The Performance Max brand leak: how to size it

Performance Max reports the best return in the account — and that is exactly what should make you suspicious. In most ecommerce accounts a meaningful share of that result is bought on queries containing your own store name — orders that would have arrived anyway. The reported number goes up, incremental sales do not.

PMax is a bidding system pointed at the target you set, and it takes the cheapest route there — no bad faith involved. Brand queries convert several times better and cost less per click, so your own name is the most efficient inventory it can find. The report then blends demand you created with demand you already had — and you cannot set a target against a blend.

Joshua Uebergang laid out the step-by-step audit on PPC Hero. It takes one working day:

• export the PMax search terms report and segment it by ad format;
• classify terms with a regex into five buckets: pure brand, brand+product, brand+qualifier, brand+competitor, nonbrand;
• calculate two numbers — brand share of cost and brand share of conversion value;
• rebuild the return without brand — that figure, not the blend, is what a target belongs against;
• corroborate in Search Console: flat organic brand demand with rising paid clicks tells the story.

Then the levers. Negative keywords cover Search and Shopping, at campaign or account level. Brand exclusion lists also reach YouTube search, with a checkbox that keeps Shopping ads on brand queries — for retailers, the key setting here. Sequence is critical: launch the brand campaign first, confirm it is serving, then apply the exclusion. Reverse that and the top of your own results page goes to whoever was bidding second.

For the budget: campaign return will drop once brand comes out, and that is expected, not a failure. Read total account spend, revenue and orders over at least four weeks. Revenue holding while spend falls means the money really was leaking.