Marketplace Ads and the Incrementality Trap: Rethinking CAC
Blended CAC hides more than it reveals. A practical framework for measuring what your marketplace ad spend is actually buying.
8 min read
Marketplace advertising has a structural problem most dashboards quietly paper over: a meaningful share of 'converted' ad clicks would have converted anyway, because the shopper was already searching for your brand or category on that platform. Blended CAC treats every conversion as incremental. Usually it isn't.
Branded vs. non branded is the first cut
The single highest leverage change most teams can make is splitting spend and reporting between branded and non branded keywords or placements. Branded search ads on a marketplace often just win a sale you'd have gotten organically anyway. You're paying rent on your own demand. Non branded placements are where genuine incrementality tends to live, because you're winning attention you wouldn't otherwise have had at all.
Geo holdouts beat attribution models
Platform reported attribution is built by the same platform selling you the ads, so treat it as directionally useful, not authoritative. A geo holdout test, where you pause spend in a subset of comparable markets and compare sales deltas, gives a cleaner read on true incrementality than any last click or even multi touch model a marketplace hands you natively.
What a healthier CAC framework looks like
- Report branded and non branded CAC separately. Never as one blended number.
- Run periodic geo or platform holdouts to sanity check platform reported incrementality.
- Weight budget allocation toward channels with higher measured incrementality, not just a lower reported CAC.
None of this means marketplace ads aren't worth running. For most brands they clearly are. It just means the number on the dashboard and the number that reflects true incremental growth are rarely the same figure, and only one of them should actually be driving budget decisions.