Revenue attribution vs campaign optimization: which report answers your question
In revenue attribution vs campaign optimization, the difference is the question. Attribution asks which marketing brought in your paying customers. Optimization asks where your next dollar should go. Run the wrong one and you get a confident answer to a question nobody asked.
- Revenue attribution
- Looks back. Gives each payment one owner, so every sale is counted once.
- Campaign optimization
- Looks ahead. Compares what each campaign cost with what it returned, to decide what to grow, pause or fix.
- Use both
- When you are deciding whether a channel is worth keeping, and then how to spend inside it.
Good marketing attribution reporting starts by naming the question. This guide defines both reports in plain terms, runs the same Stripe and Meta Ads data through each, and gives you a simple way to decide which attribution report to use.
What revenue attribution actually measures (and what it cannot)
Revenue attribution answers one question: which marketing brings in paying customers? It starts from the money. Each payment is traced back to the marketing that customer touched, so the total of all channels adds up to the revenue you actually collected.
That fixes a real problem. Every ad platform takes credit for the same customer. Google, Meta and your email tool can each claim one sale, and they cannot all be right. Attribution gives you one count.
What it cannot tell you
- Whether the sale would have happened anyway. A customer who already planned to buy may still click an ad on the way.
- What happens if you spend more. Credit for past sales is not a forecast.
- Anything about sales it cannot connect to a touchpoint. To attribute, you need a link between a click and a payment.
What campaign optimization measures: spend efficiency, not credit
Campaign optimization answers a different question: which campaigns should you grow, and which should you stop? It starts from the spend. Each campaign's cost sits next to what it produced, so you can compare them on the same scale.
It is about efficiency, not credit. A campaign can earn little credit and still be your cheapest way to find new customers. Another can collect plenty of credit by showing ads to people who were buying anyway. The useful measures are:
- Kept revenue per dollar spent, after refunds.
- Cost per new paying customer.
- The same two numbers at the ad set and creative level, where most decisions get made.
Same data, two conclusions: a worked Stripe + Meta Ads example
Here is one month of two Meta Ads campaigns, matched to Stripe payments by the customer's email. The numbers are a made-up illustration, not any client's data.
| One month | Campaign A: new audiences | Campaign B: past visitors |
|---|---|---|
| Ad spend | $2,000 | $1,000 |
| Purchases Meta reports | 30 | 40 |
| Paying customers matched in Stripe | 16 | 20 |
| …of whom had paid you before | 1 | 15 |
| Charged in Stripe | $3,200 | $3,000 |
| Refunded | $400 | $0 |
| Kept revenue | $2,800 | $3,000 |
The attribution reading
Campaign B touched more paying customers (20 against 16) and more kept revenue ($3,000 against $2,800). It cost half as much. As a record of where this month's money came from, B looks like the winner.
The optimization reading
Look at new customers instead. Campaign A found 15 new paying customers for $2,000, about $133 each. Campaign B found 5 for $1,000, $200 each. Most of B's revenue came from people who had paid you before, and some of them would likely have bought without the ad. If the goal is growth, A is where the next dollar goes.
Both are right
Neither report is wrong. They answer different questions. The only way to settle how much of B's revenue was extra is a test: pause B for a period, or hold some of its audience back, and see what happens to their purchases. Note also that Meta reported 70 purchases while Stripe shows 36 paying customers. That gap is the problem attribution exists to fix.
A decision tree: pick your report from the question you were asked
Start from the question you were asked, not from the report you usually run:
| The question | The report | What it needs |
|---|---|---|
| Where did last quarter's revenue come from? | Revenue attribution | Every payment, and the marketing each paying customer touched first. |
| Our ad platforms claim more sales than the bank shows. Which is right? | Revenue attribution | Payments matched to clicks, so each sale is counted once. |
| Should we move budget from one campaign to another? | Campaign optimization | Spend by campaign next to kept revenue and new customers. |
| Which ad set or creative should we pause? | Campaign optimization | Spend and outcomes at the ad set and creative level. |
| Is this channel worth it at all? | Both, in order | Attribution first, to see if it brings paying customers. Then optimization inside it. |
If the question is “what happens if we spend more?”, neither report answers it alone. Use optimization to pick the campaign, then test a higher budget on it.
Where each analysis breaks down (short windows, offline conversions, B2B lag)
Short windows
Meta counts conversions under its own window: seven days after a click and one day after a view, by default. Google Ads credits a conversion back to the day of the click, so recent days keep rising. A sale that lands after the window drops out of the platform's count. Matching to Stripe catches it, but the longer the gap between click and payment, the weaker the link.
Offline conversions
Sales by phone, by invoice or at a counter never reach the ad platform's tag. Stripe sees the payment if it runs through Stripe, but the match still needs an email or phone on both sides. If customers buy in person, read our guide to stitching online and in-store identity.
B2B lag
When a deal closes weeks after the first click, optimization on this month's spend sees cost and no revenue yet. Judge campaigns on qualified leads in the meantime, and run attribution on a longer look-back once deals close. LinkedIn, for example, counts conversions within 30 days of a click by default and never sees whether a lead became a deal. Our post on LinkedIn Ads and HubSpot attribution covers that case.
Refunds and payout dates
Stripe records each charge and refund as it happens, and payouts land days later. Use the charge date for both reports, and subtract refunds, or a campaign that sells and then refunds will look better than it is.
How to run both on one unified dataset
The two reports should never disagree about the facts, only about the question. That happens when both are built from one table:
- Pull spend by campaign, ad set and creative from Meta Ads.
- Keep the campaign tags and click IDs on every visit, so a click can be tied to a person.
- Pull charges and refunds from Stripe, with the customer's email.
- Clean the email on both sides, lowercased and trimmed, and match on it.
- Build one row per customer: first touch, later touches, payments, refunds, and whether they had paid before.
- Read attribution off the payments, and optimization off the spend, from the same rows.
askotter does this with live connections to Stripe and Meta Ads. See the revenue attribution and campaign optimization analyses for that pair, or use the marketing ROI calculator for a quick first pass with your own numbers. A person reviews every suggestion before it changes your budget.