The Paid Media Reporting Cookbook: ROAS, ROI and Honest Numbers

Return on ad spend is revenue attributed to ads divided by ad cost. A ROAS of 4 means four units of revenue per unit spent. Whether that is good depends entirely on how much of each unit of revenue you keep. Revenue is not profit, and the platform has no idea what your product costs to deliver.

Break-even ROAS is the ROAS at which an ad-driven sale makes back its media cost and nothing more. The formula is one divided by your gross margin, expressed as a fraction. If gross margin is 40 percent, break-even ROAS is 1 / 0.40, which is 2.5. Below 2.5 the media spend costs more than the margin it earns. The 40 percent is an illustrative figure for the example; use your own, calculated from revenue less the direct cost of delivering the product or service, and be explicit about what you included.

Set the break-even number before the campaign launches and write it where everyone reporting will see it. It stops a team from celebrating a 3.0 that is quietly losing money, and from panicking at a 2.0 on a product with a high margin. For businesses with recurring revenue, the first-order margin and the lifetime margin give different answers, so decide which one governs and record the reason. A defensible default is to use first-order margin for decisions about cash and lifetime margin for decisions about how much growth to fund.

Customer acquisition cost is the total cost of winning a new customer divided by the number of new customers won. The decision that matters is what goes in the numerator. Media spend alone gives a platform CAC. Adding agency or freelancer fees, creative production, tooling, and the sales time spent on paid leads gives a fully loaded CAC. Both are useful, but they answer different questions, so label which one you are reporting. Platform CAC tells you whether the channel is efficient; loaded CAC tells you whether the business is.

Payback period is how long it takes the gross margin from a new customer to repay their CAC. A customer who costs 300 to acquire and contributes 50 in gross margin a month pays back in six months (illustrative numbers). ROAS cannot show this because it is a ratio with no time in it. A business with annual contracts and a cash constraint may need a payback of a few months, while a business with strong retention and cheap capital can afford a longer one. That threshold is a finance decision, not a marketing one, so get it agreed in writing.

Be careful with lifetime value. It is a forecast, and forecasts built on a short history flatter the channel. If your customers have been around for eight months, a three-year lifetime value is an assumption, and the report should label it as one. Prefer cohort tables that show actual cumulative margin by months since acquisition, and extend the forecast only as far as the data supports.

Ad platforms report the conversions they can see and credit under their own attribution rules. Your CRM or finance system holds what was actually invoiced. These two will not agree, and the gap is information. Platforms may credit a sale that a different channel started, may credit a conversion that falls inside a long conversion window or under a generous attribution model, depending on the settings you chose, and may miss sales that closed offline or after the tracking window. The CRM may undercount because sources were typed in by hand or never recorded.

The honest approach is to report both and reconcile them on a schedule. Take a period that has fully closed, list revenue by source from the CRM, list revenue the platform claims, and compute the ratio. Do not average the two or pick the larger. Choose the CRM as the system of record for revenue, since it ties to invoices, and treat the platform as the system for fast signals such as spend, clicks, and early conversions. Record the ratio over time. If the platform claims twice what the CRM can trace, your ROAS target should be set on CRM-traced revenue, and the platform number is a leading indicator rather than a result.

Fix the plumbing that makes reconciliation possible. Pass the click identifier or UTM values into the lead record at creation, make the source field a controlled list rather than free text, and record the date a lead became a customer so revenue can be tied back to the month of spend. Without that, any blended number is an estimate with a precise-looking decimal point.

Attribution assigns credit. It does not show what would have happened without the spend. Branded search is the standard example: people searching your name were likely to buy anyway, so a high ROAS there says little about the value of the ads. The same logic applies to retargeting people who were already on their way to purchase. Treating attributed revenue as caused revenue overstates the channel, and the overstatement is largest where the campaign targets people closest to buying.

You do not need a research team to run a sanity check. The practical options, in rising order of effort, are: a before-and-after comparison around a spend change while watching total business revenue, not just attributed revenue; a geographic holdout, where comparable regions get the ads and others do not, for a fixed period; and a platform-provided lift or conversion lift study where the platform offers one and your volume qualifies. Each has limits. Before-and-after comparisons are confounded by seasonality and promotions, geographic tests need regions that behave alike, and platform studies are run by the party selling the ads. State the limitation next to the result.

Use these checks to challenge the biggest assumptions, not to audit every campaign. If you test one thing, test whether branded and retargeting spend is carrying its weight. Report the outcome as a range with the caveats, and be willing to conclude that a test was inconclusive. An honest "we could not tell" is more useful than a confident number that falls apart later.

The weekly report exists to support decisions about spend and creative, so keep it short: spend against plan, leads and cost per lead by campaign, qualified leads where the CRM allows, the one or two changes made last week, and the decisions needed this week. Its audience is the person who can move budget. Put the definitions of every metric at the bottom of the page, including which system each figure comes from, so nobody argues about what a number means during the meeting.

The monthly report is for direction. Include revenue traced in the CRM against spend, ROAS against the break-even line, loaded CAC and payback by channel, the platform-to-CRM reconciliation ratio, any incrementality test results with their caveats, and a short list of what was learned and what will change. Show trends against the previous few months rather than a single month. Mark every illustrative or estimated figure as such, and flag any period where tracking broke. A reader who sees the weak spots disclosed will trust the rest of the report.