The Ecomm Analyst

A few months back I was helping a brand evaluate their Meta creative. They had one ad that was clearly the winner. Inside Ads Manager it was crushing: 4.2 ROAS, low CPM, scaling cleanly. Their media buyer wanted to triple the budget on it the following week. It was the obvious call.

Then we pulled the same ad up in our independent attribution tool, and the picture flipped. On a multi-touch model that gave appropriate credit to other channels, that ad’s actual contribution to incremental revenue was sitting around 1.4 ROAS. It was getting last-click credit for conversions email had been warming up for weeks. The “best ad in the account” was, in reality, a mediocre middle-funnel asset that Meta was happily taking credit for.

We didn’t triple the budget. We left it where it was, and shifted the proposed spend to a creative Meta was reporting at 1.8 ROAS that was actually doing 3.1 on a multi-touch view. Within a month, new customer revenue from paid social was up about 22% on roughly the same total spend.

This is the case for ad-level attribution.

Most operators have made peace with channel-level attribution at this point. They know Meta over-attributes. They know Google over-attributes. They run blended ROAS as a sanity check and use a tool to pick a single source of truth. That’s good. It’s also not enough.

Channel-level data tells you how much to spend on Meta versus Google versus email. It doesn’t tell you which Meta ads to scale and which to kill. And the answer to that question, if you take it from inside Ads Manager, is going to be wrong in a predictable way: the platform will always favor the assets that converted the most users on its own pixel, which has nothing to do with which assets are actually pulling new customers in.

A few things ad-level attribution lets you see that channel-level can’t:

Which creatives are acquiring versus which are closing. A retargeting ad will always look great on platform-reported ROAS because everyone who sees it has already shown intent. Strip that out and ask which prospecting ads are bringing in genuinely new customers, and your top-performer list usually changes.

Which ad sets are duplicating credit. When two ad sets are running against overlapping audiences, both will claim the conversion. Independent attribution distributes credit across the actual customer journey, so you can see that one of them might be doing 80% of the real work and the other is just collecting receipts.

Which creatives have long conversion windows. Some ads convert in 24 hours. Some convert in 14 days through three other touches. Inside Ads Manager you can’t tell the difference. With proper multi-touch attribution at the ad level, you can.

The reason this is hard is that ad-level data needs to be stitched across platforms. You need the click on the Meta ad, the email open three days later, the Google search that finally drove the purchase, all tied back to a single customer. Most attribution tools stop at the channel level because going deeper requires actual customer-level identity resolution, not just UTM tagging.

ThoughtMetric is the tool I’ve been using for this. It exposes attribution down to the ad and ad set level, with the same multi-touch logic applied consistently across every channel, so you can compare a Meta ad against a Google search ad against an affiliate placement on equal footing. That comparability is the part platform-native reporting can never give you, because each platform is grading its own homework.

If you’re already past channel-level attribution and feeling like you’ve solved the problem, my honest read is that you’ve solved the easy half. The decisions that actually move the needle on creative spend live one level deeper.