The Search Console Branded Queries Filter: Brand vs Non-Brand Traffic

#01What counts as a branded query?

Google's branded queries announcement from 20 November 2025 says branded queries are those that include your brand name, variations or misspellings, and unique products or services associated with the brand. Non-branded queries are everything else, for example "invoice approval software" rather than "ledgerline invoice approval".

#02How is it classified?

Google uses an AI-assisted system that considers the brand name in multiple languages and close variants. That means you don't maintain a list, but also that you can't correct it. The Insights report help page is explicit: "Query classification is powered by AI and may mislabel branded and non-branded queries."

#03Who gets it?

#04Where do you find it?

• Performance report: add a query filter and choose Branded or Non-branded, then use every other tab as normal.
• Insights report: the Branded and non-branded traffic card shows the split for the selected period.

#05Which marketing questions does it answer?

Our view: the non-branded line is the most honest SEO metric most teams have. Report it separately from total clicks, especially when brand campaigns are running.

#06What does the filter change in your reporting?

Splitting brand from non-brand changes how you read almost every trend.

#07Troubleshooting

#08A fallback regex when the filter isn't available

If your property doesn't get the filter, use a regex on Queries. Include the brand, common misspellings, product names and the domain:

(ledgerline|ledger line|ledgerlin|ledgerline\.example|ll ap)

Run it once as Matches regex and once as Doesn't match regex. Remember that the two won't add up to the total because anonymised queries drop out when you filter by query. Review the list quarterly as you launch products.

#09How to report the split to leadership

A brand split changes the conversation, so present it carefully.

1. Lead with non-branded clicks as the SEO metric, trended weekly or monthly.
2. Show branded clicks separately and call them a brand-demand signal shaped by PR, product and paid media.
3. Annotate launches, campaigns and PR moments on both lines.
4. Explain the caveats once: classification is automatic and can mislabel queries, and filtered totals won't add up to the overall total.
5. Don't compare with last year's regex numbers without checking how the definitions differ.

#10Edge cases

• Generic brand names. If your brand is also a common word, expect more misclassification in both directions. Sample queries every month.
• Acquired brands and product renames. The old name may still be classified as branded for a while, or not at all. Keep both names in your regex fallback.
• Competitor-plus-category searches such as "Competitor A alternatives" are non-branded for you. They often belong in your comparison content programme.
• Multiple languages. Google says the system considers multiple languages, but check transliterations and local spellings yourself.

#11A combined view hack

Use the branded filter and a comparison-intent regex together: Non-branded + (vs|alternative|compare) shows people evaluating the category without naming you. That segment is often the highest-value SEO opportunity for a B2B site.

#12Ledgerline walkthrough: setting up the split

1. The SEO lead opened the Domain property, not the /blog/ path property, because the filter isn't available on sub-properties.
2. They applied Non-branded, set 12 months at Monthly granularity and exported.
3. They repeated with Branded and placed both lines in the monthly sheet with campaign dates.
4. They sampled 30 queries from each side and found the product nickname mislabelled, so they kept a regex check for it.

#13Where this fits in the monthly routine

In week 1 of the routine (lesson 21), export the branded and non-branded lines and add them to the monthly sheet. Flag any month where branded rises sharply, then check it against campaigns and PR.

#14Worked example: Ledgerline's brand split

Ledgerline is a fictional accounts-payable software company. Details and numbers are illustrative.

Ledgerline's total clicks rose after a funding announcement, and the board deck credited SEO. With the branded filter, the lift was almost entirely in branded queries in the two weeks after the news, while non-branded clicks were flat. Sampling showed the classifier labelled "LL AP" (an internal product nickname used by customers) as non-branded, so the team added it to their regex check. The board deck now shows non-branded clicks as the SEO metric and branded clicks as a brand-demand signal.

Walkthrough

1. Open Performance > Search results in your Domain property.
2. Click + Add filter and look for the Query > Branded / Non-branded option.
3. Choose Non-branded and set the date range to the last 6 months at Weekly granularity.
4. Click the Queries tab and scan the top 50 rows. Note any that are actually branded.
5. Switch the filter to Branded and scan again for misclassified generic queries.
6. Open Insights and read the branded and non-branded traffic card for the same period.
7. Export both views. Record the share of clicks that are non-branded as your baseline.

#15Try it yourself

Build a brand-split baseline:

1. Export branded and non-branded clicks by month for the last 12 months (or as far as your data goes).
2. Mark months with launches, PR or campaigns.
3. Note whether non-branded clicks grew, fell or held steady across the year.
4. Sample 30 queries from each side and count misclassifications. If more than a few are wrong, add a regex check to your routine.

#16Quick quiz

1. Is the branded queries filter rule-based?

Answer: No. It uses an AI-assisted system and can misclassify queries.

2. Is the filter available on https://example.com/blog/?

Answer: No. It isn't available for sub-properties such as path or subdomain properties.

3. Does being labelled branded affect rankings?

Answer: No. The classifications are for information only.

4. Why might branded + non-branded regex not add up to the total?

Answer: Anonymised queries are dropped when you filter by query.

#17What should you check before the next lesson?

• You know whether your property has the branded filter.
• You have a non-branded clicks baseline for the last 6-12 months.
• You sampled queries on both sides for misclassification.
• For brand demand beyond Google, read Tracking ChatGPT referral traffic and brand mentions.

Next: Insights report and query groups. Or return to the Search Console for Marketers track.

#18Common questions

When did the branded queries filter launch? Google announced it on 20 November 2025 and later made it available to all eligible sites.

Does it work on Discover data? The filter applies to Search results data. Check the filter options in each report you use.

Can I edit what counts as branded? No. The classification is automatic. Use a regex filter alongside it if you need a custom definition.