Marketing payback period is the number of months it takes for the gross profit produced by a group of new customers to equal the marketing spend that won them. It answers a different question from cost per customer. The blended CAC calculation tells you what a customer costs; payback tells you how long the company is out of pocket before that customer has funded its own acquisition. In a sponsor-owned business the second question is usually the binding one, because the marketing budget competes with debt service and the hold period has an end date.
The formula in general circulation was written for subscription software. It divides acquisition cost by monthly recurring revenue multiplied by gross margin, and it assumes a customer who pays the same amount every month until they cancel. A services business has no such ledger. Customers buy a job, then possibly another job a year later, and a real share of them arrive by word of mouth with no traceable spend behind them. The calculation still works, but every input has to be rebuilt from transaction data rather than lifted from a subscription report. This guide covers that rebuild: what belongs in the numerator, which margin belongs in the denominator, how to produce the number channel by channel, what length is defensible, and what to do when a channel pays back after the date the sponsor expects to sell. How much to spend in the first place is a separate question and is covered in the marketing budget guide.
The numerator is every cost incurred to win the customers in the cohort, and the definition should be written down once and not revisited quarterly. Include media, lead purchase, agency and contractor fees, creative and production, the acquisition share of tooling and data, and the loaded cost of the people who run the channels. Include any selling cost incurred before the first order, which in a field services business often means the cost of quoting work that never converts. Exclude the cost of delivering the work, which belongs in the denominator.
The harder problem is timing. Spend does not produce customers in the month it leaves the account. Paid search converts in days, a purchased lead in weeks, and a search presence or a referral engine in quarters. Matching a calendar month of spend against the same month of new customers measures a coincidence rather than a return, and it systematically favors the fastest channel in the mix. Lag the spend to the cohort it produced, using each channel's own observed interval between spend and first order.
Then settle the word-of-mouth question in writing. In the businesses we have audited, roughly 38 percent of new customers arrived through word of mouth that nothing in the system was tracking. Those customers carry gross profit and close to no attributable spend, so leaving them inside a blended cohort makes every paid channel look better than it is. Either hold them in their own cohort with the referral program's cost against them, or exclude them and state that the payback figure covers acquired customers only. Both are defensible. Switching between them between board meetings is not.
Use gross profit, not revenue. Revenue in the denominator is the most common way a payback number gets flattered, and on a low-margin service line it will show a channel paying back in weeks while every customer it produces loses money. Gross profit here means revenue less the direct cost of doing the work: job labor, materials, subcontractors, and the vehicle, fuel, or equipment cost that moves with volume. Contribution margin, which also deducts the variable selling cost of the repeat order, is the better input where a company can produce it reliably. Most cannot yet, and waiting for it is not a reason to go another quarter without the number.
The horizon matters more than the margin definition. Three versions exist and they answer different questions. First-order payback asks whether the company recovers acquisition cost on the initial job, which is the version a lender or a cash-constrained operator cares about. Twelve-month payback counts the gross profit the cohort generates across its first year and is the version that supports an operating decision. Lifetime payback counts everything the cohort will ever produce, and it is close to useless as a control because it can justify any level of spend. Report the first two together. If a payback period is longer than the interval at which customers in that cohort have actually been observed to buy again, it is not a payback period, it is a forecast.
A single company-wide payback figure is a solvency check, not a management tool. The channel view is where the decisions are, and it needs three joins: transaction to customer, customer to first-order source, and source to the spend that funded it. The binding requirement is that the acquisition source is stamped on the customer record at the first order and never overwritten, which is where most of this work fails before it starts.
With that in place the method is mechanical. Fix the cohort by first-order month. Pull gross profit for every transaction that cohort generates over the following 24 to 36 months, grouped by month of service. Sum the channel's lagged spend for the same cohort. Read the month in which cumulative gross profit crosses the spend line. Run it per channel, per brand, and for a multi-site platform per location, because a payback period that is comfortable on average is often carried by two locations and destroyed by four. Where a channel is too small for the cohort to be meaningful, say so and report it as unmeasured rather than folding it into a larger group and calling the result precise.
No published benchmark is worth anchoring a decision to. The figures in circulation come from subscription software and consumer retail, and they assume a recurring margin stream and a cost structure a services business does not have. Derive the target from three of the company's own constraints instead.
What the business can pre-fund. Payback length is a working capital commitment. At the customer volume the plan calls for, the company is carrying acquisition cost for the full payback period before the cohort is whole. Multiply the planned monthly spend by the payback period in months and that is the balance the business is financing at steady state.
What that cash costs. If acquisition is funded from a revolver, a 24-month payback is a financing decision as much as a marketing one, and the interest belongs in the conversation. A payback period that is defensible in a company with cash on the balance sheet is not automatically defensible in a debt-financed one.
How much hold remains. A payback period is only useful if the result lands inside the window in which someone is accountable for it. Twenty-four months is a reasonable underwriting horizon in year one of a hold and an unreasonable one in year five.
Those three constraints produce a range, and the range should be set per channel at the moment money is allocated rather than assessed afterward. That is the allocation discipline described in the investment marketing guide: every material line carries an expected payback period, and actuals are reported against the expectation that was set.
The awkward case is a channel whose payback lands after the date the sponsor expects to sell. Rejecting it on that basis alone is the reflex, and it is usually wrong, because it systematically starves the channels that raise the exit price. Apply two tests instead. Does the spend build something a buyer pays for, and is the commitment reversible if the timetable moves.
Owned acquisition passes both. In one multi-site platform, shifting the mix toward owned channels cut customer acquisition cost by 60 percent and took the owned share of new customers to three times its prior level, and the mix takes two to four quarters to move materially. In another, the same commercial work supported 82 percent footprint growth in 15 months. Neither of those returns is visible inside a single quarter's payback window, and both are the kind of asset a buyer underwrites. The sequencing argument for building them early rather than late is set out in the CAC reduction guide.
A workable rule: in the first half of a hold, a channel can be underwritten on payback alone. In the final 18 months it has to clear payback inside the remaining window or be defensible as something that shows up in the price. Spend that does neither is the first thing to cut.
The blended figure across all channels tells you whether growth is solvent. It does not tell you whether the next dollar is, and that is the question a budget decision actually turns on. The marginal dollar in any channel pays back more slowly than the average dollar, because the cheapest and most intent-driven demand is bought first. A channel with a ten-month blended payback can easily be at twenty months on its most recent increment of spend.
Getting to the marginal number does not require new tooling. Compare the payback of the most recent tranche of spend against the trailing average, or split the channel by geography, campaign tier, or keyword group and read the payback of the weakest tier that is still funded. Then make scaling decisions on that number. Companies that scale on the average find out a quarter after they have committed the money.
One page, four quarters of history. For each material channel: lagged spend, customers acquired, gross profit per customer, payback in months on first-order margin and on twelve-month margin, the payback expectation set when the money was allocated, and the variance against it. Below that, the blended payback trend, the payback of the most recent increment of spend, and the reallocations made since the last meeting with the reasoning attached.
The reason to build this before being asked is that the question is arriving anyway. S&P Global's 2026 research found 71 percent of general partners and 53 percent of limited partners now treat operational value creation as a priority, which means commercial numbers are being read more closely at the fund level than they were two years ago. A company that can show payback by channel, and show it moving in the right direction, is in a materially stronger position than one that can show spend and leads. In one portfolio company, rebuilding the commercial engine on this basis produced a 36 percent revenue increase over 24 months on flat marketing spend, which came from reallocating money against payback evidence rather than from spending more of it. Assembling the first honest version of the number is normally part of an eight-week commercial audit, because the data joins are the slow part and they are the same joins several other commercial questions depend on.
Marketing payback period is the number of months it takes for the gross profit produced by a group of new customers to equal the marketing spend that won them. It answers a cash question rather than an efficiency question: acquisition cost tells you what a customer costs, and payback tells you how long the company funds that customer before the customer funds itself. In a private-equity-backed company the cash question is usually the binding one, because the marketing budget competes with debt service and the hold period has an end date.
Define a cohort by the month customers placed their first order. Sum the marketing spend that produced that cohort, lagged to the period when the spend actually ran rather than the month the customers appeared. Then sum the gross profit that cohort generates month by month from transaction data, and find the month at which cumulative gross profit crosses the spend. That month is the payback period. The subscription formula of acquisition cost divided by monthly recurring revenue does not apply, because a services customer buys jobs at irregular intervals rather than paying a fixed monthly fee.
Gross profit. Using revenue is the single most common way the number gets flattered, and it can make a channel that loses money on every customer look like it pays back in weeks. Deduct the direct cost of delivering the work, which in a services business means job labor, materials, subcontractors, and any vehicle or equipment cost that varies with volume. Contribution margin is a better input if the company can produce it reliably, but gross profit at the job level is the workable version and it is enough to make decisions on.
There is no benchmark worth anchoring to, because the figures in circulation are medians drawn from subscription software and assume a margin structure a services business does not have. Derive the target from three of the company's own constraints instead: how many months of acquisition the business can pre-fund at the volume being planned, what that cash costs given the capital structure, and how much of the hold period remains. A defensible payback period is one the company can actually finance and still show the result before the sponsor sells.
Test it against two questions rather than rejecting it automatically. Does the asset the spend builds show up in the price a buyer pays, and is the commitment reversible if the exit timetable moves. A durable search presence, a working referral engine, and a customer base that a buyer can see is not rented all survive both tests, which is why they belong early in a hold rather than late. Spend that produces neither a payback inside the hold nor an asset in the price is the spend to cut first.
Have a revenue problem the board is asking about? Start a conversation.