How branded search bleed happens
Someone who already knows your business searches your name. That click is cheap, converts well, and would very likely have happened anyway through the organic result. When an automated campaign is allowed to match those searches, it takes them — and its reported return climbs, because it is buying the easiest traffic there is.
Broad match and automated campaign types make this hard to avoid by default. Performance Max in particular has no keyword list to inspect, so brand queries can flow into it quietly while a dedicated brand campaign sits alongside doing the same job. The money is still spent. It simply moves between campaigns, and the campaign it moves to looks like the star of the account.
Why branded search bleed matters
It corrupts the decision the numbers are used for. A campaign reporting a strong return looks like the obvious place to add budget, so more money goes to the campaign that is hardest to grow and least likely to introduce a new customer. Meanwhile the non-brand campaign that actually creates demand looks expensive and gets cut.
It also hides performance from whoever is paying. Blended figures rise, the account looks efficient, and nothing in the report mentions that the growth came from people who had already decided to buy. Questions about incrementality — what the advertising added that would not have happened anyway — are exactly the questions this masks.
Where it goes wrong
The most common error is reporting a single blended return without separating brand from non-brand. The second is the opposite overreaction: switching brand coverage off entirely on the grounds that organic will catch it. Sometimes it does. When a competitor bids on your name, or your organic listing sits below a block of ads, it does not — and that is a test to run, not an assumption to hold.
A third is assuming a setting fixes it. Brand exclusions help, but they act on a brand list you maintain, and misspellings, product names, former trading names and staff names routinely slip through.
What to do about it
Separate the demand first. Keep a brand campaign with exact and phrase coverage of your own name, and add your brand terms as negatives in the non-brand campaigns so the split holds. Apply brand exclusions to automated campaigns, then read the search terms report for the variants that were missed and add those too.
Report the two lines separately from then on. Non-brand cost per acquisition tells you whether the account is finding new customers. Brand cost per acquisition tells you what it costs to defend ground you already hold. Both are useful; the average of them is not. If you want to know whether the brand spend is adding anything at all, pause it in one region for a full sales cycle and watch total orders in your own system rather than platform-reported conversions.