CAC Payback Period: Formula, Benchmarks, and Beating LTV:CAC
CAC payback period is the number of months of gross-margin-adjusted revenue it takes to recover what you spent acquiring a customer. The formula: CAC ÷ (New MRR per customer × Gross Margin %). Under 12 months is the classic target from David Skok's early SaaS work, though the 2024 SaaS median sat around 18 months before improving toward 16 on 2025 actuals. It wins over LTV:CAC for one blunt reason. It doesn't ask you to guess how long a customer will live before you can defend the number.
The formula, spelled out
Here's the canonical version. Copy it straight into your model:
CAC Payback (months) = CAC ÷ (New MRR per customer × Gross Margin %)
The gross-margin adjustment is the part people skip, and it's the part that decides whether your number is honest. If a customer pays you $1,000 in MRR but your gross margin is 78%, you are not recovering $1,000 a month toward that CAC. You're recovering $780. The rest goes to hosting, support, payment processing, and the humans who keep the service running. Benchmarkit's own definition is explicit that CAC payback is measured "on a Gross Margin adjusted basis" — not raw revenue — precisely because raw revenue flatters every SaaS business on the planet.
Then there's timing. Sales and marketing spend that closes a deal in Q2 usually got spent in Q1. Divide this quarter's CAC by this quarter's new customers and you're matching spend to the wrong cohort. Lag the S&M spend by your average sales cycle so the dollars line up with the deals they actually produced. On a two-month sales cycle, the March cohort's true CAC lives partly in January's ad invoices.
The two mistakes that make your number lie
Both common errors point the same direction. They make you look better than you are. The first is using raw revenue instead of gross-margin-adjusted revenue, which understates payback by whatever your cost of goods runs. The second is not lagging S&M, which credits a cohort with spend it never triggered.
A CFO I worked with at a Series B company caught the raw-revenue version live in a board deck. We'd printed a 9-month payback. He asked one question — "is that gross margin adjusted?" — and when the answer was no, the real number came back at 12. Nobody enjoyed the next ten minutes. The lesson stuck: state the basis before someone else asks.
Worked example: watching the margin recover, month by month
Abstract formulas hide the thing that actually matters, which is the crossover point. So let's walk one customer out until the cumulative margin recovers the CAC.
Say you spent $500 to acquire a customer paying $100 in MRR at an 80% gross margin. Each month contributes $80 of recovered margin, not $100. Here's the full recovery table:
| Month | Margin recovered this month | Cumulative margin | CAC remaining |
|---|---|---|---|
| 1 | $80 | $80 | $420 |
| 2 | $80 | $160 | $340 |
| 3 | $80 | $240 | $260 |
| 4 | $80 | $320 | $180 |
| 5 | $80 | $400 | $100 |
| 6 | $80 | $480 | $20 |
| 7 | $80 | $560 | recovered |
The crossover lands in month 7. Cumulative margin ($560) finally exceeds the $500 CAC. Run this on raw revenue at $100/month and the model shows recovery in month 5 — two months of fiction, entirely from ignoring gross margin.
Assumptions in this example — disagree with any of them:
- Gross margin is a flat 80% and never moves.
- Zero expansion revenue; the customer never upgrades.
- MRR is flat $100 with no price increases and no churn inside the window.
- CAC is fully loaded and already lagged to match the cohort.
Real businesses violate at least two of these. Expansion revenue in particular can pull the crossover forward. If your net revenue retention runs above 100%, your true payback is faster than a flat-MRR model shows. Swap in your own numbers and the shape holds. It's the crossover row you're hunting for, not the tidy assumptions.
The good / ok / bad benchmark band
No single median tells you anything useful, so here's the range with sources attached. Skok's rule of thumb from For Entrepreneurs is to recover CAC in under 12 months, on the logic that a long payback strains cash in a capital-scarce startup. That's the aspirational line.
The reality post-2021 drifted. Per Benchmarkit's 2025 report as cited by Drivetrain, the median CAC payback for SaaS companies was 18 months in 2024, up from 14 the year prior. The 2026 Aleph x Benchmarkit benchmarks, drawing on full-year 2025 actuals from 342 companies (198 of them reporting this metric), put the median B2B SaaS payback back down to 16 months, with the top quartile at 6 months or fewer and the bottom quartile at 24 months or more.
Synthesizing those into a band you can actually judge yourself against:
| Rating | CAC payback | Grounding |
|---|---|---|
| Excellent | ≤ 6 months | Aleph x Benchmarkit 2025 top quartile |
| Good | 7–12 months | Skok's <12-month rule |
| OK / median | 13–18 months | Benchmarkit 2024 median 18mo; 2025 actuals ~16mo |
| Concerning | 19–24 months | Aleph x Benchmarkit 2025 bottom quartile |
| Bad | 24+ months | bottom quartile floor |
Before you slot yourself into a row, keep reading. A raw median ignores the one variable that swamps everything else.
Segment by ACV or the number means nothing
Deal size dominates payback more than almost any operational choice you make. Benchmarkit's 2024 data, reported by G-Squared Partners, shows companies with an ACV of $5,000 or less recovered CAC in a median 9 months, while companies with ACV above $100,000 took 24 months. Same metric, nearly a 3x spread, driven entirely by motion.
That's why comparing your enterprise sales motion to a blended or PLG-heavy median is survivorship-flavored nonsense. The public benchmark pools that publish crisp medians skew toward companies willing to report, which skews toward companies proud of their numbers. If your $150K-ACV field-sales business benchmarks against a 9-month median built on sub-$5K self-serve deals, you'll conclude you're broken when you're actually normal for your band. Find your ACV row first. Then judge.
Why payback beats LTV:CAC when cash is tight
Here's the argument that matters. CAC is paid upfront. You spend the whole acquisition cost before the customer sends you a second invoice. Revenue arrives in monthly trickles. Skok's framing of the "cash flow trough" captures the counterintuitive result: "The faster the business decides to grow, the worse the losses become. Many investors/board members have a problem understanding this, and want to hit the brakes at precisely the moment when they should be hitting the accelerator." Growth digs the hole deeper before it fills it.
The remedy isn't slowing down. It's shortening payback. Skok's own modeling makes the stakes concrete: halving the months-to-recover-CAC cut capital consumed to $226.8k against a base case of $801k, and shortened time to cash-flow breakeven from 26 months to 12. That's not a marginal efficiency gain. That's the difference between raising another round and not needing to.
Now weigh that against LTV:CAC. The 3:1 "healthy" benchmark also traces back to Skok, and it's a fine long-horizon story. The trouble is that LTV depends on churn, expansion revenue, and contract length — inputs that, as GrowthSpree puts it, "take 12–24 months to measure reliably." You cannot defend a metric built on inputs you won't observe for two years.
My defensible take: payback is an observable metric and LTV:CAC is a forecast wearing a metric's clothes. Payback is a number you can point at in a table by month 7. LTV:CAC is a projection with a churn assumption baked in, and every board member in the room knows churn assumptions are where hope goes to hide. In a downturn, you defend the observable one. Keep both in your unit economics dashboard — just know which one holds up under cross-examination.
Tracking it without lying to yourself
The blended-average trap is the quiet killer here. Divide total CAC by total new customers each quarter and you get a number that mixes your fast-recovering self-serve cohorts with your slow enterprise ones, telling you nothing actionable about either. Track it by acquisition cohort instead. Watch the January-acquired customers recover their margin month over month, then February's, then March's. The crossover per cohort is the signal. The blended average is noise wearing a suit.
That's an infrastructure problem more than an analysis one. You need acquisition data and revenue data in the same view, keyed to the same cohort, or you spend every quarter reconstructing the join by hand in a spreadsheet. When acquisition cohorts and product-revenue data live in one place — tools like Kixo put funnels, cohorts, and attribution in a single dashboard — the monthly recovery curve becomes something you watch rather than rebuild from exports. Whatever you use, the requirement is the same: cohort in, revenue out, one view. If you're still stitching channel-level attribution together to even get to CAC, sort the upstream attribution question first, because a wrong CAC makes every payback number downstream wrong too.
The one-line version for your CFO
Say this out loud in the meeting: "Payback tells us when this cohort turns cash-positive, on a gross-margin basis, by month. LTV:CAC is our long-horizon story, but it rests on churn we won't confirm for a year."
One of those sentences survives a budget defense. It's the one with a crossover row you can point at. Segment it by ACV, adjust for gross margin, lag the spend, and you'll walk into the room with a number that holds — not a forecast you have to apologize for later.