Your shopping cart is empty!
Performance Max is testing household income exclusions at campaign level
Publication date: 26.07.2026
Performance Max has gained a lever advertisers have wanted for a long time: household income segments can now be excluded at campaign level while Google's automated optimisation keeps running. PMax used to offer almost no audience control, so even a narrow filter is a noticeable step.
What exactly appeared
PPC specialist Thomas Eckel spotted the setting first in a European campaign. There is no official announcement from Google — it looks like a limited test rollout, covered by Search Engine Land. The control sits in campaign settings and removes the selected income groups from serving.
Seven brackets are available for exclusion: top 10% of incomes, the 11–20%, 21–30%, 31–40% and 41–50% ranges, the lower 50%, and "unknown income" as a separate group.
Does it work in Ukraine
Household income targeting is a tool Google offers only in some markets: the data exists where it can be modelled from statistics. Ukrainian campaigns usually do not have these segments at all, so the first step is to open audience settings in your own account and check whether the bracket list exists for the country you target.
One more nuance: even where the brackets exist, this is a modelled estimate, not an income declaration. Google infers the group from a mix of signals — location, device type, behaviour — which is why the "unknown income" share passes half of all traffic in some countries. That is not a misconfiguration, it is the accuracy limit of the data itself.
In practice the feature matters for those running campaigns into external markets — the EU, the US, the UK. For campaigns inside Ukraine the same job is done by other signals: geography down to a city district, device type, ad schedule and, above all, the assortment on the landing page itself.
Who actually needs the filter
| Business type | What to exclude | Why |
|---|---|---|
| Premium, luxury, financial services | lower brackets | stop paying for an audience that cannot buy |
| Discounters, value brands | top 10–20% | aim at your own audience instead of random clicks |
| Small business and agencies | selectively, after a test | get a little manual control back inside the black box |
How to test it safely
The main trap is the "unknown income" segment. It often holds a large share of traffic, and excluding it can cut reach sharply. The sequence is simple:
- record baseline campaign metrics for 28 days: impressions, CPA, conversions, branded traffic share;
- enable the exclusion on one campaign, not across the account;
- leave "unknown income" alone in the first test — work with the extreme brackets only;
- compare CPA together with conversion volume, not CPA on its own: the filter almost always improves one at the expense of the other;
- if reach drops by more than a third, put the bracket back and work on the feed and creatives instead.
If the numbers contradict each other after the test, the bracket is almost never the real problem: campaign structure, conversion signals or feed quality usually are. Those show up in a contextual advertising audit, where you look at the whole account instead of one setting.
Income exclusions are not a profitability button. They remove clearly irrelevant traffic where the cost of a mistake is high. Everything else in PMax still comes down to feed quality, assortment and landing pages.
