On one account I audited, over 90% of the revenue being claimed as incremental was not incremental, just taken from core. The business had paid £25k in commission on sales their existing campaigns were already generating. The reporting made everything look like it was working perfectly.
A growing number of eCommerce businesses are paying a percentage of revenue to incremental Shopping feed tools. The pitch isn't unreasonable: find products that aren't converting, get them in front of buyers, charge a cut of the new revenue. Straightforward enough. The problem is that "new revenue" is being defined in a way that makes the numbers almost impossible to question, unless you know what to look for.
To understand why, you need to see exactly how these tools generate their results. Because the methodology has a structural bias baked in from the start.
How it works
Your core campaign is running well. The large variant of the Classic Cotton T-Shirt is one of your top sellers, converting reliably through Google Shopping.
In the feed this is the Large variant SKU, but the ad shows no size information to the shopper.
The tool scans your feed for SKUs with zero conversions in the last 30 days. The small variant of the same t-shirt hasn't converted on paid. It gets flagged as an "incremental opportunity", duplicated with a slightly different feed label, and pushed into a new Shopping campaign alongside your existing one.
The large variant runs as before. Same ad, same budget.
The small variant is pushed. To the shopper on Google, this ad is identical. Same image, same title, same price. When they click it, they land on the same product page and can select any size they like. The click is recorded against the small variant SKU, so commission is charged regardless of which size they actually end up buying.
The incremental feed cannibalises the core campaign's ad. Shoppers who would have bought through core now land on the incremental listing instead. Total revenue doesn't move. The split does.
The product hadn't suddenly found a new audience. The sale just moved from one campaign to another, and a commission got charged on the way through.
If this only happened occasionally, it might be noise. But there's a pattern in the data that gives it away every time.
What it looks like in the data
The tell is in the timeline. Total revenue for those parent products stays flat, sometimes for months. The incremental campaign's share grows. The core campaign's share shrinks by almost exactly the same amount, in lockstep.
Illustrative example. Total revenue is unchanged. The split shifts after onboarding.