Google Adds Branded and Multimodal Data to the Search Console API

The Search Analytics API now supports branded query classification, multimodal search data, and filtering by multiple countries or devices in a single call.

· 4 min read · 5 comments

Google has added three Performance report features to the Search Analytics API in Search Console: branded query classification, multimodal search data, and filtering by multiple countries or devices in a single call. As Search Engine Journal reports, Google announced the changes on November 16. Developers can use these features in automated reports without stitching together separate API requests or maintaining complex lists of regular expressions.

Branded queries: a breakdown with limitations

The new brandedClassification dimension lets you group or filter results by BRANDED and NON_BRANDED. This makes it possible to analyze traffic from queries Google classifies as branded and non-branded separately, without maintaining your own list of company name variations. But you cannot simply swap the old filter for the new field across every report.

The breakdown is available only for top-level properties and only from February 21, 2026. You cannot use it to split data from earlier periods. If a property does not have enough queries and impressions, the API returns an empty table with no rows. So check what an empty response means before drawing conclusions about demand.

There are also restrictions on how the dimension can be used: it supports only the equals operator. You cannot filter or group by brandedClassification if the same request already includes a query filter. Teams that build reports around individual keyphrases will need to account for this when setting up their queries.

Google determines a query’s category using an AI-assisted process and has previously warned that classification errors are possible. If your team already uses a regular expression to identify branded queries, it is worth comparing the two breakdowns to see where they differ. The new field is easier to maintain, but its classifications should not be treated as error-free.

Multimodal data and filtering by country or device

The API now has a dedicated webMultimodal search type for multimodal search. According to the reference documentation, it covers searches through Google Lens, Circle to Search on Android, and image uploads to Google Search. This data appeared in the Search Console report on September 24; at the time, the API reference did not yet list a separate search type value. The reference documentation updated on November 16 does include it.

The scope of this data is not yet fully documented. Google’s September announcement mentioned image search through Chrome’s context menu, and it appears in the Performance report documentation. But neither the API announcement nor its reference documentation mentions Chrome. They also do not say how far back webMultimodal data is available.

Google also said multimodal data would be added to BigQuery for properties with bulk export enabled, but gave no timeline. As of November 16, the export table reference listed six search types, none of them multimodal. Another open question is whether the standard web metrics include webMultimodal data. Until that is clear, adding the two together without checking risks an inaccurate total.

filterLists: multiple values in one call

The filterLists field lets you select multiple countries or multiple device types in a single API request. Google gives EMEA, LATAM, and DACH as examples of country groups. For reports using these selections, that means fewer separate calls and less work combining the results.

The field supports only the country and device dimensions, and values must be exact matches. You cannot specify the same dimension in both regular filters and filterLists within one request. So before changing an existing reporting setup, check how its filters are configured.

What this means for website reports

Website owners and marketers can more easily isolate branded demand in Search Console data. But period comparisons need to account for the breakdown’s start date—February 21, 2026—and the possibility of an empty response for some properties. If a chart suddenly starts on that date, that does not mean the API update created demand.

Reporting developers can add a separate selection for multimodal search and reduce the number of requests needed to analyze countries or devices. Before publishing a combined chart, it is best to compare the web and webMultimodal values with the text and multimodal metrics in the Performance report. Otherwise, you risk counting data twice where the reference documentation has not yet explained whether the two overlap.

My take: automation does not replace verification

I like that recurring reports will take less work to query and less fiddling with branded regexes. But I would compare Google’s classification with the old filter first, and only then present the difference between branded and non-branded demand as a firm conclusion.

The same goes for multimodal figures: check what the metrics include before adding them up. The API can make a report tidier, but it will not decide how to interpret the data for us. And a new reporting field will not magically create useful, properly indexed pages—unfortunately, there is still no button for that.

Sources

Where the news comes from. The text is a retelling in the author’s own words; the facts come from the source, the opinion is the author’s.

  1. Google Adds Branded & Multimodal Data To Search Console API via @sejournal, @MattGSouthern www.searchenginejournal.com

About the author

Pavlo Ustymenko

Junior SEO specialist

I manage content and product listings for a small store while learning technical SEO. I check page speed more often than the store owner would like.

All posts by the author

Leave a comment

Not published. We’ll send a link to confirm your comment.

No links or ads. Your first comment is published after review. By sending a comment you agree to the privacy notice.

  1. In store reporting, I’ve found branded query lists become messy as product and brand names vary, so Google’s classification could save maintenance work. I’d still compare it with a manually reviewed query sample before using it to explain changes in non-branded visibility.

    1. @hanna.hakala, in my home goods store reports, product and brand-name variations make manual lists costly to maintain; I’d use Google’s split to reduce that work, but verify a query sample before attributing a visibility change.

      1. One extra check: I’d keep a small, reviewed set of ambiguous queries—especially product names that are also common words—and see how Google labels them over time. The classification can cut list maintenance, but it’s still Google’s interpretation, not a guarantee that every borderline query reflects brand intent.

      2. I’m not sure this reduces the work as much as it sounds: brandedClassification is limited to top-level reports, so I’d still keep the manual list for checks where that breakdown isn’t available.

        1. @igor.p, do you mean keeping the manual list only for reports where the brandedClassification breakdown is unavailable? For a small center, that distinction matters: I’d rather not maintain a second list unless it serves a specific check.