Google Just Rewired How Demand Gen Finds New Customers — What DTC Brands Need to Know

What Actually Changed in Demand Gen Lookalike Targeting

Previously, Demand Gen Lookalike segments worked the way most advertisers expected: you'd upload a seed list (existing customers, purchasers, email subscribers), set a reach slider — narrow (2.5%), balanced (5%), or broad (10% of the target location) — and the system would cap delivery to users within that similarity threshold. Tight control, predictable audience size.

Starting March 15, 2026, that model changed. Lookalike segments now operate as audience suggestions, not hard targeting constraints. Your seed list and reach level tell Google AI who to prioritize — but the algorithm is no longer restricted to a defined similarity band. It can go wider if it thinks that's the path to conversions.

The reach slider still exists and still has meaning: narrow pushes the algorithm toward users most similar to your seed list, balanced gives it more flexibility, broad gives it the most room to roam. But none of these settings cap your audience the way they used to.

Why This Matters for DTC Prospecting

Demand Gen has quietly become one of the better prospecting channels for DTC brands running Google Ads. YouTube, Discover, and Gmail inventory at reasonably efficient CPMs, with visual creative formats that work well for product-forward brands. A lot of accounts have built their upper-funnel strategy around tight Lookalike targeting as a way to stay efficient while still reaching new users.

That approach needs to be recalibrated. With Lookalikes as suggestions rather than constraints, your audience can expand significantly — which means your CPMs and CPCs could shift, your CPA might move in either direction, and the users you're reaching are no longer guaranteed to sit within a defined similarity band relative to your best customers.

What You Should Do in Your Account Right Now

A few concrete steps worth taking this week:

• Audit your Demand Gen performance post-March 15. Look at CPM, CPA, and conversion rate trends before and after. If volume went up but CPA degraded, the audience expansion is the likely cause.
• Review your seed list quality. Since Google AI is now using your seed list as a signal rather than a strict template, the quality of that list matters more than before. A seed list heavy on low-value customers will push the algorithm toward the wrong users at scale. Segment your seed lists — purchasers above AOV threshold, repeat buyers, high-LTV cohorts — and use those as separate Lookalike inputs.
• Add exclusions explicitly. Lookalike segments operating as suggestions means they're more porous. Layer in explicit exclusions (recent purchasers, existing email subscribers, retargeting audiences) to avoid budget waste on users who should be in a separate retargeting flow.
• Test bid adjustments against the new behavior. If your Demand Gen campaigns are running tROAS, the algorithm may be bidding more aggressively into expanded audiences. Consider temporarily lowering your tROAS target by 10–15% to give the system room to learn the new signal environment before tightening back up.

The April 30 API Deadline Worth Knowing

There's a secondary change that primarily affects accounts using the Google Ads API to manage Lookalike lists: starting April 30, 2026, Google will enforce a uniqueness check on Lookalike user lists. Any list that duplicates an existing one — same seed list, same expansion level, same country targeting — will be flagged and potentially paused.

For most DTC brands managing campaigns through the UI, this isn't a major concern. But if you're working with a media buying partner or automation layer that programmatically creates and manages audience lists, flag this to them now. Duplicate lists getting auto-paused mid-campaign is the kind of thing that doesn't surface until performance drops and you're hunting for the cause.

Is the New Model Better or Worse?

Honestly, it depends on your account maturity. For brands with clean conversion data, strong seed lists, and well-structured exclusions, giving Google's AI more flexibility to find converting users is probably a net positive. The algorithm has gotten genuinely good at identifying high-intent signals across Google's network — better than a rigid similarity threshold in many cases.

For brands earlier in their Demand Gen journey, with thinner conversion data or messier seed lists, the expanded reach introduces more noise before there's enough signal to steer it. In those cases, tighter ROAS targets and explicit exclusions become your substitute for the constraints that used to come from the reach slider.

Either way, the change has happened. The accounts that adapt quickly — auditing performance, cleaning up seed lists, layering in proper exclusions — will come out ahead. The ones that don't notice until their Demand Gen CPA drifts 25% in the wrong direction will have more ground to recover.