Understanding Meta Attribution Settings (So You Can Trust Your ROAS Again)

I was inside an ad account this week and pulled up the attribution comparison panel. You know the one — the little box where you tick 7-day click, 28-day click, 1-day view, all of it. And it reminded me how many people run real money through Meta every month without ever knowing what those settings are quietly doing to their numbers. So let's sort that out. No jargon dump, just how I actually read it when I'm looking at an account.

What are Meta's attribution settings actually doing?

Short answer first — they decide which conversions Meta is allowed to take credit for. That's it. See that line in the panel that says selections here are for reporting only and don't change optimisation? That's the most important sentence on the whole screen. Changing your attribution window does not make your ads perform better or worse. It only changes the story Meta tells you about what already happened. Makes sense right.

What does 7-day click attribution mean?

This is the default, and most accounts are reading it without realising. 7-day click means if someone clicks your ad and then buys anytime in the next seven days, Meta claims that sale. Even if they clicked on Monday, forgot about you, came back through a Google search on Thursday and bought then — Meta still raises its hand and goes that one was mine. For a lot of D2C brands with a quick consideration window, that's roughly fair. For higher-consideration stuff, B2B, anything where people sit and think for a while, that seven-day window starts scooping up sales the ad only partly nudged.

Should you turn on 1-day view attribution?

Be careful with this one. 1-day view counts a conversion when someone just saw your ad — didn't click, didn't tap, it simply scrolled past in their feed — and then bought within a day. View-through attribution is exactly where reported ROAS quietly puffs up. I've watched accounts where flipping on view-through added what looked like a lovely lift, the founder got excited, and then the actual revenue didn't move at all. Because those buyers were going to buy anyway. The ad got served, it counted the sale, but it didn't cause the sale. Use view-through to understand reach and brand effect — not to decide whether a campaign is pulling its weight.

All conversions or first conversion — which one?

All conversions counts every purchase inside the window. First conversion counts only the first. If you sell something people buy again and again, all conversions will stack several sales onto a single ad click and your ROAS will look bigger than the new revenue the ad actually brought in. When you're trying to judge whether an ad is winning you new customers, first conversion is the more honest read. For a total revenue picture, all conversions is fine — just be clear in your head which question you're answering before you celebrate the number.

Why does my Meta ROAS look better than my bank account?

This is the bit nobody enjoys, but it's the whole game. Meta is grading its own homework. Every setting in that panel is a lever that lets the platform claim a bit more credit — last-touch, view-through, longer windows, all-conversions. Each one pulls your reported ROAS up without a single extra sale happening in the real world. I had a note in my own journal months ago, watching people optimise for add to carts or for traffic and then quote a ROAS like it meant something — but how are you tracking the actual sales behind that number? You can't be delusional in performance marketing. It's all numbers. You have to keep asking why a figure is what it is, and solve it like a maths problem from a few different angles.

The fix I keep coming back to is MER — marketing efficiency ratio. Total revenue divided by total ad spend across everything you're running. It doesn't care which platform claims what. For early-stage brands especially, MER is a far more honest signal than the ROAS Meta hands you, because it's measured against the one number that can't be massaged: what actually landed in the bank.

So how should you actually set this?

Practical version. Keep 7-day click as your main reporting view so you're at least comparing like with like over time. Pull up 1-day click every now and then as your conservative read — that's much closer to what the ad almost certainly caused on its own. Treat view-through as context, never as proof. And sanity-check the whole lot against MER once a month. When your 7-day-click ROAS says one thing and your MER says another, MER wins. Every single time.

All of this matters because the numbers are meant to help you make decisions, not flatter you. If you're staring at a Meta dashboard that looks great while the revenue doesn't quite add up, that gap is very often hiding in exactly these settings — and it's genuinely fixable once you know where to look. This is the kind of thing I untangle for clients all the time, so if you'd like a second pair of eyes on your account, book a free strategy call.