01What the 50% number actually means
Across 300+ DTC accounts Tegra has audited since 2017, the median PMax campaign now sends 42% of its impressions to branded search queries. The top quartile is at 61%. For a brand spending $50K/mo on PMax, that translates to roughly $21K being spent on queries the user was already going to type.
Here's what makes this insidious: PMax doesn't show you the branded share in its native reporting. You have to derive it by joining search-term data against your brand list — a query that takes 30 seconds to write and that almost no agency runs.
The number itself isn't the problem. The problem is that nobody is reporting on it, so nobody is acting on it.— Operator audit log · Q1 2026
02How to diagnose in 4 minutes
The exact query we run on every new audit. Works in the standard Google Ads UI; no Looker or API required.
If branded share exceeds 35%, you have the problem. If it exceeds 50%, you have a crisis. Most DTC operators discover they're in the second bucket the first time they run this.
- Open Google Ads → Insights → Search terms (the PMax-specific view, not the legacy one).
- Filter to the last 14 days, segment by asset group.
- Export to CSV. Add a column: is_branded = "yes" if the term contains your brand name, hero-SKU name, or any obvious misspelling.
- Pivot impressions and conversions by is_branded.
Brand exclusions on every asset group, refreshed monthly — not added once and forgotten. Add your brand spellings, your hero-SKU names, and any creator collab names the moment they go live.
03The fix · brand-exclusion architecture
The instinct most operators have is to 'just add a negative.' That works for the obvious brand spellings, but it misses three patterns that account for most of the leak.
Run a search-term audit weekly for the first month after applying the list, then monthly.
04Real account · what changed
One Tegra client — a supplements brand at $80K/mo in Google Ads spend — came in with their Shopping ROAS drifting from 3.2× to 1.9× over nine months while the team kept pouring budget in. Their branded share was sitting at 61%.
We re-applied brand exclusions across 4 of 6 asset groups (the others already had them) and rebuilt their attribution windows. Ten weeks later, Shopping ROAS was back to 3.1× and total revenue had climbed from $22K/mo to $42K/mo without changing spend.
There was nothing magical about the work. It was a structural fix that took an operator 90 minutes to design and an agent 4 minutes to deploy. That's the entire pattern.
We'd been at 1.9× for six months. Three weeks in and we're at 3.1×. The agent caught a feed schema gap nobody on our team would have ever found.— Head of Growth · supplements brand
05TL;DR
If you take exactly one action this week: open your PMax search-term report, segment by asset group, and check the branded share. If it's above 35%, you have homework to do.
The diagnostic is free. The fix is structural. The agent runs it in 4 minutes.
Ruslan co-founded Tegra in 2017. Runs the Google Ads practice — feed, PMax, search, attribution. Writes weekly about the parts of paid search operators are afraid to touch.