Most advertisers set up keywords, write ads, pick a bidding strategy, and then leave the audience tab completely untouched. That is a mistake. Keywords tell Google what someone is looking for. Audiences tell Google who that person is, what they have been researching for weeks, and whether they have already visited your site. Combine both and you get cleaner data, smarter bidding, and less wasted spend. The Ultimate Guide to Audiences in Google Ads is a useful companion to this.
This guide breaks down every audience type available in Google Ads, explains exactly when to use Observation instead of Targeting, and shows how to layer segments on Search and Display campaigns without shrinking your reach to a trickle. You will also get copy-and-adapt examples of custom segments built from competitor URLs and real search terms.
What Is Google Ads Audience Targeting?
Google Ads audience targeting is the practice of showing (or adjusting bids for) your ads based on who a user is rather than only what they typed into the search bar. Google builds these audience segments from signals such as browsing behaviour, app usage, YouTube activity, purchase intent signals, Google account data, and your own first-party data. The piece Google Ads Audience Targeting: How it Works makes a good next read.
In practice, audiences do three jobs:
- Refine: narrow a broad campaign to the people most likely to convert.
- Inform: feed performance data back to Smart Bidding so it bids up on high-value profiles.
- Reveal: expose which types of people actually buy from you, which is gold for creative and landing page work.
The Full List of Google Ads Audience Segments
Here is every audience category you can apply at the campaign or ad group level, with the realistic use case for each.
| Audience Type | What It Means | Best Used On | Funnel Stage |
|---|---|---|---|
| Affinity segments | Long-term interests and lifestyle (Cooking Enthusiasts, Business Professionals) | Display, YouTube, Demand Gen | Top |
| In-market segments | People actively researching or comparing products right now | All campaign types | Middle / Bottom |
| Life events | Moving, getting married, graduating, starting a business | Display, YouTube, Demand Gen, Gmail | Middle |
| Detailed demographics | Parental status, homeownership, education, employer size, marital status | All campaign types | Any |
| Custom segments | Audiences you build from keywords, URLs, and apps | Display, YouTube, Demand Gen, PMax signals | Middle / Bottom |
| Your data (remarketing) | Site visitors, app users, YouTube viewers, customer lists | All campaign types | Bottom |
| Customer Match | Uploaded emails, phones, addresses matched to Google accounts | Search, Shopping, Display, YouTube, Demand Gen | Bottom / Retention |
| Similar / lookalike expansion | Automated expansion from your seed lists inside Demand Gen and PMax | Demand Gen, PMax, Display | Top / Middle |
| Combined segments | AND / OR logic stacked across several of the above | Display, YouTube, Demand Gen | Any |
Affinity Segments: Who They Are
Affinity audiences describe stable, long-term interests. Think of them as the digital version of TV programme demographics. “Home Decor Enthusiasts” will still be home decor enthusiasts in six months.
When to use them: brand awareness, YouTube views, Display prospecting, and category education where nobody is searching for your product yet.
When to skip them: lead generation with an aggressive CPA target. Affinity traffic converts slowly and will drag your blended cost per acquisition if you judge it on last-click alone.
In-Market Segments: Who Is Buying Right Now
In-market segments are built from active purchase-intent signals: recent searches, clicks on comparison pages, time spent on category content. Someone in “In-Market for Business Software > CRM Software” is shortlisting vendors this month, not daydreaming.
Pro tip: add relevant in-market segments in Observation mode to your existing Search campaigns first. After three to four weeks you will see a clear bid-adjustment story, for example a 40% higher conversion rate for “In-Market for Office Furniture” on a generic keyword group that otherwise looks mediocre.
Life Events: Timing Beats Interest
Life events cover a short window: recently moved, about to move, newly engaged, starting a new job, retiring. Reach is small compared to affinity, but the timing relevance is enormous for movers, insurers, furniture retailers, banks, and local services.
Detailed Demographics: The Quiet Winner
Household income brackets, parental status, homeownership, education level, and company size are frequently the fastest win in an audit. B2B advertisers in particular should be layering employer size and industry data in Observation and then bidding up where the numbers justify it.
Your Data Segments: The Highest ROI Audience You Own
Website visitors, cart abandoners, past purchasers, video viewers, and uploaded customer lists consistently outperform every interest-based segment. If you have a Google tag with enhanced conversions and consent mode configured properly, build these lists first:
- All visitors (last 30, 90, and 540 days)
- Product or pricing page viewers who did not convert
- Cart or form abandoners (last 14 days)
- Converters, split by value tier
- Existing customers for exclusion or upsell
- YouTube channel viewers and engagers
Targeting vs Observation: The Setting That Decides Everything
This is the single most misunderstood control in the audience tab, and it appears in the People Also Ask box for a reason.
| Targeting mode | Observation mode | |
|---|---|---|
| Effect on reach | Restricts delivery to that audience only | No restriction at all |
| Bid adjustments | Allowed (manual bidding) | Allowed (manual bidding) |
| Reporting | Only for the selected audience | Full breakdown by segment |
| Typical use | Remarketing campaigns, RLSA on broad terms, tight prospecting on Display | Standard Search campaigns, data collection, Smart Bidding support |
| Risk | Severe volume loss if layered carelessly | Almost none |
The Simple Rule
- Search campaigns: use Observation by default. The keyword already qualifies intent. Adding a second qualifier in Targeting mode multiplies restrictions and starves the campaign.
- Display, Demand Gen and YouTube: use Targeting, because there is no keyword doing the qualifying work for you.
- Switch a Search ad group to Targeting only when you deliberately want a narrow experience: broad or generic keywords limited to past visitors, a dedicated “win back” ad group, or an aggressive bid on cart abandoners with different ad copy.
Important note on Smart Bidding: with tROAS, tCPA or Maximise Conversions, your manual audience bid adjustments are ignored (except negative adjustments on remarketing lists in some cases). Observation audiences still matter, because the segment membership becomes a signal the bidding model can use, and because the reporting tells you where to build dedicated campaigns.
How to Add Audiences to a Campaign or Ad Group
- Open the campaign or ad group in Google Ads.
- Go to Audiences, keywords and content, then Audiences.
- Click Edit audience segments.
- Choose whether you are editing at campaign or ad group level (ad group level gives finer bid control, campaign level is easier to maintain).
- Select Targeting or Observation.
- Browse, search, or use the “Ideas” tab to pick segments. Add them, then save.
- Wait for statistical significance before touching bids. For most accounts that means at least 30 conversions per segment, not three days of data.
How to Layer Audience Segments Without Shrinking Reach
Google applies AND logic between different targeting categories and OR logic within the same category. That single sentence explains almost every “my campaign stopped serving” panic.
- In-market A OR In-market B (same category) = wider reach.
- In-market A AND Age 25-34 AND Household income top 10% (three categories) = tiny reach.
- Add a topic or placement restriction on top of that and you may serve almost nothing.
A Safe Layering Framework
- Layer one, always on: add all plausible audiences in Observation across Search campaigns. No reach cost, pure intelligence.
- Layer two, tier your remarketing: create separate ad groups or campaigns for high-intent lists rather than restricting existing ones. A cart abandoner campaign in Targeting mode does not remove reach from your generic campaign, it runs alongside it.
- Layer three, cap the stack: on Display and Demand Gen, never combine more than two restrictive layers. One audience plus one geographic or device qualifier is plenty.
- Layer four, use exclusions instead of narrowing: excluding existing customers, job applicants, or converters keeps reach intact while removing waste.
- Layer five, monitor the reach estimator: if the projected weekly impressions drop below what your budget can spend, unstack one layer.
Layering on Search: A Worked Example
Take a B2B accounting software advertiser bidding on “accounting software” (broad, expensive, competitive).
- Ad group 1 (Observation): all keywords open to everyone. Audiences added for observation: In-Market for Accounting & Financial Software, In-Market for Business Services, Company size 1-249 employees, custom segment of competitor researchers.
- Ad group 2 (Targeting): same keywords but restricted to the “Visited pricing page, no trial” list, with ad copy that pushes a demo booking and a stronger offer.
- Exclusion at campaign level: existing customers list.
Total reach is unchanged. You simply routed the highest-intent slice of traffic to a better message.
Custom Segments: Where the Real Advantage Lives
Custom segments let you define an audience using your own inputs instead of Google’s pre-packaged categories. You can build them from:
- Interests and search terms: people who searched any of these terms on Google
- URLs: people with browsing behaviour similar to visitors of these websites
- Apps: people who use apps similar to these
- Places: people who visited certain types of locations
Building a Custom Segment From Competitor URLs
You are not remarketing to competitor visitors (that is not possible). Google reads the content of the URLs you submit and finds users whose browsing profile resembles the audience of those sites.
Recipe:
- List 10 to 15 direct competitor domains, plus the comparison and review pages that rank for “best [your category]”.
- Include category-defining publications, not just competitors, for example industry review directories and trade magazines.
- Avoid giants like amazon.com or wikipedia.org. They dilute the signal into meaninglessness.
- Use deep URLs where possible (competitor.com/pricing beats competitor.com).
- Name it clearly: CS | Competitor Browsers | CRM | 2026.
Example for a project management SaaS:
- asana.com/pricing
- monday.com/pricing
- clickup.com/compare
- wrike.com/product
- g2.com/categories/project-management
- capterra.com/project-management-software
Building a Custom Segment From Search Terms
This is the version most people underuse. Pull your actual converting search terms out of the Search terms report and paste them in. You are telling Google: find me people on Display, YouTube and Demand Gen who behave like the people who search these exact things.
Recipe:
- Export the Search terms report for the last 90 days.
- Filter to terms with at least one conversion.
- Strip out branded terms into their own segment (they behave completely differently).
- Group the rest into 3 to 5 thematic segments, 15 to 30 terms each. Do not dump 300 terms into one segment.
- Keep phrasing natural and commercial, as users would type it.
Example segments for a dental clinic:
| Segment name | Sample search terms | Where to use |
|---|---|---|
| CS | Implants Research | dental implant cost, full mouth implants price, implant vs bridge | Demand Gen, YouTube, Display |
| CS | Cosmetic Intent | teeth whitening near me, veneers before after, smile makeover | Demand Gen, PMax audience signal |
| CS | Urgent Care | emergency dentist open now, broken tooth what to do | Display remarketing, Demand Gen |
Custom Segments as Performance Max Audience Signals
In Performance Max, audiences are signals, not restrictions. Google uses them as a starting point and then expands. That makes custom segments built from your converting search terms one of the most valuable inputs you can give a PMax asset group, especially in the first four to six weeks while the campaign is learning.
Rank your signals in this order:
- Customer Match lists of high-value buyers
- Converters and cart abandoners
- Custom segments from converting search terms
- Custom segments from competitor and category URLs
- In-market segments
- Affinity and detailed demographics
Audience Strategy by Campaign Type
| Campaign type | Default mode | Priority audiences |
|---|---|---|
| Search | Observation | Your data, in-market, detailed demographics, Customer Match |
| Shopping | Observation | Cart abandoners, past purchasers, in-market |
| Display | Targeting | Remarketing, custom segments, in-market |
| Demand Gen | Targeting | Custom segments, lookalikes from first-party lists |
| Video | Targeting | Affinity for reach, custom segments for consideration |
| Performance Max | Signal only | Customer Match, converters, custom segments |
Seven Mistakes That Quietly Waste Budget
- Using Targeting mode on Search by accident. Check every campaign you inherited. This alone can cut impressions by 80%.
- Stacking three or more restrictive layers. Remember the AND logic between categories.
- Judging affinity audiences on last-click conversions. Use a view-through or data-driven attribution window before cutting them.
- Building one giant custom segment. Thematic groups of 15 to 30 inputs outperform a single dump of 200.
- Never excluding converters. Especially on Display and Demand Gen, where you will pay to re-reach the same people for weeks.
- Letting remarketing lists expire. A list with fewer than 1,000 users will not serve on Search. Extend the membership duration or widen the rule.
- Applying bid adjustments on Smart Bidding campaigns and expecting them to work. They will not. Restructure instead.
A 30-Day Implementation Plan
Week 1: Foundation
- Verify the Google tag, consent mode and enhanced conversions are firing correctly.
- Build your core remarketing lists and upload a Customer Match file.
- Audit every existing campaign for accidental Targeting mode.
Week 2: Observation Everywhere
- Add 8 to 12 relevant in-market and detailed demographic segments in Observation to all Search campaigns.
- Add exclusions for existing customers and irrelevant profiles.
Week 3: Custom Segments
- Export the search terms report and build 3 to 5 thematic custom segments.
- Build one competitor and category URL segment.
- Feed both into Performance Max signals and Demand Gen targeting.
Week 4: Read and React
- Pull the audience report by conversion rate and cost per conversion.
- Split out the top two segments into dedicated ad groups or campaigns with tailored copy and landing pages.
- Apply negative adjustments or exclusions to segments with volume but no conversions.
FAQ
What is the difference between audience targeting and observation in Google Ads?
Targeting restricts your ads so they only show to people in the selected audience. Observation places no restriction at all: your ads keep serving to everyone, but you get performance data broken down by segment and, with manual bidding, the ability to apply bid adjustments. On Search campaigns, Observation is the safe default.
Does adding audiences reduce my reach?
Only in Targeting mode. Observation never reduces reach. If reach drops after you add audiences, you almost certainly have multiple restrictive layers ANDed together, or you switched the ad group to Targeting without realising it.
What are the four main types of target audiences in Google Ads?
The four groups advertisers rely on most are: who they are (detailed demographics), what their interests are (affinity), what they are actively researching (in-market and life events), and how they have interacted with you (your data, remarketing and Customer Match). Custom segments sit across these categories as a build-your-own option. A comparable breakdown sits on wordstream.com.
How many users does a remarketing list need to serve?
Roughly 1,000 active users for Search and Shopping, and around 100 for Display. Customer Match generally requires a larger uploaded file to reach a usable match rate. If your lists are too small, extend membership duration up to the maximum allowed and broaden the qualifying rule.
Can I target competitor website visitors with Google Ads?
No, you cannot remarket to people who visited a competitor’s site. What you can do is create a custom segment using competitor URLs, which tells Google to find users whose browsing behaviour resembles that site’s audience. It is a proxy, and a very effective one for Display, YouTube and Demand Gen.
Do audience bid adjustments still work with Smart Bidding?
Not in the way you might expect. With tCPA, tROAS or Maximise Conversions, Google’s algorithm sets bids per auction and overrides your manual audience adjustments. Your audiences still matter as inputs and reporting dimensions, but if you want genuinely different bidding behaviour for a segment, isolate it in its own campaign or ad group.
Is a small daily budget enough to use audience targeting?
Yes, and small budgets often benefit the most. With limited spend, Observation audiences tell you quickly which profile converts, so you can concentrate budget instead of spreading it. Just avoid heavy Targeting layers on a small budget, since low volume plus narrow reach means you will never gather enough data to make a decision.
Key Takeaways
- Use Observation on Search, Targeting on Display, Demand Gen and Video.
- Different audience categories are combined with AND, which is why over-layering destroys reach. Stack no more than two restrictive layers.
- Your own data beats every Google-built segment. Build remarketing lists and Customer Match first.
- Custom segments from converting search terms and competitor URLs are the fastest way to give Google a better starting point, especially in Performance Max and Demand Gen.
- Do not narrow campaigns to improve quality. Split them, exclude waste, and write better ads for each segment instead.
Set your audience layer up properly once and it keeps paying you back, because every conversion you record teaches the bidding model something about the people worth paying for.