Audience Targeting

This page is updated every two months with current best practices for Google Ads audience targeting. As Smart Bidding and Performance Max take over delivery, the audience signals you feed Google increasingly shape who sees your ads and how efficiently your budget converts. Each update draws on our own experience plus authoritative industry sources and verified real-time research. Bookmark this page and check back for the latest audience targeting best practices. Each update includes worked examples with the arithmetic shown.

Last updated: 21 August 2026

In This Guide

  1. Executive Summary
  2. Benchmarks & Numbers at a Glance
  3. Audience Types Available
  4. In-Market, Affinity & Custom
  5. Detailed Demographics
  6. First-Party Data & Customer Match
  7. Remarketing & Similar Segments
  8. Targeting vs Observation
  9. Audiences Across Campaign Types
  10. Privacy & Consent
  11. Common Mistakes to Avoid
  12. What Changed Recently
  13. References

Reading this to build your own campaign? Skip the learning curve — we’ll build the whole thing for you for $129, delivered within 24 hours as a spreadsheet you own. See how it works →

1. Executive Summary

Google Ads audience targeting in 2026 is built on five principles that distinguish high-performing accounts from those relying on outdated segment-first thinking.

  • First-party data leads every strategy. Customer Match lists, CRM-derived segments, and website remarketing pools are more durable and more accurate than any Google-provided audience segment. Build your targeting stack from the inside out: your data first, Google segments second.[1][2][13]
  • Observation before Targeting, always. Adding audiences in Observation mode costs nothing in reach and generates the performance data you need to make defensible decisions. Move a segment to Targeting only when the evidence clearly supports restricting delivery to that group.[1][2][34]
  • Intent signals outrank interest signals for conversion campaigns. In-market and custom segments reflect recent behavioural intent. Affinity segments reflect lifestyle interests. For campaigns measured on cost per acquisition, prioritise intent signals and use affinity for awareness and reach extension only.[2][6]
  • Audience quality depends on measurement quality. Smart Bidding uses audience signals as inputs. If conversion tracking is broken, under-counting, or measuring the wrong event, audience optimisation will train on bad data regardless of how well the segments are constructed.[1][4][10]
  • Privacy compliance is a prerequisite, not an afterthought. Customer Match and remarketing lists are only activatable where you hold a compliant consent record. Segments built on improperly consented data must be excluded before they cause regulatory and platform-level consequences.[12][13][19]

2. Benchmarks and Numbers at a Glance

Metric Typical range or threshold Applies when Source
Minimum active users for a Search audience list to be eligible for targeting 100 active users in the last 30 days Search campaigns; vendor claim, not an independently published study [71]
Minimum active users for a Display or YouTube audience list 100 active users in the last 30 days Display and YouTube campaigns; vendor claim [71]
Customer Match minimum eligible members for list activation 100 members added or updated within the last 540 days Customer Match lists across all eligible campaign types; vendor claim [71]
Customer Match maximum membership duration 540 days All Customer Match list types; vendor claim [71]
Remarketing list maximum membership duration 540 days Website visitor and customer lists; vendor claim [64]
Privacy threshold below which list size is hidden in the API Fewer than 100 members — size displays as 0 Customer Match lists accessed via Google Ads API; vendor claim [67]
Maximum combined bid adjustment (upward) +900% Manual CPC and eligible Smart Bidding campaigns with audience bid adjustments; vendor claim [65]
Maximum combined bid adjustment (downward) −90% Manual CPC and eligible campaigns with audience bid adjustments; vendor claim [65]
Similar segments — targeting support Deprecated from 1 May 2023; no longer available for targeting or reporting All campaign types; confirmed Google platform change [11]
Recommended Customer Match list refresh frequency Weekly minimum; real-time API sync preferred for high-velocity businesses Any account using Customer Match with Smart Bidding; vendor best-practice guidance [13]
Observation mode — impact on reach Zero reduction in eligible impressions All Search, Display, and YouTube campaigns using Observation; vendor claim [34]
Minimum targetable Customer Match list — practical threshold cited in third-party guides 1,000 matched users Customer Match in Targeting mode; note: official Google threshold is 100 members — 1,000 is a conservative third-party recommendation for reliable delivery [10]

3. The Audience Types Available

Google Ads offers six principal audience categories as of August 2026. Understanding what each category measures — and what it does not — prevents the most common targeting errors.

Affinity Segments

Affinity segments group users by long-term interests and lifestyle habits inferred from sustained browsing patterns, app usage, and content consumption. They are broad by design and best suited to awareness campaigns where reach matters more than purchase proximity.[2][6]

In-Market Segments

In-market segments identify users who are actively researching or comparing products or services in a defined category. Google refreshes membership continuously based on recent search and browsing behaviour, making these segments useful for consideration and conversion campaigns.[2][6]

Custom Segments

Custom segments are built by the advertiser using keywords, URLs, and app names. They capture intent themes that Google’s predefined segments do not cover, including competitor research, niche product queries, and industry-specific vocabulary.[2][6]

Detailed Demographics

Detailed demographics extend standard age and gender targeting to include parental status, household income, education level, homeowner status, and life events such as moving house or starting a business. These are most valuable as exclusion layers or qualification filters rather than primary targeting dimensions.[6][15]

Your Data Segments (formerly Remarketing)

Google now uses the term “your data segments” to describe audiences built from first-party signals: website visitors tagged via the Google Ads tag, Customer Match lists uploaded from CRM data, and app user lists. These are the highest-value segments in any account because they are based on real interactions with your business.[2][3][5]

Combined Segments

Combined segments allow you to stack two or more audience conditions using AND/OR logic. A combined segment might require a user to be both in-market for home loans AND in a high-income household bracket. This is useful when you need tighter qualification but must be balanced against the risk of over-constraining delivery.[15][8]

Worked example

Choosing the right segment type for a financial services campaign

  • Setup: A Melbourne-based mortgage broker account spending $12,000 per month wants to introduce audience targeting across three Search campaigns: a brand campaign, a first-home-buyer campaign, and a refinancing campaign.
  • Numbers: The broker has a CRM list of 3,400 past enquirers. At a conservative 40% match rate, that yields approximately 1,360 matched users — above the 1,000-user practical threshold cited for reliable Customer Match delivery.[10] The refinancing campaign targets users who have previously visited the rate-comparison page; that tag-based list currently holds 620 active users in the last 30 days, which is above the 100-user platform minimum.[71] The first-home-buyer campaign has no existing audience data.
  • Decision: Brand campaign — add Customer Match (past enquirers) in Targeting mode at +30% bid adjustment. Refinancing campaign — add the rate-comparison remarketing list in Targeting mode; exclude the Customer Match purchaser list. First-home-buyer campaign — add Google’s “In-Market: First Home Buyers” segment in Observation mode with no bid adjustment until 30 days of data are collected.
  • Why: The 1,000-matched-user threshold[10] is met for Customer Match Targeting; the remarketing list exceeds the 100-user platform minimum[71]; Observation is required for the in-market segment because there is no performance evidence to justify restricting reach on a new campaign.[34]

4. In-Market, Affinity and Custom Segments

These three segment types are the primary tools for prospecting to users who have not yet interacted with your business. Selecting the right one for each campaign objective is the most consequential audience decision in a prospecting account.[2][6]

In-Market Segments

In-market segments are the default starting point for conversion-focused campaigns targeting new users. Google refreshes category membership frequently, so the audience reflects current research behaviour rather than historical interest. The practical limitation is that predefined categories may not match your specific product; in that case, custom segments are more precise.[2][6]

Best-practice use: add in-market segments in Observation on all active Search and Display campaigns immediately. After 30 days, review the audience report for conversion rate index and CPA. If a segment shows a cost per acquisition 20% or more below the campaign average, consider promoting it to Targeting or applying a positive bid adjustment.[1][2]

Affinity Segments

Affinity segments are better matched to Display, YouTube, and Demand Gen campaigns where building brand familiarity is the goal. Using affinity segments as a primary targeting layer on a conversion Search campaign is a common error because the segments capture interest, not intent. The correct role for affinity in a conversion account is as a reach extension on upper-funnel Display or video, or as a negative exclusion layer where clearly irrelevant interest groups can be filtered out.[1][2][6]

Custom Segments

Custom segments built on keywords are the closest Google Ads equivalent to keyword targeting for Display and Demand Gen campaigns. Build separate custom segments for: (a) your own product category keywords, (b) competitor brand terms, and (c) problem-aware queries that precede category searches. Test each custom segment independently in Observation before combining them, so you can isolate which intent theme performs.[2][6]

Worked example

In-market vs custom segment test on a Display campaign

  • Setup: A Brisbane e-commerce account selling ergonomic office furniture spends $4,500 per month on Display. The account manager wants to compare Google’s predefined “In-Market: Office Furniture” segment against a custom segment built on 35 keywords including “standing desk Australia,” “ergonomic chair back pain,” and three competitor brand URLs.
  • Numbers: After 30 days of Observation on both segments, the in-market segment accumulates 1,840 impressions and 6 conversions (conversion rate: 0.33%). The custom segment accumulates 940 impressions and 7 conversions (conversion rate: 0.74%). Campaign CPA target is $95 AUD. In-market CPA: $4,500 × (1,840/total impressions proportion) ÷ 6 — calculated from spend allocated: approximately $112 AUD. Custom segment CPA: approximately $58 AUD based on same proportional spend.
  • Decision: Promote the custom segment to Targeting mode. Keep the in-market segment in Observation for a further 30 days to accumulate more data before deciding whether to apply a −20% bid adjustment or exclude it.
  • Why: The custom segment’s 0.74% conversion rate versus 0.33% for in-market and its $58 AUD CPA versus the $95 AUD target provide sufficient evidence to restrict delivery; the in-market segment has not yet reached a volume that justifies exclusion.[2][6]

Worked example

Using affinity segments correctly for a brand awareness YouTube campaign

  • Setup: A Sydney-based meal kit subscription account allocates $8,000 per month to a YouTube non-skippable awareness campaign running from 1 September 2026 to 31 October 2026. The campaign goal is reach and brand recall, not direct sign-ups.
  • Numbers: The account manager adds three affinity segments in Targeting mode: “Cooking Enthusiasts,” “Health & Fitness Buffs,” and “Foodies.” Combined estimated reach across these three segments in the Sydney DMA is approximately 1.2 million unique users per month at the current CPM. Adding all three in Targeting reduces reach from 2.8 million (run-of-network) to 1.2 million but improves relevance alignment. The awareness campaign has no CPA target; the KPI is cost per thousand impressions (CPM) and view-through rate.
  • Decision: Use affinity segments in Targeting mode for this YouTube awareness campaign only. Do not apply affinity segments to the conversion-focused Search campaigns running in the same account.
  • Why: Affinity segments measure long-term lifestyle interest rather than purchase intent[2][6]; they are appropriate for reach-based awareness objectives but should not be the primary targeting layer on campaigns measured by CPA.[1]

5. Detailed Demographics

Detailed demographics in Google Ads cover parental status, household income decile, education level, homeowner status, marital status, and life events. They extend well beyond the standard age and gender options and allow advertisers to qualify or exclude users based on life-stage signals.[6][15]

Using Detailed Demographics as Exclusions

The highest-value use of detailed demographics is exclusion. If your product is unsuitable for a specific demographic — for example, a premium investment product that is not appropriate for users in the lowest household income brackets — applying a demographic exclusion reduces wasted impressions without requiring a separate campaign. This is preferable to narrowing all targeting dimensions simultaneously, which can over-constrain delivery and limit Smart Bidding’s ability to find converting users.[6][15]

Using Detailed Demographics as a Qualification Layer

When combined with in-market or custom segments, detailed demographics can tighten relevance for products with a defined buyer profile. The recommended approach is to add the demographic layer to an existing campaign in Observation first, confirm that the targeted demographic sub-group converts at a materially better rate than the excluded group, and only then apply the demographic dimension as a hard Targeting filter or negative bid adjustment.[6][15]

What to Avoid

Avoid using detailed demographics as your only or primary targeting layer. Google’s predefined demographic categories are broad inferences, not verified attributes. A user categorised as “likely homeowner” may not own property; a user in “upper 10% household income” may be misclassified. Demographic-only campaigns typically have lower match rates and higher CPAs than campaigns that lead with intent signals.[6][15]

Worked example

Household income exclusion on a premium home renovation campaign

  • Setup: A Perth luxury home renovation account spending $7,200 per month on Search and Display wants to reduce spend on users unlikely to afford a project value above $50,000 AUD. The campaign currently shows to all income brackets.
  • Numbers: After adding household income as an Observation layer for 60 days, the audience report shows: “Lower 50% household income” segment accounts for 28% of clicks but only 6% of conversions, producing a CPA of $620 AUD against the account average of $185 AUD. “Upper 30% household income” accounts for 41% of clicks and 72% of conversions, producing a CPA of $97 AUD.
  • Decision: Apply a −70% bid adjustment to “Lower 50% household income” on Display (do not exclude entirely, to preserve some auction participation). Apply a +25% bid adjustment to “Upper 10% household income” on both Search and Display. Do not move to Targeting mode on Search — retain Observation to avoid restricting keyword-matched reach.
  • Why: The −70% bid adjustment keeps the account competitive for lower-income users who may still convert while dramatically reducing cost exposure; the +25% upward adjustment is well within the −90% to +900% permitted range[65] and rewards the segment with the strongest demonstrated CPA.[6][15]

6. First-Party Data and Customer Match

First-party data is the most durable targeting asset in Google Ads because it is not subject to third-party cookie deprecation, is specific to your customer relationships, and provides the clearest signal of user value and lifecycle stage.[1][2][13]

Building a Segmented Customer Match Strategy

The single most common first-party data mistake is uploading one “all customers” list and treating it as a Customer Match strategy. Effective Customer Match requires separate lists for distinct lifecycle groups so that bids, messages, and exclusions can be calibrated to each group’s value and intent.[1][2][7][8]

Recommended minimum segmentation: high-LTV purchasers (top 20% by revenue), recent purchasers (last 90 days), lapsed customers (no purchase in 365–540 days), qualified leads (lead form submitted, not yet converted), and trial users (active trial, not yet paid).[1][2][7][8]

Data Quality and Match Rates

Upload the fullest set of identifiers available — email address, phone number, and postal address — because Google’s matching system combines multiple identifiers to increase list size. Remove duplicates, validate email formats, and normalise phone numbers to E.164 format before upload. Stale or malformed data reduces match rates and degrades the signal quality available to Smart Bidding.[13][9][10]

Lists with fewer than 100 matched members will show a size of 0 in the API and are not targetable.[67] A conservative practical minimum for reliable delivery in Targeting mode is 1,000 matched users.[10] Members are counted only if they were added or updated within the last 540 days.[71]

Refresh Cadence

Google explicitly states that stale Customer Match lists reduce optimisation quality.[13] For accounts with daily or weekly transaction volumes, automated API-based list syncing is the preferred approach. For smaller accounts, a weekly manual upload is the minimum acceptable cadence. At minimum, lists should be refreshed whenever a significant new cohort of purchasers, lapsed users, or leads has been added to the CRM.[13][10]

Using Customer Match with Smart Bidding

Customer Match lists function as audience signals for Target CPA, Target ROAS, and Maximise Conversions bidding strategies. Google’s guidance is explicit that the system uses Customer Match signals to improve automated bid decisions.[13] This means Customer Match improves performance even when lists are small, because the signal informs bidding rather than restricting reach. Upload lists before launching Smart Bidding campaigns rather than adding them post-launch.[13]

Consent Requirements

Customer Match data must only be activated where you hold a consent record that covers advertising use of that individual’s data. For Australian advertisers collecting data from users in the European Economic Area, Google’s Ads Data Hub guidance requires acknowledged consent for first-party data activation.[19][4] Ensure your CRM process captures and stores consent status, and that your list exports exclude contacts who have not consented to advertising use.[12][16][19]

Worked example

Segmenting a single CRM export into five actionable Customer Match lists

  • Setup: An Adelaide SaaS account spending $15,000 per month has a CRM with 9,200 contacts. Currently, all contacts are in one Customer Match list called “All CRM.” The account manager is rebuilding the list architecture ahead of a Q4 2026 acquisition campaign launching 1 October 2026.
  • Numbers: CRM breakdown: 1,840 contacts purchased in the last 90 days (high-recency buyers); 920 contacts are high-LTV (lifetime value above $4,800 AUD, top 10%); 2,100 contacts lapsed — no login or purchase in 365–540 days; 3,200 contacts are qualified leads (demo booked, not converted); 1,140 contacts are active trial users. At an estimated 42% match rate: high-recency list yields ~773 matched users (below 1,000 practical threshold — combine with high-LTV); high-LTV list yields ~386 matched users; combined high-recency + high-LTV list yields ~1,157 matched users (above 1,000 threshold[10]); lapsed list yields ~882 matched users (below threshold — expand window to 540 days or add email-only contacts to reach 100 platform minimum[71]); leads list yields ~1,344 matched users; trial list yields ~479 matched users (use as signal only, not Targeting).
  • Decision: Create four active lists: (1) High-value customers (recency + LTV combined, 1,157 matched) — Targeting mode on brand Search, +40% bid adjustment. (2) Lapsed customers (882 matched) — Observation mode until list reaches 1,000; use as exclusion on prospecting campaigns immediately. (3) Qualified leads (1,344 matched) — Targeting mode on competitor keyword campaigns. (4) Trial users (479 matched) — add as audience signal to Performance Max only, not as a Targeting list. Refresh all lists weekly via API from 1 October 2026.[13]
  • Why: Segmentation by lifecycle stage allows bids and messages to be calibrated to each group’s value; the 1,000-user practical threshold[10] determines which lists enter Targeting versus Observation; the 540-day membership cap[71] defines the maximum lookback for lapsed customers.

7. Remarketing and Similar Segments

Remarketing — now referred to by Google as “your data segments” in the platform interface — remains one of the highest-return audience strategies available, because users who have already interacted with your business convert at materially higher rates than cold prospects.[2][5]

Segmenting Remarketing Lists by Behaviour and Recency

A single “all website visitors” remarketing list is insufficient for accounts with meaningful traffic volumes. The recommended minimum segmentation is by funnel stage and recency: product page viewers (7-day window), add-to-cart abandoners (3-day window), checkout abandoners (1-day window), and past purchasers (90-day window for cross-sell, 540-day maximum for re-engagement).[2][5][8]

Shorter recency windows capture higher intent but produce smaller list sizes. The 100-user platform minimum[71] applies to each list independently, so very short windows (24–48 hours) may not reach threshold for smaller accounts. In those cases, extend the window to 7 days or combine funnel stages until the minimum is met.[71]

Exclusions as a Revenue Protection Tool

Excluding recent purchasers from prospecting and competitor campaigns is not optional — it is standard practice. Serving a prospecting ad to a user who purchased 48 hours ago wastes spend and damages brand perception. Apply purchaser exclusions across all non-remarketing campaigns as a baseline setting.[5][8][16]

Similar Segments — Deprecated

Google discontinued support for similar segments (lookalike audiences) for targeting and reporting from 1 May 2023.[11] As of August 2026, similar segments are no longer available in any campaign type. Accounts that previously relied on similar segments for prospecting scale should replace that function with custom segments built on high-intent keywords and competitor URLs, combined with optimised targeting in Display and Demand Gen campaigns.[11][2]

Optimised Targeting as the Lookalike Replacement

For Display and Demand Gen campaigns, Google’s optimised targeting feature uses your audience inputs — including your data segments and Customer Match — as seed signals to find additional converting users beyond the defined audience. This is the closest current equivalent to the deprecated similar segments function.[2][6] Provide high-quality seed audiences (high-LTV purchasers, not all visitors) to improve the signal quality optimised targeting receives.[1][2][7]

Worked example

Building a tiered remarketing list structure for a mid-size e-commerce account

  • Setup: A Canberra outdoor gear retailer spending $22,000 per month on Google Ads receives approximately 18,000 website sessions per month. The account currently uses one remarketing list: “All visitors – 30 days,” containing 9,400 active users.
  • Numbers: Session breakdown by funnel stage (last 30 days): product page views only — 11,200 users; add-to-cart events — 3,600 users; checkout initiated — 1,800 users; purchase confirmed — 1,100 users. All four groups exceed the 100-user platform minimum.[71] Proposed windows: product viewers — 14-day list; cart abandoners — 7-day list; checkout abandoners — 3-day list; past purchasers — 90-day list (for cross-sell), 540-day list (for win-back). Estimated list sizes after window adjustment: product viewers (14-day) ~5,600; cart abandoners (7-day) ~2,520; checkout abandoners (3-day) ~540; purchasers (90-day) ~3,300; purchasers (540-day) ~6,600.
  • Decision: Create five separate remarketing lists with the windows above. Apply checkout abandoners (3-day) to Search in Targeting mode with a +50% bid adjustment. Apply cart abandoners (7-day) to Search in Targeting mode with a +30% bid adjustment. Apply product viewers (14-day) to Display in Targeting mode. Apply purchasers (90-day) to Shopping in Observation mode for cross-sell messaging. Exclude purchasers (90-day) from all prospecting campaigns.
  • Why: Recency and funnel depth determine bid aggressiveness; the +50% adjustment for checkout abandoners is within the permitted −90% to +900% range[65]; purchaser exclusion on prospecting eliminates spend on users who have already converted.[5][8]

8. Targeting vs Observation Mode

The choice between Targeting and Observation mode is the most operationally consequential audience setting in a Google Ads account. Selecting the wrong mode either wastes the learning period (Targeting too early) or delays delivery restriction indefinitely (never promoting from Observation).[1][2][34]

What Each Mode Does

Mode Effect on reach Effect on bidding When to use
Observation No restriction — all eligible users can see the ad regardless of audience membership Allows audience-level bid adjustments; Smart Bidding uses the segment as an input signal Default for all new audience additions; testing phase; when you need data before restricting reach[1][34]
Targeting Hard restriction — only users in the defined audience are eligible to see the ad Bid adjustments apply only within the targeted pool; Smart Bidding operates within the constrained audience Remarketing-only Search campaigns; Customer Match re-engagement campaigns; after Observation data confirms the segment justifies restriction[2][5][11]

The Observation-First Workflow

Add every new audience segment in Observation mode. Run for a minimum of 30 days or until the segment accumulates at least 100 conversions at the campaign level, whichever comes later. Review the audience performance report: segments with a conversion rate index above 1.20 (20% above campaign average) or a CPA more than 20% below the campaign target are candidates for promotion to Targeting or a positive bid adjustment. Segments with a conversion rate index below 0.70 are candidates for a negative bid adjustment or exclusion.[1][2][34]

Bid Adjustments in Observation Mode

When a campaign uses manual CPC or a Smart Bidding strategy that permits audience bid adjustments, you can apply a percentage adjustment to each observed segment. The permitted range is −90% to +900%.[65] In practice, bid adjustments of −50% to +50% are the most common range for audience layers in Search. Adjustments outside this range should be treated with caution because they can cause delivery concentration or near-elimination of a segment before there is sufficient evidence to justify it.[65]

When Not to Use Bid Adjustments

If a Search campaign is governed by Target CPA or Target ROAS Smart Bidding, manual audience bid adjustments work against the algorithm. Smart Bidding already incorporates audience signals into its per-auction bid calculation. Adding a manual +30% adjustment on top of a Target CPA strategy creates double-counting and can push bids above the efficient frontier. In fully automated campaigns, add audiences in Observation as signals and let Smart Bidding determine the effective bid premium.[4][8]

Worked example

Promoting an in-market segment from Observation to Targeting on a Search campaign

  • Setup: A Gold Coast travel agency account spending $9,500 per month on Search has been running “In-Market: Luxury Travel” in Observation mode since 1 July 2026. The campaign uses Target CPA bidding at $140 AUD. The review date is 1 August 2026 (31 days of data).
  • Numbers: Over 31 days, the in-market segment recorded: 4,200 impressions, 210 clicks, 38 conversions, CPA = $78 AUD. The remaining (non-segment) traffic recorded: 9,800 impressions, 490 clicks, 41 conversions, CPA = $184 AUD. Segment CPA ($78) is 44% below the campaign Target CPA ($140). Segment conversion rate: 38 ÷ 210 = 18.1%. Non-segment conversion rate: 41 ÷ 490 = 8.4%. Conversion rate index: 18.1% ÷ 8.4% = 2.15 — well above the 1.20 promotion threshold.
  • Decision: The campaign uses Target CPA Smart Bidding — do not apply a manual bid adjustment. Instead, create a separate remarketing-only Search campaign with the “In-Market: Luxury Travel” segment in Targeting mode, set a separate Target CPA of $85 AUD for this campaign, and exclude the segment from the original broad campaign to prevent auction overlap.
  • Why: Manual bid adjustments conflict with Target CPA Smart Bidding[4][8]; the 2.15 conversion rate index (above the 1.20 threshold) justifies restricting reach to this segment in a dedicated campaign; separating the campaigns prevents the algorithm from cannibalising itself.[1][2]

9. Audiences Across Search, Performance Max and Demand Gen

Audience mechanics differ significantly across campaign types. Applying a Search-centric audience model to Performance Max or Demand Gen leads to structural errors that reduce campaign effectiveness.[6][7]

Search Campaigns

In Search, audience targeting is layered on top of keyword matching. The keyword remains the primary eligibility filter; the audience layer either refines bids (Observation) or adds a second eligibility condition (Targeting). The recommended default is Observation on all audience types except deliberate remarketing or Customer Match re-engagement campaigns, where Targeting is appropriate from the start.[1][2][34]

Performance Max Campaigns

Performance Max does not use traditional Observation and Targeting mode controls in the same way as Search. Audiences are provided as signals — inputs that guide where Google’s automation looks for likely converters, not hard constraints on delivery eligibility. The quality of the audience signal matters more than the quantity of signals provided. Supply your highest-quality Customer Match lists (high-LTV purchasers, recent converters) as the primary signal. Supplement with in-market and custom segments relevant to your category.[6][7]

Performance Max will expand delivery beyond the provided signals when it identifies conversion opportunities. This is by design. If you observe irrelevant placements, the correct intervention is to add URL exclusions and brand safety settings rather than attempting to constrain audience signals further.[6][7]

Demand Gen Campaigns

Demand Gen campaigns support audience targeting in a mode closer to Display than Search. You can specify audiences as inclusion or exclusion targets. Optimised targeting is available and recommended for prospecting objectives, using your data segments and Customer Match as seed inputs. Affinity and in-market segments are appropriate for Demand Gen when brand awareness or consideration is the objective.[6][16]

Campaign type Audience role Recommended mode Key restriction
Search Secondary filter on keyword-matched traffic Observation (default); Targeting for remarketing campaigns Audience restricts, it does not replace, keyword matching
Performance Max Audience signal to guide automated delivery Signal input — no classic Observation/Targeting toggle Google expands beyond signals; use brand safety and URL exclusions to manage placements
Demand Gen Targeting input with optimised expansion available Targeting or optimised targeting for prospecting; exclusion for suppression Optimised targeting may expand beyond defined audience; monitor reach vs relevance trade-off
Display Primary eligibility and bid layer Observation for new segments; Targeting for remarketing and Customer Match Over-layering multiple Targeting dimensions can reduce delivery below minimum thresholds

Worked example

Structuring audience signals for a Performance Max campaign launching in September 2026

  • Setup: A Hobart homewares retailer launching a Performance Max campaign for a Spring 2026 sale (campaign live 1 September 2026 to 30 November 2026) with a monthly budget of $18,000 AUD. The account has three existing Customer Match lists: high-LTV customers (1,600 matched users), cart abandoners (2,100 matched users), and all past purchasers (4,400 matched users). The account manager wants to configure audience signals before launch.
  • Numbers: High-LTV list: 1,600 matched users — above 1,000 practical threshold[10], strong signal quality. Cart abandoners: 2,100 matched users — above threshold, high-intent signal. All purchasers: 4,400 matched users — broad but useful as a supplementary signal. Proposed custom segment: 28 keywords including “linen duvet cover Australia,” “Scandinavian homewares online,” and two competitor URLs. In-market segment: “Home Décor” (Google predefined).
  • Decision: Add audience signals in this priority order: (1) High-LTV Customer Match list as primary signal. (2) Cart abandoners Customer Match list as secondary signal. (3) Custom segment (28 keywords + 2 URLs) as intent signal. (4) In-Market: Home Décor as contextual signal. Do not add “all past purchasers” as a signal — the list is too broad and dilutes signal quality. Do not attempt to set Observation mode — Performance Max uses signals, not classic audience mode controls.[6][7]
  • Why: Performance Max audience signals guide automation toward users resembling the provided groups[6][7]; providing the highest-quality, most specific lists first (high-LTV, then cart abandoners) gives the algorithm the clearest intent signal; adding a broad “all purchasers” list reduces signal specificity without adding meaningful incremental information.[1][2]

10. Privacy, Consent and Audience Eligibility

Privacy regulation and platform policy changes have materially changed what is permissible in audience targeting. As of August 2026, the key constraints affecting Australian advertisers are consent requirements for first-party data use, Apple’s App Tracking Transparency (ATT) impact on iOS-sourced list membership, and Google’s ongoing enforcement of sensitive category targeting restrictions.[3][12][13][19]

Consent Requirements for Customer Match and Remarketing

Google’s policy requires that Customer Match data is only used where you have obtained consent from the individual for advertising purposes. Your privacy policy must disclose that you share customer data with advertising platforms. Your consent capture mechanism must be specific enough to cover this use. Contacts who have not provided advertising consent must be excluded from all Customer Match exports before upload.[12][13][19]

For Australian advertisers with European customers, Google’s Ads Data Hub guidance requires acknowledged consent for first-party data activation on EEA users. Implement Google’s Consent Mode v2 signals so that consent status is communicated to the Google Ads tag in real time, and audit your tag configuration to confirm consent signals are being passed correctly.[19][4]

Apple ATT and iOS Audience Gaps

Apple’s ATT framework limits the ability of the Google Ads tag to observe and cookie iOS users who have not granted tracking permission to the app or browser in question. This means your website remarketing lists and Customer Match exclusions may systematically under-represent iOS users. Google’s own documentation acknowledges that ATT can impact your data segments and Customer Match on iOS 14+ traffic, including exclusions.[3] The practical implication is that your remarketing lists are likely smaller than the true eligible audience, and purchaser exclusion lists may miss some iOS converters.

Sensitive Category Restrictions

Google prohibits audience targeting based on sensitive categories including health conditions, financial hardship, and political beliefs. Custom segments built on keywords that imply sensitive status — for example, debt-related queries or specific medical condition searches — may be restricted or disapproved. Review custom segment keyword lists against Google’s personalisation policies before activation, particularly for financial services, health, and legal advertisers.[17]

Minimum Size Thresholds and Privacy Aggregation

Lists with fewer than 100 matched members display a size of 0 in the API and are not usable for targeting.[67] This threshold also applies to audience reporting: segments with fewer than the privacy threshold of members will not show disaggregated performance data, protecting individual user privacy. Build lists large enough to clear the 100-member platform minimum[71] before attempting to activate them in Targeting mode.[71][67]

Worked example

Auditing a Customer Match workflow for consent compliance before a November 2026 campaign launch

  • Setup: A Melbourne health insurance comparison account plans to upload a Customer Match list of 5,800 CRM contacts for a campaign launching 1 November 2026. The compliance team has flagged that the CRM consent records are inconsistent — some contacts opted in only to transactional emails, not to marketing or advertising data sharing.
  • Numbers: CRM audit result: 5,800 total contacts. Of these, 3,200 have explicit consent records covering “marketing and advertising personalisation.” 1,400 have consent covering “email marketing only” — advertising data sharing is not covered. 1,200 have no documented consent record (legacy contacts pre-dating the consent capture process). Compliant export: 3,200 contacts. At a 40% match rate: approximately 1,280 matched users — above the 1,000 practical threshold[10] and the 100-member platform minimum.[71]
  • Decision: Upload only the 3,200 consented contacts to Customer Match. Exclude the 2,600 non-compliant contacts entirely. Flag the 1,200 legacy contacts for re-consent outreach via email before December 2026. Do not upload the full list of 5,800 on the grounds that a larger list improves performance — non-compliant data cannot be activated regardless of list size benefit.[12][13][19]
  • Why: Google’s Customer Match policy and Australian Privacy Act obligations require advertising data sharing consent before activation[12][19]; the 3,200-contact compliant list still yields 1,280 matched users, which clears both the platform minimum[71] and the practical delivery threshold.[10]

11. Common Mistakes to Avoid

The following errors recur across accounts of all sizes and represent the most common sources of wasted spend and misconfigured audience strategy.[1][2][6][11]

Starting in Targeting Mode Without Data

Applying a new audience segment in Targeting mode before collecting performance evidence restricts delivery based on assumption rather than observation. The correct workflow is always Observation first, Targeting only after the segment has demonstrated a conversion rate or CPA that justifies the delivery constraint.[1][2][34]

Using a Single “All Customers” Customer Match List

A combined list of all CRM contacts conflates high-LTV purchasers with lapsed customers, unqualified leads, and opted-out contacts. This destroys the value of segmentation and prevents lifecycle-appropriate bidding and messaging. Minimum segmentation: purchasers, high-LTV, lapsed, and leads as four separate lists.[1][2][7][8]

Neglecting Purchaser Exclusions on Prospecting Campaigns

Serving prospecting ads to existing customers wastes budget and creates a poor user experience. Apply a recent purchaser exclusion list (minimum 90 days) to every prospecting campaign as a default setting, not an optional extra.[5][8][16]

Applying Manual Bid Adjustments on Top of Smart Bidding

Smart Bidding already incorporates audience signals in per-auction bid calculations. Adding a manual +30% or +50% audience bid adjustment on a Target CPA or Target ROAS campaign instructs the algorithm to apply the adjustment on top of its own calculation, which can overpay for the segment and undermine campaign efficiency. In Smart Bidding campaigns, add audiences as Observation signals only and let the algorithm determine the appropriate bid premium.[4][8]

Over-Constraining Delivery with Too Many Targeting Layers

Stacking multiple Targeting dimensions — for example, in-market segment AND detailed demographic AND combined segment, all in Targeting mode simultaneously — can reduce the eligible audience below viable delivery levels. Combined segments and multiple Targeting layers are appropriate only when each individual component has demonstrated performance in Observation, and only when the combined pool remains large enough to support the campaign’s daily budget at a competitive CPM or CPC.[15][8]

Ignoring the 540-Day Membership Cap

Customer Match and remarketing lists have a maximum membership duration of 540 days.[71][64] Users who entered a list more than 540 days ago are automatically removed. Accounts that rely on lists seeded from historical CRM exports without regular updates will find their targetable audience steadily shrinking. Set calendar reminders for quarterly list audits and implement weekly automated refreshes.[13]

Treating Affinity Segments as a Conversion Driver in Search

Affinity segments measure long-term lifestyle interest, not near-term purchase intent. Applying affinity segments in Targeting mode on a conversion-focused Search campaign restricts the auction to users whose interest-profile matches — not users who are actively searching to buy. The result is reduced reach without a corresponding CPA improvement. Use affinity segments for awareness on Display, YouTube, and Demand Gen, not as a Targeting layer in Search.[1][2][6]

Worked example

Diagnosing an over-constrained campaign caused by stacked Targeting layers

  • Setup: A Sydney legal services account spending $6,000 per month on Search has seen impression share decline from 68% to 31% over six weeks from 1 July 2026 to 11 August 2026. The account manager investigates and finds three audiences in Targeting mode on the same campaign: (1) In-Market: Legal Services (Targeting), (2) Combined segment: In-Market: Legal Services AND Household Income Upper 30% (Targeting), (3) Customer Match: past enquirers (Targeting).
  • Numbers: Each Targeting layer independently reduces the eligible auction pool. Estimated eligible impressions with no audience restriction: 45,000 per month. After In-Market Targeting layer: ~18,000 (40% of total). After combined segment Targeting layer (stacked AND condition): ~7,200 (40% of 18,000). After Customer Match Targeting layer (further restriction): ~2,880 (40% of 7,200). With a $6,000 monthly budget and average CPC of $12 AUD, the campaign needs at least 500 clicks per month; at a 3% CTR on 2,880 eligible impressions, available clicks are approximately 86 — far below the 500 needed to spend the budget.
  • Decision: Move In-Market: Legal Services from Targeting to Observation mode. Move the combined segment from Targeting to Observation mode. Retain Customer Match in Targeting mode (the only list with strong prior CPA evidence). Set a +25% bid adjustment on the in-market segment in Observation. Re-evaluate after 30 days (11 September 2026).
  • Why: Stacking three Targeting mode audiences applied an AND eligibility condition that reduced the eligible pool to approximately 2,880 impressions — insufficient to spend $6,000 at $12 AUD CPC[15][8]; retaining only the Customer Match list in Targeting mode preserves the highest-value restriction while restoring auction reach.[2][34]

12. What Changed Recently (Last 30 Days)

The following section covers documented Google Ads audience changes and announcements as of August 2026. Where a change is confirmed in Google’s official help documentation, it is marked as confirmed. Where a change is reported only in third-party sources and is not corroborated in official documentation, it is flagged as unverified.[4][11][14]

Confirmed Changes

Budget and bid behaviour update — 17 August 2026. Google announced a change to campaign budget and bid behaviour taking effect on 17 August 2026. While primarily a bidding-system change rather than an audience-targeting change, it affects how Smart Bidding distributes spend across audience segments within a campaign. Accounts using Target CPA or Target ROAS should review post-17 August performance data to confirm audience-level CPAs have not shifted materially as a result.[4]

Continued “your data” terminology enforcement. Google’s platform interface and help documentation consistently use “your data segments” in place of “remarketing” across all campaign types and reporting views. Advertisers referencing “remarketing lists” in internal documentation or third-party tools should update their nomenclature to align with current platform labelling.[2][3][5][12][16][17]

Consolidated Audiences reporting page. Google continues to direct all audience management — demographics, segments, exclusions, and bid adjustments — to the unified Audiences page within the Campaigns menu. This is the single correct location for reviewing audience performance, applying bid adjustments, and managing segment membership. There is no separate audiences interface; it is fully consolidated.[2][3][5][12][16][17]

Similar segments remain deprecated. Similar segments were deprecated for targeting and reporting from 1 May 2023 and remain unavailable as of August 2026. No re-introduction has been announced.[11] Accounts still referencing similar segments in campaign plans should replace this with optimised targeting seeded from high-LTV Customer Match lists.[11][2]

Unverified — Treat with Caution

100-user audience minimum across all campaign types. Several third-party sources published in mid-2026 claim Google has lowered the minimum audience list size to 100 active users for Search, Display, and YouTube.[6][8][10] This figure is consistent with the threshold published in Google’s official help documentation as of August 2026.[71] However, because the threshold change is reported as recent in third-party sources but is not accompanied by a dated Google announcement in the official help centre, advertisers should verify the current threshold in their own account UI before relying on it for campaign planning. The conservative position is to treat 1,000 matched users as the practical minimum for reliable Targeting mode delivery, with 100 as the platform minimum for list eligibility.[10][71]

Worked example

Auditing audience settings following the 17 August 2026 bid behaviour update

  • Setup: A Newcastle automotive dealership account spending $11,000 per month on Search uses Target CPA bidding at $220 AUD per lead across three campaigns. The account manager reviews performance on 19 August 2026, two days after the announced bid behaviour change took effect on 17 August 2026.
  • Numbers: Pre-update performance (1–16 August 2026, 16 days): total leads = 64, total spend = $5,867 AUD, blended CPA = $91.67 AUD per day average spend. Post-update (17–19 August 2026, 3 days): total leads = 8, total spend = $1,124 AUD, blended CPA = $140.50 AUD. Audience-level review: “In-Market: Cars & Trucks” (Observation, +20% bid adjustment) — pre-update CPA $78 AUD, post-update CPA $115 AUD over 3 days (small sample, statistically unreliable). Customer Match: past service customers (Observation, no bid adjustment) — pre-update CPA $65 AUD, post-update data insufficient (2 conversions).
  • Decision: Pause the +20% manual bid adjustment on the in-market segment immediately, because the 17 August 2026 bid behaviour change affects how Smart Bidding interacts with manual adjustments. Monitor for a full 14-day post-update window (until 31 August 2026) before drawing conclusions or making further audience changes. Document the pre/post CPAs for each audience segment to identify any persistent shift caused by the update.
  • Why: Three days of post-update data is insufficient for statistical significance — a 14-day window is the minimum reliable evaluation period for a Smart Bidding campaign following a system-level change[4]; pausing the manual bid adjustment eliminates a potential interaction effect between the update and the stacked adjustment, consistent with best practice for Smart Bidding campaigns.[4][8]

References

  1. [1] https://aleckirby.com/blog/google-ads-audience-targeting/ aleckirby.com
  2. [2] https://www.clickedon.co/insights/google-ads-audience-targeting www.clickedon.co
  3. [3] https://clicksgeek.com/better-targeting-for-google-ads/ clicksgeek.com
  4. [4] https://www.clairejarrett.com/top-5-google-ads-best-practices-for-2022/ www.clairejarrett.com
  5. [5] https://lmgroup.au/blog/mastering-google-ads-audience-targeting-2026 lmgroup.au
  6. [6] https://roa-marketing.com/blog/google-ads-audience-targeting-options-2026/ roa-marketing.com
  7. [7] https://videoadsstudio.com/google-ads-targeting-2026-marketer-s-edge/ videoadsstudio.com
  8. [8] https://videoadsstudio.com/google-ads-target-smarter-not-harder-in-2026/ videoadsstudio.com
  9. [9] https://www.mbadv.agency/google-ads/audience-targeting www.mbadv.agency
  10. [10] https://twosquares.co.uk/blog/google-ads-audiences twosquares.co.uk
  11. [11] https://ppchero.com/advanced-google-ads-techniques-to-master-in-2026/ ppchero.com
  12. [12] https://blog.adnabu.com/google-ads/google-ads-audience-targeting/ blog.adnabu.com
  13. [13] https://support.google.com/google-ads/thread/413560562/google-ads-best-practices-for-2026?… support.google.com
  14. [14] https://www.attnagency.com/blog/advanced-google-ads-audience-targeting-ai-optimization-202… www.attnagency.com
  15. [15] https://business.google.com/us/resources/articles/reach-larger-new-audiences/ business.google.com
  16. [16] https://almcorp.com/blog/google-demand-gen-best-practices-february-2026/ almcorp.com
  17. [17] https://support.google.com/google-ads/answer/2476691?hl=en support.google.com
  18. [18] https://spideraf.com/articles/10-tips-to-maximize-google-ads-targeting-and-segmentation spideraf.com
  19. [19] https://leadsbridge.com/blog/google-ads-best-practices/ leadsbridge.com
  20. [20] https://www.linkedin.com/posts/blasonde_google-ads-2026-complete-guide-for-advertisers-act… www.linkedin.com
  21. [21] https://www.overthetopseo.com/google-ads-strategy-2026-maximum-roi/ www.overthetopseo.com
  22. [22] https://socialadsstudio.com/audience-targeting-drive-2026-roi-with-1st-party-data/ socialadsstudio.com
  23. [23] https://developers.google.com/ads-data-hub/marketers/guides/audience-activation developers.google.com
  24. [24] https://www.thesarahstemen.com/blog/google-ads-algorithm-training www.thesarahstemen.com
  25. [25] https://roa-marketing.com/blog/first-party-data-google-ads-strategy-2026/ roa-marketing.com
  26. [26] https://www.unled.net/blog/google-ads-first-party-data-management-2026.html www.unled.net
  27. [27] https://www.groas.com/post/google-ads-first-party-data-customer-match-without-cookies-guid… www.groas.com
  28. [28] https://support.google.com/google-ads/answer/6154846?hl=en support.google.com
  29. [29] https://transcenddigital.com/blog/first-party-data-google-ads-setup-guide/ transcenddigital.com
  30. [30] https://business.google.com/us/accelerate/resources/articles/google-ads-ai-essentials-cust… business.google.com
  31. [31] https://twominutereports.com/blog/google-ads-best-practices twominutereports.com
  32. [32] https://easyinsights.ai/blog/strategies-to-use-first-party-audiences-on-meta-and-google-ad… easyinsights.ai
  33. [33] https://developers.google.com/ads-data-hub/guides/audience-activation developers.google.com
  34. [34] https://support.google.com/google-ads/answer/7365594?hl=en support.google.com
  35. [35] https://www.searchenginejournal.com/targeting-observation-ppc/430438/ www.searchenginejournal.com
  36. [36] https://developers.google.com/google-ads/api/docs/targeting/targeting-settings developers.google.com
  37. [37] https://support.google.com/google-ads/answer/7374253?hl=en support.google.com
  38. [38] https://lionelz.com/en/blog/audience-layering-google-ads/ lionelz.com
  39. [39] https://learn.jyll.ca/blog/should-you-use-targeting-or-observation-mode-in-google-ads learn.jyll.ca
  40. [40] https://growthmindedmarketing.com/blog/targeting-vs-observation/ growthmindedmarketing.com
  41. [41] https://aimers.io/blog/google-display-ads-targeting-options aimers.io
  42. [42] https://hawksem.com/blog/audience-targeting-vs-observational/ hawksem.com
  43. [43] https://www.reddit.com/r/googleads/comments/1iu7ri8/google_search_audiences_targeting_or_o… www.reddit.com
  44. [44] https://www.linkedin.com/pulse/google-ads-audience-targeting-strategies-work-dana-ditomaso… www.linkedin.com
  45. [45] https://support.google.com/google-ads/answer/2454137?hl=en support.google.com
  46. [46] https://support.google.com/google-ads/answer/1704368?hl=en support.google.com
  47. [47] https://support.google.com/google-ads/answer/2497941?hl=en support.google.com
  48. [48] https://support.google.com/google-ads/answer/19888?hl=en support.google.com
  49. [49] https://support.google.com/google-ads/answer/2472738?hl=en support.google.com
  50. [50] https://pinpoint-media.global/insights/google-ads-lowered-audience-size-requirements pinpoint-media.global
  51. [51] https://help.optmyzr.com/en/articles/4984177-change-history-widget help.optmyzr.com
  52. [52] https://www.webtonic.io/blog/google-ads-audience-targeting www.webtonic.io
  53. [53] https://www.seroundtable.com/google-ads-100-customer-lists-39472.html www.seroundtable.com
  54. [54] https://www.linkedin.com/posts/maulikpanchal_googleads-ppcupdate-marketingnews-activity-74… www.linkedin.com
  55. [55] https://support.google.com/sa360/answer/13388565?hl=en support.google.com
  56. [56] https://support.google.com/google-ads/answer/2701222?hl=en support.google.com
  57. [57] https://jyll.ca/insidegoogleads/72 jyll.ca
  58. [58] https://support.google.com/google-ads/answer/17061251?hl=en support.google.com
  59. [59] https://support.google.com/google-ads/answer/2580383?hl=en support.google.com
  60. [60] https://support.google.com/google-ads/answer/2497940?hl=en support.google.com
  61. [61] https://developers.google.com/google-ads/api/docs/query/date-ranges developers.google.com
  62. [62] https://steerads.com/cs/blog/audiences-affinite-in-market-custom steerads.com
  63. [63] https://lafactory.com/google-ads-audiences-in-market-affinity-custom-combined-segments/ lafactory.com
  64. [64] https://support.google.com/google-ads/answer/6276125?hl=en support.google.com
  65. [65] https://support.google.com/google-ads/answer/2732132?hl=en support.google.com
  66. [66] https://support.google.com/google-ads/answer/10566573?hl=en support.google.com
  67. [67] https://developers.google.com/google-ads/api/docs/remarketing/audience-segments/customer-m… developers.google.com
  68. [68] https://www.bigeyeagency.com/insights/google-ads-bid-adjustments-in-2026-what-still-works-… www.bigeyeagency.com
  69. [69] https://support.google.com/google-ads/answer/6299717?hl=en support.google.com
  70. [70] https://services.google.com/fh/files/misc/audience_product_experts_academy_lesson_four.pdf services.google.com
  71. [71] https://support.google.com/google-ads/answer/6379332?hl=en support.google.com
  72. [72] https://organicgrowthstudio.com/google-ads-boost-ctr-15-with-2026-segmentation/ organicgrowthstudio.com

Don’t want to do this yourself?

Get a complete, ready-to-launch Google Ads campaign — keywords validated against real search data, copy written to the best practices on this page, and a proper negative-keyword list — built for your business and delivered within 24 hours.

Build my campaign — $129

One-off price · Human-reviewed · No subscription · No access to your Google Ads account

Share the Post:

Related Posts