Why Payback Period Beats Last-Click ROAS as Your North-Star Metric
As digital advertising concentrates in social, video and commerce media, last-click ROAS is getting worse at describing reality. Payback period — how long it takes to earn back what you spent to acquire a customer — gives marketing and finance a single number both sides can defend.
GrowthGrowth teams do not suffer from a shortage of numbers. They suffer from a shortage of numbers that a CFO will still believe on a bad Tuesday. Return on ad spend, as reported inside a platform, is easy to produce and easy to game. It usually credits the last click. It rarely tells you whether the customer was going to buy anyway. And it gets less honest as more of the industry’s money moves into formats that do not look like a search click.
That shift is no longer a forecast. The Interactive Advertising Bureau, in its PwC-conducted report on 2025, said U.S. internet advertising revenue reached $294.6 billion, up 13.9 percent. Social advertising grew 32.6 percent to $117.7 billion. Digital video grew 25.4 percent to $78 billion. Commerce media grew 18 percent to $63.4 billion. Creator advertising spend reached $37 billion. Search, still huge at $114.2 billion, grew more slowly (11 percent) than it did the year before. IAB’s own read is that the market “has reoriented around performance channels” and that buyers want spend tied to business results. (See the IAB summary, April 16, 2026.)
If the mix is heavier in social, video and creators, last-click ROAS will systematically flatter search and shopping and systematically punish the channels that create demand. Finance notices the mismatch long before marketing admits it. The way out is not a prettier dashboard. It is a shared definition of payback.
What payback actually means
Customer-acquisition payback is a time question: how many months of gross profit from a new customer does it take to recover what you spent to acquire that customer?
A workable formula looks like this:
- Payback (months) = CAC ÷ (monthly gross profit per new customer)
Where:
- CAC (customer acquisition cost) is all fully loaded sales and marketing cost to win a new customer in the period — media, creative, tools, agencies, and the people who run them — divided by the number of new customers. Do not hide salaries. Do not count existing customers who happened to click an ad.
- Monthly gross profit per new customer is what that customer pays you, minus the variable cost to serve them, in a typical month. For a subscription, that is roughly ARPU × gross margin. For a retailer, it is contribution after discounts, returns and cost of goods, averaged over the first months.
If CAC is $360 and a new customer produces $90 of gross profit a month, payback is four months. That sentence can sit in a finance meeting without translation. “ROAS of 3.2x on Meta,” by contrast, requires a glossary, a grain of salt and an argument about view-through windows.
This is a framework, not a universal benchmark. Pick one window that fits your cash cycle, write it down, and make every channel report against it.
Why ROAS keeps starting fights
Last-click ROAS asks a different question: of the revenue the platform is willing to claim, how does it compare with the media dollars spent inside that platform? Three problems follow.
First, channels that assist but do not close look worse than they are. That is how a company “optimizes” its way into an all-search mix and then wonders why the top of the funnel dried up.
Second, platforms measure themselves. In a market where programmatic alone was $162.4 billion in 2025, according to IAB, you should expect each vendor’s number to put that vendor in a good light.
Third, cookies were never going to save you. Google announced in April 2025 that Chrome would keep its current third-party cookie choice rather than introduce a new standalone prompt (Privacy Sandbox update). Tracking did not snap back to 2018. Incrementality — what happens if you turn a channel off — is still the adult measurement.
Payback does not replace incrementality tests. It gives those tests a unit everyone already understands: months until the cash comes back.
How to install payback without a six-month data-science project
You can stand this up in two weeks if you are willing to be slightly crude at first.
- Week 1: agree on the ingredients. Finance and marketing sit down with last quarter’s actuals — not the media plan, the bank-and-payroll actuals. Add up sales and marketing cost. Count new customers (not leads, not “engaged users”). Compute monthly gross profit per new customer using the last 90 days of cohort data if you have it; if you do not, use blended gross margin × average revenue per new customer in month one, and flag it as a temporary input.
- Week 2: report every channel two ways. Keep the platform ROAS; you will still use it for in-flight optimization. Add an estimated payback by assigning new customers with a simple, written rule (for example: last non-direct click for performance channels, plus a tested share for brand/video based on the latest holdout). The rule will be imperfect. Writing it down is what makes it improvable.
- Ongoing: run one kill test a month. Take your most expensive disputed channel and pause or geo-split it long enough to see whether new-customer volume and payback move. Move budget only when the test contradicts the last-click story.
If you raise money, this discipline is not cosmetic. Carta’s Q3 2026 private-markets data shows a fundraising market that is richer and pickier at the same time: companies on Carta raised $49.3 billion across 1,179 rounds in the quarter, but total deal count was down 21 percent year over year. Seed dollars were up 21 percent while seed rounds were down 27 percent. Investors are writing larger checks to fewer companies (Carta, State of Private Markets: Q3 2026). A slide that says “ROAS is healthy” is easy to shrug at. A slide that says “we recover CAC in five months on gross profit, and here is the cohort” is a conversation.
Guardrails so the metric does not rot
- Do not stretch payback by inventing lifetime value. Payback should use near-term gross profit you can observe, not a 36-month LTV model with a hopeful retention curve.
- Do not let paid search “own” every branded conversion. Separate branded search from non-branded. Brand demand often comes from everything else you did.
- Do not optimize payback by cutting creative and brand to zero. That improves the next 30 days and damages the next 300. Hold a fixed share of spend for demand creation and judge it on incrementality, not on last-click payback.
- Recompute after pricing or cost shocks. If your input costs jump — as September 2026 CPI reminded anyone watching energy prices — gross profit per customer moves, and so does payback. The BLS September CPI showed a 0.6 percent monthly rise and a 3.7 percent twelve-month rise in the all-items index, with energy up 20.4 percent over the year. Unit economics are not a set-and-forget file.
Key takeaways
- Last-click ROAS systematically over-credits closers (especially search) and under-credits the social, video and creator channels that now absorb a growing share of a $294.6 billion digital ad market.
- Payback period — CAC divided by monthly gross profit per new customer — is a single number marketing and finance can share.
- Stand it up with last quarter’s actuals in two weeks, keep platform ROAS as a tactical input, and decide budget shifts with monthly incrementality tests.
- In a venture market that is sending more money to fewer companies, clean payback math is part of how you stay fundable — or how you stay comfortably independent.
Keep reading
More in Growth →
GrowthShopify’s Q2: 34% revenue growth and an 18% free-cash-flow margin
On August 5, 2026, Shopify said second-quarter revenue rose 34% to $3.583 billion and GMV reached $115.567 billion, with free cash flow of $654 million, an 18% margin.
GrowthShopify’s Q1: $3.17 billion in revenue and more than $100 billion in GMV
On May 5, 2026, Shopify said first-quarter revenue rose 34% to $3.170 billion and GMV reached $100.743 billion, with free cash flow of $476 million, a 15% margin.
GrowthShopify closed 2025 at $11.6 billion in revenue and authorized a $2 billion buyback
On February 11, 2026, Shopify said 2025 revenue was $11.556 billion, up 30%, with $2.007 billion in free cash flow, and its board authorized up to $2 billion in share repurchases.