Key details

  1. Announced November 16, 2026 in Google's Search Central blog.
  2. brandedClassification groups or filters BRANDED and NON_BRANDED only for eligible top-level properties.
  3. Branded history starts February 21, 2026; brandedClassification cannot be combined with a query filter.
  4. type=webMultimodal covers Lens, Circle to Search and image uploads.
  5. filterLists combines multiple countries or devices in one API request.
  6. Some client-library generated types may lag the live discovery document.

What builders should take away

  1. Update automated reporting only after checking the API reference and eligibility behaviour.
  2. Treat zero branded rows as possibly unsupported data, not automatically zero branded searches.
  3. Preserve a separate baseline for multimodal discovery; do not label it AI-answer traffic.
  4. Test client-library compatibility, especially typed SDKs, before deploying new filters.

What changed

On November 16, 2026, Google added three reporting capabilities to its Search Analytics API: the brandedClassification dimension/filter for branded versus non-branded queries; a webMultimodal search type for Lens, Circle to Search and image-upload searches; and filterLists to combine multiple countries or devices in a request. These features existed in Search Console's interface before API access. Google says multimodal data also appears in BigQuery bulk exports where enabled.

Why it matters

Developers building SEO dashboards and agency reporting no longer need to approximate brand splits entirely with hand-maintained keyword rules or stitch together multiple API calls for every country/device. Visual-search performance can be monitored automatically. But silently empty branded responses and type-definition lag can mislead dashboards, and none of these dimensions should be confused with proof of AI answer referrals or conversion impact.

Branded reporting is restricted, not universal

Group or filter by brandedClassification using BRANDED or NON_BRANDED. Google's API reference restricts the data to eligible top-level properties with sufficient search volume, beginning February 21, 2026. Unsupported properties return no rows, not an explicit ineligibility error. Branded classification cannot be combined with a query filter in the same request. Compare the new classification with an existing brand-keyword rule before changing longitudinal reporting; pre-February history cannot be reconstructed from this API dimension.

Visual search becomes scriptable

Set type to webMultimodal for Google Lens, Circle to Search on Android and image-upload searches. This describes image-led Google web-search entry points; it is not a direct AI Mode/AI Overview referral counter. Google says properties with active bulk export can access the multimodal data in BigQuery too. Establish a separate baseline rather than silently combining this new series with conventional web-search trends.

Fewer requests for regions and devices

The filterLists field supports several countries or several devices in a single Search Analytics query. This reduces per-country/per-device request fan-out for regional reporting, but API users should follow the documented filter-group structure and supported combinations instead of assuming arbitrary OR logic across unrelated dimensions.

Client libraries and dashboards may lag the live API

Independent implementation reporting on November 16 found that Google's live discovery document already described the new fields while some generated language bindings lagged: a Python client passed through request bodies despite an older bundled description, whereas a Node TypeScript type lacked filterLists. These are observed client-version caveats, not proof every SDK or future version behaves the same. Verify current library types and the wire request before deploying.

What to do now

Test the new query shapes against an eligible property; log and distinguish empty branded responses from zero actual traffic; compare classification with previous branded regex logic; version dashboards so historical trends remain interpretable; and retain analytics/conversion instrumentation for downstream outcomes. Google's announcement does not claim new Search ranking signals or extra organic traffic.

What to watch next

  • Whether Google expands branded-query eligibility and history.
  • Updates to generated client-library types for filterLists and webMultimodal.
  • Whether Google adds more granular AI-answer referral attribution to the API.

Still unclear

  • Google does not guarantee branded data for small or sub-property sites.
  • No independent evidence here establishes the business impact of the new reporting options.
  • SDK compatibility depends on library and version.

Sources

Direct reading behind this dossier.

4 sources
Search Analytics query API reference
Google for Developers primary_documentation

Live documentation for eligible branded properties, February 21 history and filter restrictions; date reflects inspection, not asserted page publication.

Discussion

Discussion is reader-contributed. Comments are not part of the BTN dossier or its editorial evidence.

0 visible comments

Join the discussion

Keep comments useful and relevant. Reader contributions may be moderated and are not BTN editorial evidence.

Sign in to comment