Audience Signals (Performance Max)
First-party and interest-based inputs D2C brands feed into Google Performance Max campaigns to guide the AI toward likely converters—without restricting who can see the ads.
Performance Max gives you one meaningful configuration input during campaign setup: audience signals. The field looks like targeting. Signals are training data for Google’s AI, not restrictions on who sees your ads.
How it shows up in the wild
Google (Performance Max platform) (official signal behavior documentation): Audience signals are “suggestions, not restrictions” — the algorithm expands beyond them to reach any user it predicts will convert, per Google’s official Help Center. Signal types the platform accepts include Customer Match lists, remarketing audiences, custom intent segments, and similar segments. New signals take up to two weeks to fully integrate into Google’s models.
KEH Camera (used camera reseller; Performance Max rollout managed by Inflow, August 2022–Q1 2023): Inflow structured audience signals as three tiers of purchaser seed lists — one-time buyers, loyalists, and champion customers — each placed in its own asset group. The campaign achieved a 76.3% increase in advertising revenue in Q1 2023 vs. the prior-year period on Standard Shopping, and averaged 9.93x monthly ROAS in the final phase of the rollout, per Inflow’s published case study.
Tinuiti (Q1 2026 Digital Ads Benchmark, tracking thousands of advertiser accounts): Performance Max captured 67% of Google Shopping ad spend among advertisers running both PMax and Standard Shopping, and delivered slightly better ROAS than Standard Shopping in Q1 2026, per the Q1 2026 Benchmark Report. At that level of adoption, audience signals become the primary differentiator between PMax campaigns — most other targeting levers either don’t exist in this campaign type or are controlled at the product feed level.
Why it matters
Performance Max is now the dominant format for Google Shopping. Brands feeding the system purchaser lists and Customer Match data see faster learning periods and lower early CPAs than those relying on interest or affinity segments.
The signal quality gap is real. My hunch is that “suggestions, not restrictions” is the most misunderstood thing about PMax — brands that expect signals to act like ad group targeting misread what’s happening during the learning phase.
Related terms
Frequently asked questions
Do audience signals restrict who sees my ads in Performance Max? No. Google is explicit: signals are suggestions, not targeting rules. The algorithm uses them to initialize its buyer model, then expands beyond them — unlike standard ad group targeting, where delivery can actually be restricted.
What type of audience signal performs best? For ecommerce, purchaser lists consistently outperform interest and affinity signals. A Customer Match list seeded from past buyers gives the algorithm real conversion signal. Interest segments give it browsing behavior.
Why bother with signals if the algorithm expands beyond them anyway? Signals accelerate the learning period. A PMax campaign with no signals starts cold — the algorithm discovers what a buyer looks like from scratch through trial and spend. With a purchaser list as a signal, the algorithm starts with a working model and tests from there.