The Multiplier Runs in Both Directions
Contribution margin is the per-unit cash a sale generates after the costs that vary with that sale — cost of goods, payment processing, shipping, the support hours a transaction reliably consumes, the cloud spend a query incurs. It is not gross margin dressed up; it is the marginal arithmetic of one more customer. The reason it governs the timing of growth capital is mechanical. Fixed costs are paid once and amortized across volume, so scale dilutes them. Variable costs are paid again on every unit, so scale multiplies them. When contribution margin is positive, each new customer carries the firm a step closer to covering its fixed base, and growth is the lever that finds operating leverage. When contribution margin is negative, each new customer enlarges the hole, and growth is the lever that finds insolvency faster. The same accelerator that compounds a healthy business compounds a broken one.
This is why "we'll fix the economics at scale" is the most expensive sentence in a pitch. Some costs genuinely fall with volume — purchasing leverage, server utilization, amortized onboarding. But those are improvements to a number that must already be the right sign. Scale lowers unit cost; it does not invert the relationship between price and the cost of serving. A company that loses money on the marginal unit and raises capital to sell more of them has not bought time. It has bought a larger denominator of loss and a shorter runway to discover it.
The Failure Mode Hides Inside Blended Numbers
The trap is that aggregate metrics conceal unit-level rot. Blended gross margin, total revenue growth, and a falling average cost per user can all improve while the marginal customer remains underwater — because early adopters, enterprise anchors, or a subsidized cohort are masking the rest. The discipline is to disaggregate: compute contribution margin by segment, by channel, by cohort, and by acquisition vintage, and to insist that the incremental customer — the next one the growth budget will actually buy — clears the bar, not the flattering average of customers you already have. A business can be profitable on its installed base and lethally unprofitable on its growth, and the income statement will not tell you which until the cohorts mature.
Two distortions deserve naming because they routinely launder a negative margin into a positive-looking one. The first is capitalizing or excluding costs that are genuinely variable — treating customer-acquisition or fulfillment spend as a one-time investment when it recurs with every cohort. The second is pricing against a discount or promotional rate that the cohort will never graduate from, so the modeled margin describes a customer who does not exist.
The Thresholds That Justify Acceleration
Resolving unit economics does not mean waiting for profitability; it means establishing that the unit is structurally sound before you pour capital through it. The conditions that justify pressing the accelerator are specific and jointly required:
- Contribution margin is reliably positive on the incremental customer at the price you can actually charge — typically 30 percent or more for software, where it must also absorb support and infrastructure, and unambiguously above zero for any physical-goods or services model after all per-order variable cost.
- The payback period on fully loaded acquisition cost is recovered in cash within twelve months, ideally fewer, so growth is self-funding rather than runway-funding.
- Lifetime value covers acquisition cost by a margin wide enough to survive churn and discount-rate stress — a ratio near 3-to-1 on properly discounted, contribution-based LTV, not revenue-based — and the ratio is stable or improving as you scale the channel, since acquisition cost rises as the cheap demand is exhausted.
The decision implication is sequencing, not abstinence. Capital deployed against a proven unit buys compounding; capital deployed against an unproven one buys an experiment dressed as expansion. The CFO's veto is therefore narrow and firm: prove the margin on the marginal customer first, then raise the round sized to scale it. Reverse that order and you have not financed growth — you have financed the cost of learning, at the most expensive price the cap table will ever pay.