Who this is for, and who it is not for
- For CMOs seeing an 'efficient' PMax while Search volume and the P&L look stuck.
- Not for first-time PMax launches, where there is no baseline to compare against.
- Not for channel-mix planners working outside Google.
What cannibalization looks like in the P&L (not just ROAS)
Inside the platform, cannibalization is invisible: PMax reports strong returns because it claims credit for conversions it touched, wherever the demand came from. In the P&L it shows up differently — blended cost per qualified outcome stays flat or rises, Search volume erodes, and 'growth' in one campaign is matched by decline in another.
The tell is the total. If PMax conversions climb while total account conversions and pipeline do not, budget is moving between pockets. Nothing incremental is happening, and the dashboard says otherwise.
Brand absorption vs net-new demand
Branded queries are the cheapest conversions in the account because the demand already exists. When PMax is allowed to serve on them, it buys that demand back at auction and reports it as performance. The account pays twice for a conversion it already owned.
Net-new demand looks different: non-brand queries and placements where the account had no prior presence, rising total conversion volume across the account, and Search holding or growing its own volume while PMax scales. The distinction between the two is the entire question.
A practical cannibalization test
This is a qualitative read, not a lift study with invented percentages. Three comparisons, run before and after any PMax budget change:
- Branded Search volume and CPA before and after PMax scales. If brand impressions or clicks fall as PMax budget rises, PMax is absorbing them.
- Search term overlap. Where the account's data exposes it, check whether PMax is serving on queries Search already covers, exact-match terms included.
- Total conversions versus lane shifts. If total qualified outcomes rise, the budget change worked. If conversions only move from the Search column to the PMax column, it did not.
Brand exclusions and Search protection
The principle: branded demand should be captured by the cheapest, most controllable lane — brand Search — and excluded from PMax wherever the platform supports it, through brand exclusions and account-level negatives. Google moves these controls periodically, so the specific settings matter less than the discipline of keeping brand out of PMax's reach and reporting it separately.
Protection also means budget. Brand Search should never be the campaign that runs out of money because a discovery lane overspent.
When PMax is still the right lane
None of this argues against PMax. Run as one controlled lane — clean feed, margin-aware asset groups, brand held out, and changes measured against total account outcomes — it is a good allocation engine. How SalesX runs it is covered on the Performance Max page: /services/performance-max.
Related allocation framework
Cannibalization is one symptom of a larger question: how budget should be divided among Search, Shopping, and Performance Max at $50k+/month. The role-based framework and guardrails are covered in the budget allocation guide: /google-ads-budget-allocation-search-shopping-pmax.
Get an X-Audit on Search/PMax overlap
The free X-Audit reviews your live Search/PMax mix: brand coverage and exclusions, search-term overlap, conversion definitions, and whether total outcomes rise when PMax scales. You get the findings whether or not you become a client; ongoing management is optional.
Questions we get
- Does a strong PMax ROAS prove incrementality?
- No. ROAS inside PMax counts every conversion the campaign touches, including branded and existing-demand queries that Search would have converted at a lower cost. Incrementality is a question about total account outcomes, not about one campaign's reported column.
- Should I pause PMax to test?
- A full pause is a blunt instrument and usually not necessary. Start with brand exclusions and account-level negatives where supported, then watch Search volume, impression share, and total qualified outcomes as PMax budget changes. If PMax gains are matched by Search losses, you have your answer without taking the account dark.
- Can PMax replace Search?
- No. Search is where intent is declared and where query-level control still changes outcomes. PMax has no query layer to manage. How Search is run at spend is covered on the Google Search Ads page: /services/google-search-ads.
- Is this a Meta comparison?
- No. SalesX manages Google Ads exclusively. This page is about the overlap between two Google campaign types, not about Google versus Meta or any other channel.