Are we paying back demand acquisition profitably across horizons?
№ 047 · Demand LTV/CAC
Definition
MetricDemand-side lifetime value over demand-side acquisition cost, margin-adjusted
Unitratio (x)
Same for buyers: lifetime contribution margin divided by buyer acquisition cost. Build by aggregating buyer-side per-order CM over expected lifetime and dividing by buyer CAC. Below 3 means buyer-side acquisition is too expensive. Example: $52 LTV / $14 CAC = 3.7x.
- LTV/CAC
- Lifetime value to acquisition cost ratio. LTV divided by CAC. Above 3x is the common health threshold.
- CM
- Contribution margin. Revenue minus all variable costs attributable to the unit.
- CAC
- Customer acquisition cost. Fully loaded sales and marketing spend in a period divided by new customers won in that period.
- LTV
- Lifetime value. Average gross profit per customer per period divided by that period churn rate, or summed over expected life.
Benchmarks
| Bottom 30% | Median | Top 30% |
|---|---|---|
| below 3x, or ratios propped up by extending the LTV window beyond where cohorts have actually been observed; paid-dominated mixes with rising CAC and unflattened retention curves | approx 3x floor; buyer LTV computed on net revenue from the buyer's transactions, not GMV; acceptable demand-side ratios depend on supply retention and CAC (weaker supply economics demand stronger buyer economics) | 4x+ per the a16z diligence bar, typically reached when demand retention curves flatten and repeat frequency compounds; organic and brand-led mixes push blended CAC down over time as network effects mature |
consumer marketplaces, VC diligence norms 2019-2026; LTV margin-adjusted, CAC fully loaded per first-time transacting buyer, cohort-observed rather than extrapolated where possible · Same caveat as supply side: the 4x bar is a stated norm, not a distribution. The interdependence is the trap: Casey Winters notes demand retention requirements depend on supply-side economics, so a fixed universal ratio target misleads. LTV computed on blended rather than margin-adjusted revenue overstates the ratio by roughly the inverse of contribution margin.
Category split omitted: No two independent Tier 1-2 sources publish demand-side LTV/CAC distributions by marketplace category.
When it looks bad
The ratio holds only because the LTV window keeps being extended in the model; cohort-observed curves show payback receding, and per-market splits reveal the blended ratio is carried by one mature market.
Reported blended ratio 3.2x on a 36-month projected LTV, but observed 12-month contribution covers only 60% of CAC and the newest market runs at 0.7x.
What to do about it
- Report the ratio per market and per channel on observed cohort contribution at fixed horizons (12 and 24 months), reserving projected LTV for planning only; local network effects mean the blended number is the least informative version.
- Push CAC down with compounding channels before optimizing paid: SEO on supply inventory, referral loops on both sides, and CRM reactivation, which acquire at a fraction of paid CAC and improve with scale.
- Raise LTV through the second transaction: the 30-day window concentrates half of eventual repeat behavior, so first-30-day lifecycle work moves LTV more per dollar than month-12 promotions.
- Rebalance the two sides when the demand ratio is structurally weak: cheaper, stickier supply lowers required demand economics per the Winters interdependence, so supply retention work is sometimes the best demand-LTV initiative.
Sources
- Andreessen Horowitz (Olivia Moore) 4x+ LTV:CAC sought on both sides; higher demand GMV retention increases LTV and allows more acquisition spend at a healthy ratio a16z.com ↗
- Casey Winters and Lenny Rachitsky viable demand retention and economics depend on supply retention and acquisition costs; where supply is cheap and sticky, demand can retain less and the model still works caseyaccidental.com ↗
- Gianluca Valentini (Equity Pitcher Ventures) demand CAC and supply CAC must be combined with conversion frequency weights; demand side usually sets the lifespan used in the model linkedin.com ↗
- Foundry CRO 3:1+ marketplace floor with per-market rather than blended tracking, since local network effects make blended ratios hide failing markets foundrycro.com ↗
- Eightx public marketplace buyer paybacks modeled at 1-3 months where brand-search demand keeps buyer CAC in low single dollars; paid-led private marketplaces sit far above that CAC eightx.co ↗
- Decile first-order CAC payback percentages by category (home goods 218%, fashion 162%, supplements 82%) show how category economics set the demand-side starting point decile.com ↗