Search Console Branded Queries Filter: How to Use It

A healthy-looking click total in Search Console hides the question that matters most to a content team: were these people already looking for you, or did they find you while looking for something else? The branded queries filter answers that inside the Performance report. The short version: open the filter bar, choose the Query filter, pick "Branded" or "Non-branded", compare both views under identical settings, and read the split as a clue about your audience, not as a ranking score.

This guide is for SEO specialists, content leads and agency staff who need to report brand and non-brand traffic without overclaiming. It explains where the filter lives and how it classifies, how to read non-brand growth, what the filter cannot do and how to build a manual regex substitute, then closes with a report template and an action list. The rules cited come from Google's announcement post (Opens in a new tab) and the Search Console Help pages.

What the Filter Does and Where

Before using the numbers, know exactly what is being separated and when the option will simply be missing.

Two groups, defined by the brand

The common tasks help page (Opens in a new tab) defines branded queries as searches that include your brand name, domain or a brand-specific product or service. Non-branded queries are searches that do not include your brand name. The filter works across search types: web, image, video and news. Once applied, the announcement says, metrics such as impressions, clicks, average position and CTR are limited to the selected group. It also describes a card in the Insights report that splits total clicks between branded and non-branded.

Turning it on, click by click

1. Open your property and go to the Performance report for search results.
2. Set the date range and search type. Keep them identical if you plan to compare views.
3. Click the filter bar and select the "Query" filter.
4. Choose "Branded" or "Non-branded" from the filter options.
5. Switch between the "Queries" and "Pages" tabs to see which terms and URLs make up each group.

Classification is not a keyword list

The announcement says the classification is not based on keyword matching. An internal, AI-assisted system decides, accounting for the brand name in multiple languages, typos and queries about the site's unique products or services. The help page repeats that some queries may be wrongly identified as branded or non-branded and that the labels are for information only and do not affect ranking. Practically, this means the filter may be more forgiving than a pattern you write yourself, but you cannot inspect the rule.

When the filter is missing

The help pages state that the filter is not available for sub-properties, such as a property for a single path, or for sites with a low number of impressions. The announcement says only top-level properties are supported; URL-prefix path properties and subdomain properties are not. According to the dimensions page (Opens in a new tab), the filter keeps 16 months of history. The impression threshold is not published, so do not guess it.

If you need a refresher on the report itself, the basics of query and page views are in our performance report query and page analysis guide.

Reading Non-Brand Growth

Opening two views takes seconds. Drawing a fair conclusion from them takes discipline.

Record the pair, not the total

For each period note clicks, impressions, CTR and average position for both groups. A monthly table of these eight numbers shows how the two lines move relative to each other. The branded line generally reflects people who already know you, and the non-branded line reflects how findable your pages are for topic searches. That reading is a working convention, not a rule from the documentation.

What each metric tells you in each view

Per the performance report help, average position is the average position of the topmost result from your site, and in the table it is calculated per URL or grouping row. In the non-branded view, that average shifts whenever the query mix shifts, so read row-level positions before drawing a conclusion.

Decision rule: where did the change come from?

Apply this order. First, check which view moved. If only the branded line rose, the gain is demand you already had, perhaps from a campaign, a mention or an email. If the non-branded line rose, look for new pages, refreshed content or a wider set of queries. If both rose, report which share grew more. Neither movement alone proves success; it depends on what you were trying to achieve.

Impressions versus clicks

Rising impressions with flat clicks suggest you are visible in more searches without winning the click. Rising clicks with flat impressions suggest better performance on queries you already appeared for. Name which of these happened rather than saying "traffic grew".

Seasonality needs a year-over-year view

Because the filter holds 16 months, you can compare a month against the same month last year, which strips out much of the seasonal swing. The comparison feature on the advanced filtering page (Opens in a new tab) allows one comparison at a time and adds a Difference column, which is the easiest way to see the gap.

Example scenario: a software documentation site

A fictional software company publishes setup guides. In the non-branded view the guides appear for "how to configure" style queries, while the branded view is dominated by the product name. After a release announcement, branded clicks rise for a few weeks and then settle. The non-branded line barely moves. The right reading is that the announcement drove awareness among existing audiences, not that the guides found new readers. The scenario carries no numbers.

Which pages carry the non-brand load?

Switch the non-branded view to "Pages". If a handful of guides account for most of the clicks, your discovery depends on few assets and a single decline there would be visible. If the load is spread across many pages, the program is more resilient. For the wider measurement picture, including direct visits and brand mentions, see brand awareness measurement indicators, and for what the results page looks like when someone searches your name, how to audit your brand search results.

Limits and Manual Regex Fallback

A convenient filter still has edges. Knowing them keeps you from reporting numbers that do not hold up.

Why the two views will not add up

The overview help page (Opens in a new tab) says anonymized queries are dropped when filtering by query, and that you can therefore only approximate a branded share. The advanced filtering page adds that filtering by query or URL can affect report totals because of data truncation. Do not promise that branded plus non-branded equals the unfiltered total. The documentation does not say whether anonymized queries would be labelled branded or non-branded.

Building the regex fallback

If the filter is missing, use the query filter's other options. "Queries containing" and "Queries not containing" match a single substring, which is not case-sensitive but otherwise exact. "Custom (regex)" offers "Matches regex" by default or "Doesn't match regex". Syntax is RE2, matching is partial by default, and "(?-i)" switches on case-sensitive matching.

1. List every way people write your brand: spaced, joined, hyphenated, abbreviated, common misspellings and the domain.
2. Join them into one alternation. For a fictional brand "Northwind Candles", the pattern would list "northwind candles", "northwindcandles" and "north wind" as alternatives.
3. Apply "Matches regex" for the branded view.
4. Apply "Doesn't match regex" with the same pattern for the non-branded view.
5. Add product names that do not contain the brand, since they would otherwise count as non-brand.
6. Date and version the pattern whenever you change it.

Filter versus regex

Cross-check the two when you can

If both are available, put the top queries each method calls branded side by side. Queries your pattern misses usually reveal a variant to add. Queries the filter misses but the pattern catches are the places where classification could be wrong. Repeat this every few months, not once.

When your brand is a common word

A brand name that is also an everyday word will produce false matches in any method. Read the top 20 to 30 queries by clicks in each view and decide whether the split is acceptable for reporting. This is a practical quality check, not a documented requirement.

Dashboards

If you plan to reproduce the split in a dashboard tool, test it first. The documentation does not say whether the branded filter can be reproduced through a connector, so compare the dashboard figure with the interface figure before presenting it.

Report Template

The filter is most useful when the same figures are reported in the same order every month.

A one-page monthly layout

A sentence you can adapt

"Non-branded clicks rose this month, mostly through two guide pages, while branded clicks were broadly stable. Figures will not match the unfiltered total because anonymized queries are excluded when filtering by query." Replace the generic wording with your own numbers.

What not to promise

Do not call a non-brand gain "SEO success" or a branded dip "a problem" without context. Campaign timing, season and impressions all move these lines. Avoid any wording that suggests the filter improves rankings; the documentation says the classification has no effect on how Google Search ranking works.

Archive the evidence

Save a screenshot or an export of each month's views, together with the regex pattern if you used one. With 16 months of history in the filter, you will want a longer personal archive for year-on-year trend reports.

Three Short Case Sketches

The same split answers different questions depending on the site. These sketches are fictional and carry no figures.

Example scenario: a local services directory

A directory's branded queries are mostly its name plus a city. The non-branded view shows service-and-city searches landing on category pages. If branded clicks stay flat while non-branded clicks climb after new city pages launch, the report can credit the new pages with reaching people who did not know the directory. If the directory is new and impressions are low, the filter may not appear, and a regex becomes the working method.

Example scenario: a newsletter publisher

A publisher sends a weekly newsletter and notices branded clicks jump each week after sending. That pattern shows the newsletter nudging subscribers to search for the site. The non-branded view tells the separate story of whether archive articles are being found. Reporting both on one chart without the split would make the newsletter look like an SEO engine, which it is not.

Example scenario: a store with a common-word name

A shop whose brand is an everyday noun sees noisy classification in both methods. The analyst reads the top queries each month, adds product names and a few exclusions to the regex and keeps a note describing the known gaps. The monthly report states the method and its limits at the top so nobody mistakes the split for a precise count.

Action List

Reading the report only matters if it changes what you do next.

This week

1. Check whether the filter appears on your property; if not, note whether it is a sub-property or a low-impression site.
2. Export or screenshot last month's branded and non-branded views.
3. Write your brand's spellings and draft a regex.
4. Compare the top 30 queries from the filter and the regex.
5. Copy the report template and name an owner.

Findings and next steps

Turning visibility into more brand searches

Branded query volume is an indirect indicator of awareness work. How coverage and sponsored articles can raise brand searches, and how to measure that, is covered in growing brand searches with sponsored articles. No content effort guarantees a search volume outcome.

Quarterly maintenance

Revisit brand spellings, product names and your regex every quarter. A new product line, a rename or a new abbreviation can quietly age the pattern. Plan a year-over-year review using the filter's 16 months of history.

Using the Split in Content Planning

The non-branded view is where an editorial team finds its next assignments, because it shows both what already reaches you and what sits just out of reach.

Work the visibility gap

Sort the non-branded view by impressions and flag rows with a weak CTR. Those are searches where your page appeared and was passed over. Compare each query's intent with the page's title and description. If they disagree, rewrite the snippet text, note the date and re-check the same view a few weeks later. Present the change as a test with a measurement date, not as a promised outcome.

Spot cannibalization by reading query and page together

When several of your pages surface for the same non-branded query, the structure may be overlapping. Pick the page that best fits the intent, link to it from the others and review whether the weaker pages deserve a distinct angle. You do not control which page Google chooses, so treat this as a consolidation effort, not a guarantee.

Remember what the report cannot show

The non-branded view lists queries where you already appeared. Topics where you have no visibility are absent. Pair this review with keyword research to find clusters you are missing, and decide for each whether an existing page needs expansion or a new page is warranted.

Watch the bridge between the two lines

Some people who first meet you through a non-branded search may later search your name. A single filter cannot prove that path, but if the branded line tends to rise some weeks after the non-branded line grows, record it as an observation. Avoid claiming causation from the pattern alone.

Pre-Flight Checklist

A few checks before you publish numbers prevent most misreadings.

Property and data checks

• Is the property top-level? The filter is not offered on sub-properties.
• Are there enough impressions? Low-impression sites do not get the filter.
• Which search type are you looking at? The filter covers web, image, video and news, but volumes differ by type.
• Is the last day of the range final? The newest data can be preliminary and may change within hours, according to the performance report help.

Interpretation checks

• Do both views use the same date range and search type?
• If your brand is a common word, did you sample the top queries by hand?
• Do you have the campaign and seasonal calendar for the period?
• Did the report state that anonymized queries keep totals from reconciling?

Only after these pass should numbers go into the template. Figures recorded without them tend to produce inconsistent comparisons a few months later.

Frequent Mistakes

• Treating the filter as a ranking signal. The documentation says it has no effect on ranking.
• Expecting totals to reconcile. Anonymized queries are excluded.
• Trusting the classification blindly. Spot-check the top queries.
• Comparing under different settings. Date range and search type must match.
• Never updating the regex. New spellings leak into non-brand.
• Calling a branded dip bad by default. Check season and campaigns first.
• Looking for the filter on a sub-property. It is not offered there.

Frequently Asked Questions

What is the branded queries filter in Search Console?

It is a filter in the Performance report that separates queries into branded and non-branded groups. The help page defines branded queries as searches that include your brand name, domain or a brand-specific product or service, and non-branded queries as searches that do not include your brand name. The classification is informational and does not affect ranking.

Why can't I see the filter on my property?

The help pages say the filter is not available for sub-properties or for sites with a low number of impressions. The announcement post adds that only top-level properties are supported, so URL-prefix paths and subdomain properties are out. In those cases a manual regex filter is the alternative.

How does Google decide what is branded?

According to the announcement, the classification does not rely on keyword matching. An internal, AI-assisted system makes the call, taking into account the brand name in multiple languages, typos and queries about a site's unique products or services. Some queries may be misclassified.

Will a manual regex give me the same numbers?

Not exactly. A regex only catches the patterns you write, while the filter can account for typos and other languages. The upside of regex is that the rule is transparent and fully under your control, which makes it a good fallback or cross-check.

Why do branded and non-branded clicks not add up to the total?

The help page says anonymized queries are dropped when you filter by query, so you can only approximate a branded share. Totals from two filtered views can therefore sit below the unfiltered total.

Does growth in non-brand clicks mean my SEO is working?

Not by itself. It can mean people who did not know you are finding you, but you still need to rule out seasonality, impression changes and a shift in which pages get the traffic. Read queries, pages, impressions and average position together before drawing a conclusion.

Conclusion

The limits and time windows on Google's help pages can change; check the current page before you apply anything.

The branded queries filter lets one Performance report be read through two audiences: people who search for you by name and people who find you while searching for a topic. Its AI-assisted classification is convenient but can be wrong, and it is unavailable on sub-properties and low-impression sites; a manual regex is a transparent fallback that needs upkeep. Identical settings, a fixed caveat note and a findings-to-actions table are what make the split useful month after month. No step here guarantees a ranking result. The aim is to explain where your traffic comes from.

If you are planning coverage that helps people search for your brand by name, you can look at the guest post options on Akua Adforce; scope and terms are listed on the product page.