025 benchmark is under 18 months, with under 12 months marking top-tier efficiency. CAC payback = $5,000 ÷ ($500 × 0.8) = 12.5 months for our example company.
Does CAC payback differ by growth rate, deal size, or market type?
CAC payback differs significantly by growth rate, deal size, and market type — enough that comparing your number to the flat 16-month median can be actively misleading. Companies growing 21–30% YoY post the slowest payback at 22 months — the counterintuitive outlier in the dataset. Horizontal SaaS recovers CAC in 14 months versus 18 for vertical SaaS, though vertical SaaS earns it back with a 5.6x LTV:CAC ratio versus 4.1x for horizontal.
A 22-month payback isn’t automatically a red flag — check your growth-rate cohort and your market type before deciding whether your number is a problem or just the expected shape of your segment, which is exactly the question Section 4 answers next.
How do you calculate LTV, and is 3:1 LTV:CAC still the target?
LTV is the gross-margin-adjusted revenue a customer generates over their relationship with you, divided against CAC to get the LTV:CAC ratio. The median SaaS CLTV:CAC ratio is 4.1x in 2025, above the classic 3:1 rule, with the top quartile reaching 7.8x — but 3:1 as a target comes with a condition most guides leave out.
How do you calculate LTV correctly?
LTV is calculated correctly by multiplying average revenue per account by gross margin, then dividing by the churn rate — never by revenue alone. Using our example company: LTV = ($500 × 0.8) ÷ 0.02 = $20,000. LTV:CAC = $20,000 ÷ $5,000 = 4.0x, right at the current 4.1x median.
Stripe’s own guide defines this relationship backwards: “CAC-to-LTV ratio… A CAC-to-LTV ratio of less than one is considered ideal.” Every other source — ChartMogul, Benchmarkit, Aleph, David Skok’s original framework — expresses this as LTV:CAC (value over cost, target ≥3), not CAC:LTV (cost over value). A reader who benchmarks their “0.25 CAC:LTV” against every other guide’s “4:1 LTV:CAC” standard compares the wrong numbers.
Is 3:1 LTV:CAC still a good target in 2026?
3:1 is still cited as the minimum, but it was never meant to apply to the margin-adjusted ratio everyone now benchmarks against. David Skok’s original guidance states: “Our guideline… should be higher than 3. (The guideline assumes you are using the simpler LTV formula that does not include a Gross Margin adjustment, and that you have a Gross Margin of 80% or higher.)”
Unit Economics Calculator
Always state which formula produced your LTV:CAC number before comparing it to anyone else’s — the gap between the two conventions is exactly where guides go wrong.
How do you measure churn, GRR, and NRR?
Customer churn tracks accounts lost, revenue churn tracks dollars lost, GRR shows how much revenue survived without counting expansion, and NRR adds expansion back in. The 2025 median GRR is 84%, median NRR is 102% — an 18-point spread that shows how much expansion is masking churn.
How do you calculate customer churn and revenue churn?
Customer churn is calculated by dividing customers lost by customers at the start of the period; revenue churn does the same math with dollars. For our example company: 4 customers lost out of 200 = 2.0% customer churn. $1,000 churned + $2,000 contraction out of $100,000 starting MRR = 3.0% revenue churn.
Never multiply monthly churn by 12 to get an annual figure — compound it instead: Annual retention = (1 − monthly churn)12. At 2% monthly churn, the annual figure is approximately 21.5%, not 24%.
How do you calculate GRR and NRR?
GRR = (Starting MRR − Churned − Contraction) ÷ Starting MRR = 97% for our example. NRR adds expansion back in: ($100,000 − $1,000 − $2,000 + $4,000) ÷ $100,000 = 101%. The 2025 median GRR is 84% (top quartile 91%, down from 88%); median NRR is 102%.
What is a good NRR in 2026, and why do sources disagree?
A good NRR depends entirely on which sample you’re comparing yourself to. ChartMogul’s 82% median (Dec 2025) covers a broad billing-data panel skewed toward smaller companies. Aleph × Benchmarkit’s 102% median reflects larger, venture-backed survey respondents. A “115–125% elite” figure circulating in some guides traces back to ChartMogul’s 2023 best-in-class figure for one narrow segment, not a 2026-wide benchmark.
Does AI-native SaaS retain customers differently?
AI-native SaaS retains customers dramatically worse than traditional B2B SaaS: median GRR was 40% and median NRR was 48%, against a traditional B2B SaaS median of 82% NRR — though the gap is closing fast as early “AI tourists” churn out.
Retention Calculator
Report GRR next to NRR, always — a healthy NRR on a weak GRR means expansion is covering for churn elsewhere.
See the Full NRR vs ARR Breakdown →What do the Magic Number, Burn Multiple, and Rule of 40 measure?
All three measure efficiency, not growth alone. The Magic Number measures sales-and-marketing efficiency, Burn Multiple measures total cash efficiency, and the Rule of 40 measures whether growth and profitability together clear a minimum bar. The 2025 median Magic Number is 1.37, crossing the 1.0 “invest” line for the first time in years.
How do you calculate the SaaS Magic Number?
The Magic Number is calculated as [(Current Q Revenue − Prior Q Revenue) × 4] ÷ Prior Q S&M Spend. Coined in 2005 by Rory O’Driscoll at Scale Venture Partners, it measures how much new annualized revenue each dollar of S&M generates.
Stripe’s own metrics guide defines the Magic Number as “customer lifetime value (CLV) divided by customer acquisition cost (CAC).” That’s the LTV:CAC ratio under the wrong name — a completely different formula.
Below 0.75: step back and re-examine go-to-market. Above 0.75: start scaling. Above 1.5: potential under-investment. The median rose from 0.94 in 2024 to 1.37 in 2025.
How do you calculate Burn Multiple?
Burn Multiple = Net Burn ÷ Net New ARR, with lower numbers being better. Under 1.0x is amazing, 1.0–1.5x great, 1.5–2.0x good, 2.0–3.0x suspect (David Sacks’ framing). Aleph × Benchmarkit does not publish a 2026 median for this metric — use these thresholds directly rather than an unconfirmed market figure.
How do you calculate the Rule of 40?
Rule of 40 Score = Revenue Growth Rate (%) + Profit Margin (%). Private median is 25% (up from 15%, top quartile 43%); public median is 22.6% on an EBITDA basis versus 39.1% on a free cash flow basis — a 16-point swing from the margin definition alone.
To check where your own numbers land, this free Rule of 40 calculator scores your growth and margin against 2026 benchmarks.
Calculate Your Score →How is AI changing SaaS gross margins and ARR?
AI is compressing gross margins for AI-heavy products specifically — not the median SaaS company yet — and it’s breaking the definition of ARR itself, since usage-based and contracted-but-undelivered AI revenue don’t behave like the predictable subscriptions ARR was built to measure.
Is AI actually compressing SaaS gross margins?
Benchmarkit’s 2026 report finds median software gross margin holding above 80%, stable across four years — industry-wide AI costs haven’t compressed the median yet. But for AI-native products, ICONIQ’s Q2 2026 survey of 305 executives found average AI product gross margins reached 53% in 2026, up from 45% in 2025.
Gross margin is an input to LTV, CAC payback, and the Rule of 40 — track AI COGS as its own line, separate from traditional hosting costs, or these other metrics will quietly misstate what’s happening.
Does revenue from AI features count as ARR the same way subscription revenue does?
No — this is a genuine, unresolved dispute in SaaS finance. TechCrunch reported in May 2026 that AI startups are reporting “Contracted ARR” (CARR) as if it were ARR, with CARR running up to 70% higher than actual recognized revenue in some cases.
If you’re evaluating a SaaS company’s growth, ask whether their “ARR” is contracted-and-delivered or includes CARR and trailing usage. For AI infrastructure cost and risk, see The AI Data Center Bubble and What Is AI FinOps? for how teams track this spend.
What is a good ARR per employee?
A good ARR per employee in 2026 depends on which benchmark you compare against — the median is $193K in a venture-heavy sample, up 29% from $150K, while the median across all private SaaS companies is $141,125, up from $129,724.
ARR per employee = Total ARR ÷ Full-time employees. For our example company: $1,200,000 ÷ 8 FTEs = $150,000.
Benchmarkit’s own summary page headlines $175K, up 17%, for the same panel its detailed per-metric page reports as $193K, up 29%. This article uses $193K, the figure on the report’s dedicated per-metric page.
Compare yourself against the source closest to your own company — a bootstrapped $5M-ARR business should use SaaS Capital’s $141,125 figure, not Aleph’s venture-heavy $193,420, or it will look artificially behind a benchmark that was never measuring companies like it.
Which SaaS metric should you fix first at your stage?
The right metric to fix depends on your ARR stage, not a universal checklist — pre-seed companies should fix churn, Series A companies should fix CAC payback and GRR, Series B companies should fix NRR and the Magic Number, and growth-stage companies should fix the Rule of 40 and ARR per employee.
Every metric in this article connects to the others — gross margin feeds LTV, LTV feeds LTV:CAC, retention feeds NRR, and NRR feeds the Rule of 40. The Metric Checkpoint is a four-stage sequence: at each ARR stage, one or two metrics matter more than the rest, and fixing those first makes every metric after them easier to move.
A pre-seed company with 12 customers can’t produce a meaningful NRR number. A Series B company obsessing over its 6-customer churn rate is solving a problem it already solved two stages ago. Pull up your own ARR, find your row in The Metric Checkpoint, and fix that metric before adding a dashboard full of others.
Every later-stage metric in this article — Rule of 40, LTV:CAC, ARR per employee — is downstream of getting churn and CAC payback right first (Aleph × Benchmarkit, 2026).
Frequently Asked Questions
What is a good CAC payback period for SaaS in 2026?
A good CAC payback period is under 18 months, with under 12 months marking top-tier efficiency. The 2025 median is 16 months, improved from 18 the year before (Aleph × Benchmarkit).
What is a good NRR for SaaS?
A good NRR depends on which population you’re comparing against: 82% is the median across a broad billing-data panel including small companies (ChartMogul), 102% is the median among venture-backed survey respondents (Aleph × Benchmarkit), and 110%+ is top-quartile in both samples.
Can NRR exceed 100%, and what does that mean?
NRR can exceed 100% because it includes expansion revenue from existing customers, unlike GRR, which caps at 100% by definition. An NRR above 100% means a company’s existing customer base is growing in value even without any new customers.
What’s the difference between NRR and GRR?
NRR includes expansion revenue and can exceed 100%; GRR excludes expansion and cannot. The gap between them — 18 points in the 2025 benchmark (102% NRR vs 84% GRR) — shows how much expansion is compensating for churn.
What is a good LTV:CAC ratio in 2026?
A good LTV:CAC ratio is at least 3:1, though the 2025 median has risen to 4.1x on a gross-margin-adjusted basis, with the top quartile at 7.8x. The original 3:1 guideline assumed a different, non-margin-adjusted LTV formula and 80%+ gross margin (David Skok).
Is the Rule of 40 still relevant in 2026?
Yes. The 2025 median Rule of 40 score for private SaaS is 25%, up from 15% — the largest single-year gain in five years — with the top quartile at 43%. Only 15% of public SaaS companies cleared 40 as of Q2 2026 (Aventis Advisors).
What is a good SaaS Magic Number?
A Magic Number above 0.75 signals efficient enough sales and marketing spend to justify scaling it further; above 1.5 may signal under-investment. The 2025 median crossed 1.0 for the first time in years, reaching 1.37.
What is a good Burn Multiple for a startup?
Under 1.0x is considered amazing, 1.0x–1.5x great, 1.5x–2.0x good, and 2.0x–3.0x suspect, per David Sacks’ original framing. No 2026 market-wide median exists in the primary benchmark report this article relies on elsewhere, so compare your own number against these thresholds rather than an aggregate figure.
How do you convert monthly churn to annual churn?
You compound it rather than multiplying by 12: Annual Churn = 1 − (1 − Monthly Churn)^12. At 2% monthly churn, the annual figure is approximately 21.5%, not 24%.
How is AI changing what counts as ARR?
AI is making ARR harder to define consistently, because usage-based pricing and “Contracted ARR” (CARR) — revenue from signed-but-undelivered contracts — don’t behave like traditional fixed subscriptions. One 2026 report found some AI startups reporting CARR up to 70% higher than actual recognized revenue (TechCrunch).
What is a good ARR per employee in 2026?
A good ARR per employee is around $193,000 at the median among venture-backed companies (Aleph × Benchmarkit), or $141,000 among the broader private SaaS market (SaaS Capital) — the right comparison depends on whether your company is closer to a funded, scaling business or a bootstrapped one.
Are DAU/MAU, NPS, and activation rate still worth tracking?
Yes, but no reliable 2025–2026 primary benchmark exists for any of the three as of this article’s publication. Use the standard formulas (DAU ÷ MAU for stickiness; % promoters minus % detractors for NPS; activated users ÷ new signups for activation rate) and track your own trend over time rather than benchmarking against an unverified external number.
Conclusion
SaaS metrics aren’t a checklist — they’re a chain, and The Metric Checkpoint shows which link to fix first: churn pre-seed, CAC payback and GRR at Series A, NRR and the Magic Number at Series B, the Rule of 40 and ARR per employee at growth stage.
The single most important habit from this guide isn’t any one formula — it’s checking which LTV, margin, or ARR formula a number actually used before comparing it to anyone else’s.
Pull your own numbers against the 2025 benchmarks here, find your row in the framework, and fix that one metric before adding another dashboard. For the AI cost pressure behind several of these numbers, see What Is AI FinOps?
- Aleph × Benchmarkit, 2026 SaaS & AI Performance Benchmarks (342 companies, CY-2025)
- ChartMogul, “The SaaS Retention Report: The AI Churn Wave” (December 2025)
- ChartMogul, SaaS Metrics Cheat Sheet & Metrics Library
- Aventis Advisors, “Rule of 40 in SaaS: 2026 Data, Benchmarks and Valuation Impact” (May 2026)
- SaaS Capital, Revenue Per Employee Benchmarks (15th annual survey, March 2026)
- David Skok, “SaaS Metrics 2.0 – Detailed Definitions,” ForEntrepreneurs
- David Sacks, “The Burn Multiple,” Craft Ventures
- ICONIQ Capital, “2026 State of AI: The Builder’s Economy” (Q2 2026)
- Ben Murray, The SaaS CFO — AI COGS and gross margin, via DualEntry’s SaaS COGS guide
- TechCrunch, “How VCs and founders use inflated ‘ARR’ to crown AI startups” (May 2026)
- Burkland Associates, “AI ARR You Can Defend”
- Maxio, “15 SaaS Metrics That Matter Most” & Stripe, “Essential SaaS Metrics”
- Flowjam, Rule of 40 Calculator (partner tool)






Thanks a bunch for sharing this with all people you really know what you are speaking about!
Bookmarked. Kindly additionally seek advice from my website =).
We can have a hyperlink trade agreement between us