The operator’s job is not to admire the revenue line. It is to establish what remains after the work required to create, deliver, support, collect, and retain the customer. When that remainder shrinks, the business may continue to grow while becoming less able to finance its own growth.
The weekly dashboard is a small discipline with a large effect. It fixes the definition of the unit, separates actuals from forecasts, and places every ratio beside the evidence that can explain it. A good dashboard does not make judgment unnecessary. It makes judgment arrive earlier.
Read the controls in order. The final question is always whether the unit leaves enough cash and durable value to support the next one.
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 = contributionEvery calculation is only as useful as its unit and its cost boundary.
The boundary is a decision. A support team can look like overhead in one model and variable cost in another. If serving the next customer requires additional people, hours, tickets, usage, or implementation, show that burden somewhere visible. Otherwise the dashboard will reward sales that the operating system cannot afford.
Keep the price ledger honest. List price is a positioning input. Realized price is an economic fact. Report them separately so discounting, refunds, credits, failed payments, and mix shifts cannot quietly lower the unit while the headline price remains unchanged.
Payback tells you how long growth must be financed before contribution catches up with the original investment.
Do not let LTV make payback disappear. A customer can have a profitable modeled lifetime and still require a cash window the company cannot support. Payback is not a replacement for LTV. It is the operating clock that tells you whether the balance sheet can wait for the lifetime model to become real.
Compare like with like. Use the same period and attribution logic in CAC’s numerator and denominator. For noisy weekly data, use a rolling window and write down the definition. A ratio that changes because the calculation changed is not a trend.
Cohorts prevent a young base, a large expansion, or a blended average from hiding the actual quality of the latest acquisition.
Beginning revenue minus churn and contraction, divided by beginning revenue. It shows what the existing base kept before expansion is counted.
Beginning revenue adjusted for churn, contraction, expansion, and reactivation. It shows whether the existing base grows or shrinks in aggregate.
Track activation, usage, service cost, renewal, and cumulative contribution by acquisition group. A cohort can look good at sale and bad by day 60.
Ask what the average is hiding. Customer churn and revenue churn can point in different directions. Expansion can make net retention look healthy while gross retention falls. A cohort table gives the operator a way to see who is leaving, who is shrinking, and who is carrying the average.
Use behavior as the leading edge. First value, usage, support intensity, payment behavior, and delivery completion often move before a cancellation. The dashboard should connect those operating events to the later revenue movement instead of waiting for the cancellation reason to become the only evidence.
Close the cash loop. When net burn is divided by net new ARR, the result is the burn multiple: a directional view of how much cash the business spends to create the next increment of recurring revenue. Keep the burn definition and time window explicit.
Margin death is rarely one dramatic number. It is a sequence that turns small concessions into a structural subsidy.
Pull realized price, refunds, credits, service hours, support, and usage. The sale may be larger while the unit is poorer.
Pull channel mix, sales effort, cycle length, deal size, and discounting. The acquisition machine may be buying a harder customer.
Pull account-level movement and concentration. A few expansions can hide broad contraction.
Pull collections, terms, failures, delivery timing, refunds, payroll, and burn. Contract value is not liquidity.
“The number is not the diagnosis. It is the location of the next question.”
Write the answer beside the evidence and the owner. A dashboard becomes useful when it changes a decision.
Gross margin captures direct COGS. Contribution can include other costs that scale with the unit, such as delivery, support, usage, payment fees, onboarding, refunds, and fulfillment. Showing both reveals where the burden enters.
Yes. LTV is forecast-driven and can use assumptions that have not matured in the latest cohort. Current contribution, payback, gross retention, cohort behavior, and collections are the earlier operating controls.
Investigate when two linked measures worsen for two consecutive periods, or when one change threatens cash or delivery capacity. Use internal floors and cash windows instead of treating a public benchmark as a law.
Keep the cadence, widen the window, and inspect the source events. Review invoices, deals, service hours, support queues, and bank movements. Small samples call for judgment, not invented decimal places.
Define the boundary, measure the recovery clock, read the cohort, and finish in cash.
Return To The Instruments