Most operators do not discover margin death when the company runs out of cash. They discover it after a month of explaining why revenue is up, customers are arriving, and the bank balance still feels worse. The problem is not that the business lacks a dashboard. The problem is that the dashboard reports top-line motion after the economic damage has already happened.
A weekly unit-economics review starts with the unit, not the story. It asks what a sale is really worth after the variable work required to acquire, deliver, support, collect, and retain it. Then it follows the recovery speed, the cohort quality, and the cash timing. The objective is not to memorize ratios. It is to catch the first connected signals while the operator can still change price, mix, channel, terms, or delivery.
Choose a control point to see the question, formula, and evidence that belong in a weekly operating review. Treat the sequence as one system: margin determines contribution, contribution determines payback, payback only matters when customers stay, and all four controls eventually meet in cash.
Start with realized price and every variable cost that moves with the unit. A growing top line can still be a warning when delivery, support, usage, refunds, or payment costs consume the contribution left after acquisition.
realized price − variable cost = contributionEach number answers a different question. The mistake is using one composite ratio to hide the layer that is breaking.
Define the unit in language the operating team can verify: one order, one account, one project, one location, or one month of service. Then calculate what remains after the costs that move with that unit. Gross margin is useful, but it can be too high-level when support, onboarding, delivery labor, payment fees, refunds, or usage costs scale with the customer.
Formula: contribution margin per unit equals realized revenue per unit minus variable cost per unit. Contribution margin percentage divides that result by realized revenue. Record list price separately from collected price. If the business needs discounts, credits, custom work, or founder intervention to close the sale, those are not side notes; they are evidence about the real unit.
Weekly question: did the customer become more expensive to acquire or more expensive to serve? Those are different problems and require different fixes.
CAC is a stock number. Payback is a time number. The operator needs both because a customer can be profitable in a lifetime model and still consume too much cash before the business recovers the acquisition investment. Calculate CAC by channel when possible, include attributable sales labor, and divide by monthly contribution rather than unadjusted revenue.
Formula: CAC payback months equals CAC divided by monthly contribution from the new customer. Separate contractual payback from collected-cash payback when billing terms, installments, failed payments, or refunds change timing. A channel that appears efficient on revenue may be slow on contribution and slower still on cash.
Weekly question: if the next cohort performs like the last one, how long will the company finance its acquisition before the unit pays for itself?
Aggregate retention is a dangerous comfort when the customer base is changing. A new cohort may not have reached the point at which it churns, while expansion from a few large accounts lifts net retention enough to conceal broad contraction. Track customer churn and revenue churn separately, then read gross retention before expansion is allowed to decorate the result.
Formula: gross revenue retention equals beginning revenue minus churn and contraction, divided by beginning revenue. Net revenue retention adds expansion and reactivation. Pair both with activation, first value, usage, support intensity, and cumulative contribution by acquisition cohort.
Weekly question: are customers staying because they are receiving value, or is the average being rescued by a small number of expansions?
Booked revenue is not collected cash. A company can report new contracts while funding payroll, delivery, inventory, acquisition, and implementation ahead of customer receipts. The cash control therefore reconciles what was sold, what was invoiced, what was collected, what is overdue, and what variable costs were paid before the unit returned cash.
Formula: burn multiple equals net burn divided by net new ARR; use a rolling window when weekly volume is noisy. Quick ratio compares positive recurring movements with churn and contraction. Neither is a universal score. Both are prompts to ask whether the next dollar of growth is becoming more expensive to finance.
Weekly question: what operating decision could improve cash this week without pretending that an uncollected contract is liquidity?
Margin deterioration usually arrives as a combination of signals. Use the pattern to decide where to investigate first.
| Signal pattern | Likely operating leak | First evidence to pull |
|---|---|---|
| Revenue up; contribution down | Discounts, refunds, variable delivery, support, or usage costs are rising faster than realized price. | Invoice-level price, credits, service hours, tickets, usage, and payment fees by segment. |
| CAC up; close rate flat | Spend is moving into a harder channel, sales labor is expanding, or mix is shifting toward smaller accounts. | Channel spend, attributable labor, deal size, cycle length, discount rate, and win rate. |
| Payback up; LTV:CAC stable | Lifetime assumptions are masking weak early contribution or the cohort has not matured enough to validate the forecast. | Realized cohort contribution at 30, 60, and 90 days versus modeled LTV inputs. |
| NRR stable; GRR down | Expansion from a concentrated group is hiding broad churn or contraction among the rest. | Account-level churn, contraction, expansion, customer size, usage, and renewal dates. |
| ARR up; cash down | Collections lag, annual billing is absent, delivery is front-loaded, or growth is being financed with burn. | Aged receivables, payment failures, billing terms, payroll, vendor timing, and net burn. |
The dashboard is a routing system. It should not merely color a cell red. It should tell the operator which ledger, cohort, call recording, support queue, invoice, or delivery schedule to open next. The strongest weekly review ends with one variance owner and one reversible action.
Use two confirmations before declaring a trend. A single bad week can be noise, seasonality, a billing batch, or one unusual customer. A repeated movement across linked metrics is more actionable: falling realized price plus rising service hours, or rising CAC plus longer sales cycles, or stable NRR plus worsening GRR. The point is not to delay action; it is to avoid treating a one-off as a system law.
Consistency is a control. Every weekly review should use the same definitions, time windows, cohort logic, and decision owners.
LTV is a forecast built from price, retention, margin, and expected lifetime. It can look healthy while current contribution, early cohort behavior, or cash timing deteriorates. Keep it as a planning lens, but pair it with realized contribution, payback, gross retention, cohort movement, and collections.
Use a definition that matches the decision. A media-only CAC can be useful for a narrow channel experiment, but a blended acquisition CAC should include the sales and marketing costs attributable to creating new customers. State the denominator, allocation, and period explicitly.
There is no universal list. Include costs that genuinely move with the unit: COGS, payment fees, fulfillment, usage-based infrastructure, delivery labor, onboarding labor, and support when incremental. Keep the ledger stable, then show a sensitivity view for costs that are partly variable.
Keep the weekly operating cadence, but use rolling four-week or thirteen-week windows for noisy ratios. Review individual deals, invoices, service hours, and cash events alongside the ratios. Small samples should widen judgment, not justify invented precision.
Often it is not a negative margin line. It is a widening gap between the promised economics and the realized economics: more discounting, more custom delivery, more support, slower collections, or a new cohort that takes longer to reach first value. Watch the operational inputs before the quarterly margin statement confirms the damage.
Begin with an internally approved cash window, contribution floor, target segment, and service capacity. Trigger investigation when two linked measures worsen for two consecutive periods or when one change threatens cash or delivery. Treat published benchmarks as context, never as a substitute for your own economics.
Define the unit, reconcile the variable cost, turn CAC into a recovery clock, read the cohort, and finish in collected cash.
Run The Economics Path