The ROAS report your founder probably shouldn't trust.

Every paid media dashboard lies a little. Not maliciously — the tools are just built to take credit. If your ROAS report says you're pulling 6× on Meta and 9× on Google, and your bank account says revenue is flat, the dashboard isn't wrong so much as it's answering a question no one actually asked.

After auditing dozens of accounts across DTC, luxury retail, and SaaS, we see the same three attribution gaps again and again. Each one inflates ROAS by 30–80%. Together, they can make a merely-fine media program look like a moonshot.

Gap 1: Branded search pretending to be performance.

The single most common distortion we find. Branded search — people typing "your brand name" into Google — gets bucketed into the paid search line item, and its 15–30× ROAS drags the whole channel's blended number into the stratosphere.

These are customers who already know you. They were going to buy. The click tax you're paying Google to not lose them to a competitor bid is a defensive cost, not a growth driver.

Branded search is insurance. Prospecting is investment. Measuring them in the same bucket is like averaging your fire-alarm spend with your R&D budget and celebrating the return.

The fix: split branded and non-branded into separate campaigns, separate reports, separate budget lines. Only the non-branded number should influence a scale decision.

Gap 2: View-through credit on Meta.

Meta's default reporting attributes a conversion to any ad served in the last 7 days — even if the user didn't click, didn't engage, and found you through a completely different channel. The result: Meta takes credit for conversions Google already claimed, podcast listeners already bought, and email subscribers already intended.

In a typical account we audit, turning off view-through attribution drops reported Meta ROAS by 35–55%. The revenue didn't disappear. The double-counting did.

The fix: set your attribution window to click-only, 7-day view. Benchmark incrementality with a geo-holdout test at least once a quarter. If Meta can't prove its revenue exists with the lights on, trim it.

Gap 3: The “last-click but only when convenient” problem.

Your analytics say one thing. Your ad platforms say another. Most teams resolve this by believing whichever number makes the channel they like look best. A $200k month shows $340k across platforms because three channels each claimed 50% credit.

The fix: pick a single source of truth — usually a warehouse or a post-purchase survey — and force every channel to reconcile against it. When Meta claims $500k and the warehouse sees $310k, Meta's wrong. Every time.

What a clean report actually looks like.

• Non-branded paid search broken out separately, with its own ROAS target (usually 2–4× is honest).
• Click-only attribution on Meta, reconciled weekly to a single source of truth.
• Incrementality testing at least quarterly on your biggest two channels.
• Blended MER (total revenue ÷ total spend) reported to the board. Everything else is decoration.

When we do this for a new client, reported ROAS almost always drops in month one. That's the good news — the number finally means something. From there, we can actually go scale it.