New Customer Acquisition Goal

A campaign told to maximise conversions spends your budget on the cheapest conversions it can find. Often those are the people who bought from you last month.

Existing buyers are easier and cheaper to convert, so an unconstrained bid strategy keeps reconverting them, a problem Google names in its Performance Max documentation. The New Customer Acquisition goal is the lever that leans the other way. It tells Smart Bidding to weight, or restrict itself to, buyers the account has not seen before.

”New” means “not in your last 540 days of conversions”

The setting is only as good as its definition of a new customer, and the default definition is self-referential. Google reads the account’s own conversion data from the trailing 540 days; anyone who converted inside that window is filed as existing, and everyone else is treated as new, per Google’s detection documentation.

That window is where the setting quietly misfits certain brands. A customer who bought 18 months ago sits outside the 540-day lookback, so the campaign is free to pay a first-time-buyer premium to reacquire someone who is not new at all. For a brand with a long replenishment cycle, the definition and the reality drift apart.

Uploading a Customer Match list overrides the guesswork. Google then marks anyone absent from the list as new, and it recommends running both a remarketing list and a first-party list for the most durable detection. The belt-and-suspenders advice is a tell: the automatic method alone is not something Google stands behind.

Two modes, two different volume costs

The goal ships in two settings that behave very differently. In value mode the campaign still bids on everyone, but a first purchase carries an extra assigned value — a new customer’s €200 order can be counted as €300 in the bidding signal, as Adchieve documents for Performance Max. The auction math shifts toward new buyers without shutting existing ones out.

Only mode is the hard version: the campaign bids exclusively for new customers. It typically returns lower total conversion counts and higher CPAs, and Google’s own troubleshooting page notes that campaigns using the goal may serve fewer ads than expected, in its guidance on the setting. The volume you give up is the price of the purity.

What Google’s uplift figures do and don’t cover

Google publishes aggregate results for both modes. Valuing new customers at twice average order value improved ROAS by 9% in value mode, and only mode raised the new-customer ratio 11.5% at a 3% lower acquisition cost, in Google’s own reporting on the modes.

These are platform-authored averages across many accounts, not a single advertiser’s result. They are the vendor’s own claim rather than an independent benchmark.

Named results are thinner on the ground. Cropp, the Gen-Z apparel brand under LPP Group, paired a Demand Gen campaign with the goal and reported a ROAS increase above 50% and a 4.5% lift in conversions among women shoppers, per a February 2026 write-up of Google’s Demand Gen best practices. Dime Beauty ran the goal inside Performance Max with a higher value set for new customers, though its published numbers stop at the configuration, in Adchieve’s account of the rollout.

Where the definition breaks

The accuracy ceiling is the whole story with this feature. Because detection leans on the account’s own tracking, cookie deletion and tracking opt-outs push real existing customers into the new bucket, and some conversions come back labelled unknown when Google cannot place them at all, as its detection documentation concedes. No method is fully accurate, and some existing-customer spend “getting through” is expected, in Google’s own words.

That is the gap a help page skimmed for its benefits does not assemble. A setting sold as a way to stop paying to reacquire buyers will, under ordinary privacy loss, still bid the new-customer premium on people it simply failed to recognise. The published uplift and the misclassification problem describe the same feature, both true at once, and the second caps how far the first travels in a real account.

RYZE excludes existing buyers on one ad out of 400

Before the goal automated any of this, brands did the same job by hand with customer-list exclusions. The teardown corpus shows how narrowly they reach for it.

Across RYZE’s 400 Google ads, exactly one uses customer-list exclusion — a single campaign set up specifically to avoid showing ads to people who have already bought, in the RYZE Google Ads breakdown. One more ad does the reverse and remarkets to existing customers.

The New Customer Acquisition goal moves that decision from a manual list toggle on a single ad to an account-level bidding weight. What it does not change is the harder part RYZE was already navigating: knowing, reliably, which customers are actually new.

Seen in this teardown

• RYZE Google Ads — 1 of 400 ads uses customer-list exclusion to keep prospecting ads off existing buyers; 1 remarkets to them