Efficiency & Unit Economics

What these metrics are for

Unit economics answer one question: is growth being bought at a price the business can sustain? Every metric in this section is a different angle on it, and each has a characteristic error that flatters the number in the same direction.

Acquisition cost, computed honestly

Customer acquisition cost is fully loaded sales and marketing spend divided by new customers acquired from it. Fully loaded means salaries, commissions, benefits, tooling and programme spend. Excluding headcount produces a flattering figure that an investor will recompute themselves, which is a worse outcome than reporting it correctly.

The second error is timing. Spend in a period produces customers in a later period, so dividing this quarter's spend by this quarter's customers attributes the cost of future customers to current ones. Lag the spend by roughly one sales cycle.

Payback binds before the ratio does

LTV:CAC asks whether the model works eventually. Payback period asks how much cash is consumed before it does. A company can have an excellent ratio and still run out of money, because the ratio is indifferent to timing and a bank account is not.

Both must be computed on gross profit rather than revenue. Using revenue understates payback and overstates lifetime value by the entire cost of serving the customer, which in businesses with meaningful support or infrastructure cost is not a rounding difference.

The composites, and their limits

Magic number, burn multiple and Rule of 40 are single-figure summaries designed for external comparison. Each is sensitive to definitional choices — which margin, which growth basis, lagged or contemporaneous spend — and comparing your figure to a published one without matching definitions compares nothing.

Burn multiple is the hardest of these to game, because it uses net cash burn and therefore includes everything a sales-efficiency metric excludes. A company can look efficient on CAC while burning heavily elsewhere; the burn multiple catches that.

Consistency beats correctness

There is no single correct formulation for most of these metrics. Pick one, document it, and hold it — a ratio whose definition moves cannot show a trend, and the trend is the entire value. Where a definition must change, restate the history on the new basis rather than creating a discontinuity nobody remembers.