Performance Max is a single Google Ads campaign type that runs across Search, Shopping, Display, YouTube, Gmail and Maps using one budget and one set of creative. Google's algorithm decides where each impression goes. That's the whole pitch, stripped of marketing.
Search campaigns are the older shape — keyword-targeted, text-only, run on the search results page. You know what you're paying for. You can see the search term. You can pull a lever and watch what happens. PMax hides most of that.
Three SMB profiles where PMax beats Search
PMax is strong in three scenarios. Each has a thing in common — there's enough conversion volume for Google's algorithm to learn from.
- E-commerce with a product feed : Shopify or WooCommerce stores doing $30k+ a month, with 40+ products and a clean Merchant Centre. Conversion volume teaches the algorithm fast.
- Multi-location services with strong brand search : if people already search your name in volume, PMax extends the reach without competing against your own Search campaigns.
- Mature accounts with 50+ conversions a month : the floor for the algorithm to do anything useful. Below that, PMax burns budget guessing.
Four where it doesn't, and what to run instead
PMax is wrong in more situations than Google's reps will admit. We won't run it for a small services business with under 30 conversions a month — the maths doesn't work yet.
- Lead-gen under 30 conversions a month : run tight Search campaigns instead. You'll see the keywords, you can iterate.
- High-consideration B2B : Search with negative keywords and audience layers outperforms by 40%+ on cost per lead in our accounts.
- Brand-protection campaigns : Search exact-match on your brand terms, cheap clicks, full control. PMax cannibalises this and obscures the cost.
- Anything with no conversion tracking : PMax is a black box. Without conversions feeding it, the box is empty.
The asset group lever
The one PMax lever worth pulling is the asset group split. Most accounts run one asset group, which means one creative pool, one audience signal, one budget across everything. Split by product line or service type and the algorithm gets clearer signal. We've seen cost per conversion drop 30% from this single change.
The 30-day test we run when we're unsure
When the answer isn't obvious, we run both for 30 days with separate conversion goals — Search for keyword-targeted intent, PMax for everything else. After 30 days the data picks the winner. No campaign survives on opinion past day 30.
This is the same logic behind how we run paid media generally — tests with a deadline, decisions tied to a number. The deeper pricing logic sits in the ad spend markup note.
The takeaway
PMax isn't a default. It's a tool that earns its budget when conversion volume justifies handing the levers to an algorithm. Below that threshold, Search beats it every time, with full visibility into where the money goes. If you can't tell which one your current account is running, that's the first thing to ask.