Google Ads Audience Targeting: Segments, Targeting vs Observation and What Works Where (2026)

Google Ads audience targeting reaches people based on who they are, what they are interested in, what they are actively researching, or how they have already interacted with your business. The setting that decides its effect is Targeting versus Observation: Targeting restricts who can see your ads, Observation changes nothing about reach and only reports and bids.

Most audience problems in an account trace back to that one setting being wrong, or to a segment type that simply is not available in the campaign type it was added to. This guide covers the segment types, the availability matrix by campaign, how Targeting and Observation differ, what Performance Max audience signals actually do, the minimum list sizes, and the policy limits that decide which audiences you are allowed to use at all.

The vocabulary changed, and the docs follow the new names

Google Ads renamed the building blocks, and the interface, reporting and help articles now use the new terms consistently:

• "Audience types" (custom, in-market, affinity) are now called audience segments.
• "Remarketing" is now called your data.

If you are reading an older playbook, translate as you go. A "remarketing list" is a your-data segment; an "audience type" is an audience segment. Reporting for all of it lives under Audiences in the Campaigns menu.

The segment types

Google Ads Help groups them as follows:

Two of these are worth flagging because they get misremembered. Lookalike segments are only available in Demand Gen campaigns according to Google Ads Help — they are not a general-purpose expansion tool across every campaign type. And custom segments are built from keywords, URLs and apps, which makes them the closest thing to intent targeting outside of Search.

Availability by campaign type

This is where most plans quietly break. Google Ads Help publishes the matrix, and Search is noticeably narrower than the others:

Read the two "No" columns for Search before you design a Search audience strategy: custom segments and life events are not available there. If your plan depends on them, the campaign type has to change, not the audience.

For the Display side specifically, audience targeting combines with content targeting — keywords, topics and placements — in ways that are worth understanding separately; our Google Display Network guide covers that interaction and the 2026 move of Display campaigns into Demand Gen.

Targeting vs Observation: the setting that decides everything

Google Ads Help draws the line cleanly.

Targeting. "When you're using the 'Targeting' setting for your campaigns or ad groups (like affinity audiences, placements or remarketing), you're telling Google Ads who you'd like to reach with your ads or where you'd like your ads to show on the internet." The consequence is stated just as plainly: "Targeting restricts the reach of your ad group."

Observation. "When you use the 'Observation' setting, the reach of your campaign or ad group isn't affected." It does not change who sees your ads. It lets you report on how a segment performs inside your existing targeting, and set bid adjustments on it.

Google's own recommendations differ by campaign type: Targeting is recommended for all advertisers in Display campaigns, while Observation is recommended for all Search campaigns as well as for the Display campaigns of more advanced advertisers.

There is one more line in that article that changes how you should think about Observation on Search: "For campaigns that use Smart Bidding, first-party audience segments added under the 'Observation' setting will be used as signals for Smart Bidding strategies." In other words, adding your customer and visitor segments in Observation mode on a Smart Bidding campaign is not a passive reporting choice — it feeds the bidding model without narrowing reach. If you are on automated bidding, read that alongside our Google Ads Smart Bidding guide.

Two operational notes from the same source: you cannot add an observation method when you first create an ad group, and observation data is only reported where it overlaps your existing targeting.

Performance Max: audience signals are not targeting

This is the most common misunderstanding in the entire area, and Google Ads Help addresses it directly: for Performance Max campaigns, audience signals indicating which users are most likely to convert can be added to help speed up Google AI, "However, this isn't a guarantee that ads will be served to only users within these audiences. If it's determined that other segments of users are converting well, ads will be served outside of users specified in the audience signals."

The practical reading:

• An audience signal is a starting hint, not a boundary. It accelerates learning; it does not fence the campaign in.
• If you need a hard boundary — a market you cannot serve, a customer set you must not re-approach — a signal will not deliver it. Use campaign-level exclusions and settings instead.
• Signal quality still matters, because a poor seed slows the learning phase. Your own converter and high-value customer lists are usually the strongest starting point.

Minimum list sizes

A segment that is too small simply will not serve, and the interface reports it as incompatible rather than as an error you caused. Google Ads Help states the current thresholds:

• Google Search Network: minimum 100 active visitors or users within the last 30 days.
• Google Display Network: minimum 100 active visitors or users within the last 30 days.
• YouTube: minimum 100 active visitors or users within the last 30 days.
• Customer lists: same eligibility — but the 100-user requirement "is valid only for lists uploaded or refreshed after February 1, 2024. Any list uploaded before February 1, 2024 must still meet the previous requirement of minimum of 1,000 matched users."

That last clause is the one that catches teams with older lists: a customer list sitting in the account since before February 2024 is still being judged against the old 1,000-match threshold until it is refreshed. Re-uploading it is usually the fastest fix.

One definitional caveat from the same article: "active users" counts people who engaged with your site or app and is different from total users reported in Google Analytics or Firebase. Do not size an audience from an Analytics number and assume it will serve.

Your data segments also carry a membership duration — how long a visitor stays in the segment after qualifying. Building the same intent with two durations (for example a short window for recent browsers and a longer one for past buyers) is usually more useful than one catch-all list. The same logic applies on other platforms; our Facebook Custom Audience guide covers the equivalent list-quality questions there.

Policy limits that decide which audiences you may use

Before designing anything elaborate, check whether your vertical is restricted. Google's personalized advertising policy splits audiences into two groups:

• Advertiser-curated audiences — Customer Match, your data segments, audience expansion, lookalike segments, custom segments. These cannot be used when promoting products or services in sensitive interest categories, because they may inadvertently contain sensitive user signals.
• Predefined Google audiences — in-market, affinity, demographics, detailed demographics, life events, location targeting. These can be used even in those categories, because Google states the sensitive signals are automatically excluded from them.

Two additional restrictions apply in the United States and Canada for housing, employment and consumer finance ads: certain demographics (age, gender, marital status, parental status) may not be used for targeting, and ZIP code location targeting is not available. Google also states that users under 18 are not eligible for personalized advertising of any kind.

If your account sits in one of these verticals, plan around predefined segments from the start rather than discovering the restriction at review time.

A practical setup order

1. Confirm campaign type first. Check the availability matrix before choosing a segment — a custom segment plan dies immediately on Search.
2. Check the policy group. If the vertical is restricted, design with predefined Google audiences only.
3. Build your data segments with deliberate durations, and verify each one clears 100 active users on the network you intend to serve on.
4. On Search, default to Observation. It costs nothing in reach, reports honestly, and feeds Smart Bidding when the segments are first-party.
5. On Display, use Targeting where you genuinely want to narrow, and layer Observation on the criteria you are still learning about.
6. In Performance Max, treat signals as seeds and put real constraints in settings and exclusions.
7. Filter the wrong traffic out separately. Audiences decide who; keywords decide what they typed. Our guide to Google Ads negative keywords covers the other half of that problem.

Common mistakes

• Setting audiences to Targeting on Search by default. It narrows a campaign that was already narrowed by keywords, and the volume loss is usually immediate.
• Expecting Performance Max signals to act as filters. They are hints; ads will serve outside them when other segments convert.
• Sizing a list from Analytics totals. The eligibility number is active users on the target network in the last 30 days, which is a different figure.
• Leaving pre-February-2024 customer lists unrefreshed. They remain subject to the older 1,000-matched-user threshold.
• Planning custom segments or life events into a Search campaign. Neither is available there.
• Ignoring the policy split. In restricted verticals, advertiser-curated audiences are off the table regardless of how well they would perform.

FAQ

What is the difference between Targeting and Observation in Google Ads?

Targeting restricts the reach of your campaign or ad group to the criteria you selected. Observation does not affect reach at all — it reports how selected audiences perform within your existing targeting and lets you set bid adjustments on them. Google Ads Help recommends Observation for all Search campaigns and Targeting for Display campaigns.

Which audience segments can I use in Search campaigns?

According to the availability matrix in Google Ads Help, Search campaigns support affinity segments, detailed demographics, in-market segments and your data segments. Custom segments and life events are not available in Search campaigns.

Do audience signals control who sees a Performance Max campaign?

No. Google Ads Help states that audience signals help speed up Google AI but are not a guarantee that ads will serve only to users in those audiences; if other segments convert well, ads will serve outside the signal. Use campaign settings and exclusions when you need a hard boundary.

How many users does an audience need before it can serve?

Google Ads Help states a minimum of 100 active visitors or users within the last 30 days for the Google Search Network, the Google Display Network and YouTube. Customer lists share that eligibility, except that lists uploaded before February 1, 2024 must still meet the previous minimum of 1,000 matched users until they are refreshed.