Performance Max can report growth while recycling demand unless teams isolate new customers
Search Engine Land’s Performance Max framework is a paid media article, but the strategic issue is broader: when attribution systems recycle existing demand, teams misread whether search, paid media and AI visibility are creating new buyers or just claiming them.
Editor’s Note by Dageno: Dageno AI’s editorial team structurally distilled Search Engine Land’s article “How to make Performance Max focus on net new customers” to help GEO, paid media and analytics teams understand why customer acquisition measurement must separate incremental demand from recycled warm traffic.
Understand the article in 60 seconds
- The article warns that DTC brands can show strong Performance Max ROAS while paying Google and Meta for the same conversion; for AI visibility teams, the same problem appears when branded demand created elsewhere is misattributed to the last measurable channel.
- Google’s newer controls combine brand exclusions, first-party audience exclusions and Customer Match data so Performance Max can be pushed closer to net new customer acquisition.
- The four-step framework excludes brand queries, website visitors, email subscribers and existing purchasers, then uses New Customer Bidding to bias the campaign toward people who have not meaningfully interacted with the brand.
- The author cautions that matching is never perfect and “new customer” value can inflate reported revenue if teams add fictional value inside bidding settings.
The article is really about incrementality, not only Performance Max settings
The trap is that dashboards can look healthy while contribution margin weakens. When Google, Meta, TikTok, YouTube, Gmail, Discover and Display touch the same customer, platforms may all claim credit even when the order would have happened anyway.
For Dageno readers, this has a GEO parallel: AI search, content, paid search and social can all create or capture demand. If attribution only rewards the last platform, teams will underfund the visibility surfaces that shaped the buyer’s decision.
Performance Max naturally chases warm traffic unless guardrails exist
The article says Performance Max has historically defaulted to the warmest signals: branded search, site visitors, email subscribers and existing customers. Removing brand searches alone does not change the audience if YouTube, Gmail, Discover and Display can still retarget warm users.
That is the key mechanism. Channel control requires both query-level and audience-level exclusions. Otherwise PMax can keep claiming demand created by Meta, organic content, email or brand equity.
Query guardrail
Brand exclusions and phrase-match negatives stop PMax from harvesting branded search where intent already exists.
Audience guardrail
Visitor, subscriber and purchaser exclusions reduce retargeting disguised as acquisition.
Value guardrail
New Customer Bidding should be configured so reported value does not obscure real revenue and margin.
Brand exclusions are the foundation, but they are not enough
The framework recommends brand exclusions once brand searches contribute more than 15% to 20% of cost or revenue. It also recommends adding the brand as a phrase-match negative keyword because brand lists are not foolproof.
The article adds a necessary caveat: if PMax excludes brand terms, dedicated brand Search and brand Shopping campaigns should capture that demand so competitors do not take it.
First-party audience exclusions shift PMax away from retargeting
The newer audience exclusion option lets advertisers exclude website visitors, email subscribers and other warm audiences. The article mentions Google Ads pixel, Google Analytics and ESP integrations such as Klaviyo, with real-time syncing as a benefit.
The most useful version is business-specific. Instead of excluding all visitors from the last seven days, the article recommends using an engagement metric that fits the account, such as cart adds.
Customer Match reduces existing-customer bleed but cannot remove it completely
Existing purchasers can be excluded through purchase-event audiences or customer list uploads. Shopify and Klaviyo can help feed Customer Match lists into Google.
The article warns that matching is never 100%. If Google matches 900 of 1,000 customers, some bleed remains. But reducing existing-customer targeting from everyone to a small unmatched slice can materially improve acquisition quality.
New Customer Bidding can improve focus and distort ROAS at the same time
The campaign setting offers two modes: bid only for new customers, or bid higher for new customers. The first is more aggressive. The second adds a chosen value to new-customer conversions.
The danger is reporting inflation. If a $200 sale receives an extra $100 new-customer value, Google may report $300 in revenue. That fictional value can distort ROAS and target ROAS decisions.
Net-new acquisition requires paid, GEO, analytics and finance to share one definition
| Owner | What changes | What to do next | Metric or signal |
|---|---|---|---|
| Paid media team | PMax must be configured as prospecting, not hidden retargeting. | Apply brand, visitor, subscriber and purchaser exclusions, then test New Customer Bidding. | New-customer share, CAC, contribution margin |
| GEO / SEO team | Organic and AI visibility may create demand later claimed by paid channels. | Track branded demand lift and AI answer exposure around paid tests. | Branded search lift, AI answer-share |
| Analytics team | Platform ROAS can be inflated by overlap and fictional new-customer values. | Build new-versus-returning and incrementality dashboards outside ad platforms. | Matched customer rate, MER, incrementality |
| Ecommerce growth | Customer lists and event quality determine how clean exclusions can be. | Maintain purchase events, Customer Match feeds and ESP syncs. | List freshness, match rate, bleed rate |
| Finance / leadership | Lower ROAS can be healthier if customer mix improves. | Evaluate contribution margin and net-new acquisition, not platform-reported ROAS alone. | Gross margin, payback, net-new revenue |
The right test is new-customer mix, not reported ROAS
The article expects reported PMax ROAS to fall after exclusions, while the new-versus-returning split should align more closely with actual acquisition. It also notes that without advanced configuration, even good setups may still look like a 60/40 split.
Dageno’s interpretation: the winning KPI is not prettier ROAS. It is whether the campaign produces a higher proportion of real first-time buyers, better contribution margin and cleaner channel roles.
| Signal | Why it matters | Owner | Decision it supports |
|---|---|---|---|
| New-customer share | Shows whether PMax is acquiring or recycling customers. | Paid / Analytics | Whether exclusions improved acquisition |
| Customer Match rate | Shows how much existing-customer bleed remains. | Analytics / Ecommerce | Whether list quality needs repair |
| Brand search cost share | Signals when brand exclusions become necessary. | Paid media | Whether PMax is over-harvesting brand demand |
| Contribution margin | Reveals whether reported ROAS translates into profit. | Finance | Whether spend should scale |
Use exclusions to make paid acquisition measurable before scaling spend
Separate brand demand
Create brand exclusions and phrase-match negatives, then protect brand demand with dedicated Search and Shopping campaigns.
Exclude warm audiences
Remove website visitors, subscribers and meaningful engagers based on account-specific signals.
Exclude purchasers
Use purchase-event audiences and Customer Match feeds from Shopify, Klaviyo or the CRM.
Test new-customer bidding
Compare new-customer share, margin and real revenue before judging by ROAS.
Turn AI visibility signals into execution.
Dageno AI is a complete execution platform from data monitoring → strategy → content generation → result attribution. It helps teams monitor AI answer surfaces, detect strategy gaps, generate content actions and connect visibility changes to business impact.
Alexa for Shopping