Google Ads API v25.2 adds percentile benchmarks and PMax asset group tracking

Google released version 25.2 of the Google Ads API on 23 September 2026. It is a minor release with no breaking changes, and it adds percentile rankings to the benchmarks tool, URL tracking options at the Performance Max asset group level, a way to turn a Smart campaign into a paused Performance Max draft, two new bid-too-low recommendations, specific Business Profile sync error reasons, and new hotel and travel reporting fields. The changes are listed in Google's Google Ads API release notes.

Most advertisers will never call the API themselves. They will meet these changes through the software built on top of it: bid management platforms, reporting connectors, agency dashboards, feed tools and bulk-editing utilities. So this post explains each change in terms of what it means inside an account, and says plainly when a feature only becomes useful once a tool you already use adopts it.

Where v25.2 sits in the release cycle

Google shipped the v25 major version on 22 July 2026, followed by v25.1 on 19 August 2026 and v25.2 on 23 September 2026. The release notes state that minor versions like v25.2 do not contain breaking changes. That matters for anyone paying for API-based software: nothing in this release forces a vendor to rewrite existing code, so nothing should stop working in your reports or automations because of it. New fields simply become available for tools to pick up.

Relevant Audience covered the v25 major release on 27 July 2026, when the headline was the removal of the old lifecycle-goal resources, which was a genuine breaking change. v25.2 is the opposite kind of release. It is additive and optional, and most of it is new data or new controls.

Percentile rankings arrive in the benchmarks API

Google updated BenchmarksService.GenerateBenchmarksMetrics in v25.2 so it can return a customer's competitive standing among other advertisers as a percentile tier. The results come back in a new CustomerMetrics.percentile_metrics field and are expressed using a BenchmarksCustomerPercentileTier value.

This connects to a change inside the Google Ads interface. On 17 September 2026 Relevant Audience wrote about the Spend benchmarks report in the Account Overview, which puts your spend and clicks next to a figure for businesses Google treats as similar. That report shows a peer number beside yours. What is new in v25.2 is that software working through the API can now ask for a percentile tier, a statement of where the account ranks among advertisers, rather than only raw aggregate and average figures.

There are conditions. According to the release notes, the percentile field is only filled in when the request asks for PERCENTILE_DATA in its supplemental data and sets BenchmarksSource.all_advertisers as the benchmarks source, which in turn requires a category filter. Google also added percentile support for open quarter dates, meaning a quarter that is still in progress, and a new BenchmarksError.NO_METRICS_FOUND error for input combinations that return nothing.

One limitation is easy to miss. Share metrics and the benchmark source's own rate metrics, such as the average CPM for an industry category, are not supported for open quarters. They are left out of the response unless the date range falls inside a narrower "all metrics" window. A dashboard showing current-quarter benchmarks may therefore show a percentile tier but no industry average CPM beside it. Google documents this behaviour, so the report is not broken.

For advertisers, the practical reading is the same as for the interface report. A percentile tier tells you where you rank on a metric, not whether that position is profitable. Google's release notes do not say how the peer set is defined beyond the category filter, so treat the tier as context when budget comes up. It is not a target.

Tracking templates now work at the Performance Max asset group level

v25.2 adds three URL options to AssetGroup: tracking_url_template, url_custom_parameters and final_url_suffix. Google describes them as the tracking URL template, a list of custom parameter mappings used to substitute tags in the template or final URLs, and a suffix appended to landing page URLs served with parallel tracking.

In account terms, this means Performance Max tracking can be set more precisely. A retailer running one Performance Max campaign with separate asset groups for different product lines could append a different UTM value or custom parameter per asset group, which makes it easier to tell in Google Analytics 4 which asset group sent a session. Until a tool exposes these fields, the setting has to be managed wherever your platform supports it. Anyone who has struggled to attribute Performance Max traffic in GA4 should check whether their tracking stack will use this, and a GA4 setup review is the natural place to decide the naming scheme before values start flowing.

Automated video crawl settings for Performance Max

The release adds AutomatedVideoCrawlSetting to the campaign's asset automation settings, with a new AssetAutomationType.AUTOMATED_VIDEO_CRAWL value, for Performance Max campaigns. Each entry configures a crawl URL, a source platform and an enabled opt-in status. The source platform can be LANDING_PAGE, SOCIAL or YOUTUBE.

The release notes describe the setting and its fields. They do not describe how Google uses the crawled videos, how often it crawls, or what review applies. What an advertiser can take from it is that the choice of where Performance Max may look for video is now something a tool can read and set through the API, per platform, with an explicit on or off.

Smart campaigns can be turned into Performance Max drafts

Google added SmartCampaignSettingService.GeneratePMaxDraftCampaign, which generates a Performance Max draft from an existing Smart campaign. The draft is created with a status of PAUSED and a creation status of INCOMPLETE, so nothing starts spending automatically.

A validate-only mode lets a tool check the conversion first and receive validation warnings or conversion issues without creating anything. When a draft is created, the response returns the new campaign, budget, asset group and assets. Two options, gbp_enabled and image_enabled, are not yet supported; setting them to true returns an error, according to the release notes.

This one is most relevant to small businesses and to the platforms and resellers that serve them, since Smart campaigns are Google's simplified format. A draft-first, paused-by-default path is the safer design: someone still has to review the assets, budget and goals before the Performance Max campaign runs.

Firebase ad impressions and in-app ad revenue

On conversions, v25.2 exposes two conversion action types, FIREBASE_ANDROID_APP_AD_IMPRESSION and FIREBASE_IOS_APP_AD_IMPRESSION, so they can be retrieved and modified through the API. It also exposes an IN_APP_AD_REVENUE category in ConversionActionCategory for tracking in-app ad monetization revenue.

This affects app publishers who earn from ads shown inside their own apps. Conversion audit tools and reporting connectors that list every conversion action can now see and classify these properly, which reduces the "unknown" rows that tend to appear when a tool meets a type it does not recognise.

Clearer Business Profile location sync errors

When a LOCATION_SYNC asset set fails to sync locations from Business Profile, the API previously returned a generic UNKNOWN error. From v25.2 it returns BusinessProfileLocationSyncError with a specific reason: EMAIL_INVALID, OAUTH_TOKEN_INVALID, ACCOUNT_REMOVED or NO_PERMISSION_TO_ACCESS_LOCATION_ACCOUNT.

For multi-location businesses this change is useful. Location assets are what put an address and distance next to an ad, and a silent sync failure can remove them without an obvious cause. A named reason such as an expired OAuth token or a lost permission points straight at the fix, usually on the Business Profile side. Businesses that rely on local SEO and Business Profile management should ask whoever manages their account link to surface these reasons rather than a generic failure.

Two new bid-too-low recommendations

v25.2 adds two recommendation types for Search campaigns: RAISE_TARGET_CPA_PERFORMANCE_BID_TOO_LOW and LOWER_TARGET_ROAS_PERFORMANCE_BID_TOO_LOW. Google says they support raising Target CPA or lowering Target ROAS when bids are too low for Search campaigns to enter auctions.

The CPA version provides a recommended target multiplier greater than 1.0 plus the current average target CPA. The ROAS version provides a multiplier below 1.0 plus the current average target ROAS. Both can be applied through the API.

Here the caution is about auto-apply. Any tool that applies recommendations automatically could now loosen a CPA or ROAS target because Google judges the bid too low to compete. That may be right for a campaign starved of impressions, but it also raises cost per conversion by design. Advertisers should check whether their tools apply recommendations without review, and decide whether these two belong on the list. Loosening a target is a margin decision, so it belongs with whoever owns the budget.

New hotel and travel reporting fields

The reporting additions are aimed at hotel and travel advertisers. New itinerary segments cover advance booking window in days, length of booking in days, booking start date, start day of week and whether the user selected their own dates. New vertical-ads metrics cover click-weighted average booking value in micros, potential impressions and the percentage difference between the advertiser's price and the lowest competing price for the same property. New vertical-ads segments cover price tier compared to competitors, rate rule ID and rate type. All are selectable with campaign, ad group and customer reports.

For a hotel, the price difference percentage and price tier are the ones to watch. They show whether a property is losing bookings because a competing rate for the same hotel is cheaper, which no amount of bid change fixes.

Creator insights by YouTube handle

ContentCreatorInsightsService.GenerateCreatorInsights can now search by YouTube channel handle through YouTubeChannels.youtube_channel_handles. Handles must start with the @ sign. Google describes it as similar to searching by channel ID. Planning tools can let a media buyer paste the handle they already know instead of looking up an ID.

What changed in v25.2 at a glance

The table below lists the v25.2 changes with the most direct account impact, as described in Google's release notes of 23 September 2026.

Other items in the release include the Customer Lifecycle Optimization validation errors, such as HIGH_LIFETIME_VALUE_MULTIPLIER_LESS_THAN_OR_EQUAL_TO_VALUE_MULTIPLIER, which rejects a high lifetime value multiplier that is not greater than the standard value multiplier. That is a guard against a misconfiguration rather than a new feature.

What the release notes do not say

Google's notes describe fields and behaviours. They do not give dates for when third-party tools will support them, do not explain how the benchmark peer set is built beyond the category filter, and do not say how automated video crawl results are used in ads. Any claim beyond what is written in the notes should be treated as the vendor's own interpretation until Google documents it.

What this means for Thai marketers

Nothing in v25.2 is described as limited by country, and the API is the same for accounts in Thailand as anywhere else. The useful steps are practical. Ask your reporting or bidding vendor whether they plan to support asset group tracking templates and the benchmark percentile tiers. Check whether any tool auto-applies recommendations, now that two of them can loosen CPA and ROAS targets. For businesses with many branches in Bangkok or across provinces, make sure a failing Business Profile location sync is reported with its named reason. Hotels running travel ads should look at the new price difference and price tier data once their reporting tool exposes it.

FAQ

Do I need to do anything because of Google Ads API v25.2?

No, v25.2 is a minor release with no breaking changes, so existing tools keep working. The only action worth taking is to ask your software vendors which new fields they plan to support, and to review auto-applied recommendations.

Is the percentile benchmark data the same as the Spend benchmarks report?

Not exactly: the Spend benchmarks report in the Account Overview shows your spend and clicks beside a peer figure, while v25.2 lets API software request a percentile tier showing where you rank. The API data only appears when the request asks for PERCENTILE_DATA with the all-advertisers source and a category filter.

Can I now set UTM parameters per Performance Max asset group?

Yes, v25.2 adds a tracking template, custom parameters and a final URL suffix to asset groups through the API. Whether you can set them in your own workflow depends on your tools adopting the fields.

Will my Smart campaign be converted to Performance Max automatically?

No, the new method creates a Performance Max draft that starts paused with an incomplete creation status. Someone has to review and launch it, and a validate-only mode can check the conversion without creating anything.

Should I accept the new raise Target CPA recommendation?

Only after checking your margins, because raising Target CPA or lowering Target ROAS is designed to increase what you pay per conversion. Google offers it when Search bids are too low to enter auctions. That is a volume problem, and it says nothing about whether the higher target is profitable.

v25.2 is a quiet release, but parts of it touch decisions advertisers own, especially tracking and bidding. If you want a second view on how your campaigns and automation are set up, the Google Ads management team at Relevant Audience can review the account with you.