Metrics, retention and the organization · Answered by Golden Section from more than 400 B2B software companies observed
Under 18 months, calculated on gross margin, is the line we use for a fundable growth engine, and under 12 months signals a genuinely efficient one. Compute it as fully loaded sales and marketing cost for the period, including unpaid onboarding work needed to get a customer to normal use, divided by new ARR times gross margin, in months. The number depends on gross retention: a long payback on customers who stay for a decade can work, while the same payback on a base losing 16% a year does not. At 24 months, fix before spending more. Start by recalculating payback by channel for the last four quarters.
Spend more where marginal payback stays under 18 months and gross retention is strong enough to earn the cost back several times. Where payback runs past about 24 months, fix conversion, price or cost of sale before adding spend.
| Metric | Value | What it means | Source |
|---|---|---|---|
| CAC payback, fundable | under 18 months | months of gross profit from new ARR needed to recover acquisition costUnder 12 months signals a genuinely efficient growth engine | Golden Section, publishedThe Balanced Path |
| Lending use-of-proceeds test | under 18 months | demonstrated CAC payback in the channel being fundedWe fund proven channels, not experiments | Golden Section, publishedGrowth Capital Lending |
| New-logo CAC ratio | $2.00 median; $2.82 fourth quartile | sales and marketing cost per $1 of new-logo ARRRose 14% in a year while blended CAC fell because cheaper expansion ARR grew as a share | Golden Section, publishedInvesting in Software addendum, September 2026 |
| Retention and payback cohorts | 71% vs 10% median growth | companies with NRR above 106% and payback under 10 months versus NRR below 98% and payback over 15 monthsHigh Alpha 2025 cohort analysis; S&M intensity was near flat at 25–30% of revenue across bands | External benchmarkHigh Alpha 2025 SaaS Benchmarks, as cited in GS Investing in Software addendum |
| Payback from sales efficiency | 0.7 → ~11 months; 1.5 → ~24 months | (S&M per $1 new ARR ÷ gross margin) × 12, at 75% gross marginArithmetic linking the sales efficiency ratio to payback | IllustrativeArithmetic from the Sales Efficiency Ratio play |
Payback is only comparable once cost scope and margin treatment are fixed. Use prior-period sales and marketing expense, including salaries, commissions, tools and travel, because spend precedes bookings by about one sales cycle. Our SaaS metrics play also counts unpaid services needed to get a customer to average use. Divide by gross profit, not revenue, which is why gross margin matters here.
A reasonable CAC is one the customer repays several times. Our research puts the median new-logo cost near $2.00 per $1 of ARR and the rational ceiling nearer $2.20 to $3.00 at today's margins and exit multiples. CAC usually rises for one of three reasons: broader targeting that lowers win rates, longer cycles, or a mix shift toward new logos as expansion slows. Reduce it by tightening qualification and segment focus, moving budget to channels with the lowest cost per closed dollar, and aligning price with cycle length, as the sales efficiency ratio and unit economics plays lay out.
When payback is good, spend more, in steps, and watch the marginal channel rather than the blended figure.
A company reports 14-month payback. Rebuilt with the sales team's salaries included and prior-quarter spend matched to this quarter's bookings, it is 22 months; gross margin is 70% and gross retention 88%. Outbound pays back in 30 months and referrals in 9. The CEO shifts half the outbound budget to a referral program and a partner channel and tightens qualification on outbound, bringing blended payback to 16 months within three quarters without cutting total spend. All figures are invented for illustration.
Enterprise deals with multi-year contracts and very high gross retention can justify a longer payback, because the customer's lifetime is long enough to repay it. Early channels in test should be judged on leading conversion before payback is meaningful.
From the Golden Section mistakes list, each paired with the play that prevents it.
Measuring cost per lead instead of cost per retained dollar of ARR hides where payback is actually lost.
Adding spend to a channel with long payback makes the losses arrive faster.
Pricing out of line with sales cycle length builds long payback into the model from the start.
In the order we would run them. Each is on its own page, most with a free Excel template.
Defines CAC scope, including onboarding cost, consistently.
Establishes the gross margin that payback divides by.
Models return per customer so payback is read alongside lifetime value.
Tracks spend per $1 of new ARR monthly and sets improvement targets.
Finds the stage where conversion losses are inflating CAC.
Sales & marketing plays The metrics, unit economics and efficiency plays define and improve payback together.
For most companies at this stage, yes. At 75% gross margin it means spending roughly $1.50 of sales and marketing per $1 of new ARR, past the $1.00 level where we would fix the motion before financing it. Find which channel or segment carries the long payback before cutting everywhere.
We rely on payback and gross retention rather than a single LTV:CAC multiple, because lifetime value depends almost entirely on the churn assumption. If you use it, calculate LTV on gross margin and your actual gross retention, and treat any ratio built on assumed low churn with suspicion.
Include the unpaid onboarding and implementation work it takes to get a new customer to normal use, because that is part of acquiring him. Ongoing account management belongs in cost of revenue or retention cost, not CAC.
Yes. Very short payback with stable conversion often means you are under-spending on a channel that works; our sales efficiency play treats a ratio well below 0.7 as a sign it may be time to spend more. The limit is whether the next dollar pays back as fast as the last.
Enterprise CAC runs higher per logo because cycles are longer and involve more people, and that is acceptable only when contract value and gross retention are high enough to repay it. Price and cycle should move together; our mistakes list uses about two weeks of cycle per $20K of ACV as a guide.
A channel with demonstrated payback under 18 months is what our lending funds, at $500K to $5M against ARR quality. If payback is longer, the better use of the next quarter is fixing it.
Growth capital lending →Reviewed by Dougal Cameron, CEO & Co-Founder on 2026-09-23. Golden Section observations are labeled separately from external benchmarks and illustrative arithmetic.