Do Audience Signals Actually Matter in Performance Max?

Signals Are Suggestions, Not Targeting

The first thing to understand is that audience signals in PMax are not the same as audience targeting in a Standard Shopping or Search campaign. You are not restricting who sees your ads. You are giving Google's algorithm a starting point — a shortcut to finding the right people faster. Google will use your signals to prioritize early exploration, but it will still serve ads outside those audiences if it finds conversion signals elsewhere.

This matters because a lot of advertisers either over-engineer their signals (trying to replicate the precision of a targeted campaign) or dismiss them entirely ("Google ignores them anyway"). Both approaches are wrong. Signals meaningfully compress the learning period for a new campaign — which for a DTC brand spending $500–$5K/day is a real cost difference.

What Signals Actually Speed Up Learning

Not all signals are created equal. In order of value for most DTC accounts:

• Your own customer lists — purchasers, high-LTV customers, email subscribers. This is by far the highest-quality signal you can give PMax. Google uses these to model who looks like your buyer, and it's specific to your actual business rather than a generic interest category.
• Website visitors with purchase intent — remarketing segments like "visited product pages" or "abandoned cart." These are warm signals that tell the algorithm where buyers congregate before converting.
• Custom segments based on search terms — people who searched for your category keywords (e.g. "running shoes for women" or "organic protein powder"). These proxy intent without requiring you to have a large customer list.
• In-market audiences — useful as a secondary layer, but much lower signal quality than the above. Don't use these as your primary signal.

Do Signals Matter Less as a Campaign Matures?

Yes — and this is the nuance most guides skip. Once a PMax campaign has accumulated several hundred conversions, Google's algorithm has enough real signal from your own account to optimize without relying heavily on the audience signals you provided at launch. At that point, the signals become less influential and the campaign is effectively self-directing based on observed conversion patterns.

This means the impact of your audience signal setup is front-loaded. A new campaign with weak signals (or no signals) will spend more budget learning before it gets efficient. A new campaign with strong, specific signals — especially a customer list — will typically reach target ROAS or CPA faster and with less wasted spend during the learning phase.

Practical Setup for a DTC Brand

For most eCommerce accounts, this is the signal stack we use at launch. Add one asset group per audience theme if your product catalog supports it, and keep signals tight to each theme:

• Purchasers list (all-time, or 180-day window minimum)
• Cart abandoners and product page visitors (30-day window)
• Custom segment: search terms matching your core category keywords
• If you have less than 1,000 in your customer list, add one relevant in-market audience as a secondary signal to compensate for the thin data

Keep the signal inputs reasonably specific. If you sell one product line, a single tightly defined set of signals per asset group is better than a wide collection of loosely related audiences.

Bottom Line

Audience signals in PMax do matter — most for new campaigns, less for established ones with strong conversion history. The biggest lever is uploading a real customer list before you launch. Everything else is secondary. If your PMax campaigns are already live and performing, revisiting signals is unlikely to move the needle much. But for any new campaign or account build, spending 20 minutes on signal quality upfront will save you weeks of inefficient spend in the learning phase.