Branded vs Non-Branded Keywords: What Agencies Hide

Here is the oldest trick in agency reporting, and almost nobody does it on purpose. The dashboard says organic clicks are up 140% since the engagement started. The owner is thrilled. Nobody opens the query report.

Open it, filter out the business name, and a big share of that growth is people who already knew the company and typed its name into Google. That is traffic the agency did not create and could not have lost. It is not SEO. It is memory.

This post gives you the exact filter that splits the two, in the tool you already own, in about five minutes. No agency has an incentive to hand you a filter that shrinks its own reported wins. That is precisely why we publish it.

What is the difference between branded and non-branded search?

A branded keyword contains your business name or a recognizable variant of it; a non-branded keyword does not. 'Acme Plumbing reviews' is branded. 'emergency water heater repair' is non-branded. The distinction matters because they measure two completely different things.

Branded queries measure demand you already had. Someone saw your van, got a referral from a neighbor, remembered an ad, or bought from you two years ago. Google is a phone book at that moment, not a discovery engine.

Non-branded queries measure discovery. A stranger with a problem typed the problem, not your name, and landed on you anyway. That is the entire job you hired an SEO for. Everything else is bookkeeping.

Branded variants sprawl further than most owners expect: the legal name, the trading name, misspellings, the owner's surname, the name plus a city, the name plus 'reviews' or 'hours' or 'phone number,' and the domain typed as a search. All of it is branded. All of it should come out before anyone claims credit.

Why does branded traffic make an agency report look better than it is?

Because branded queries convert at rates non-branded queries cannot touch, and they carry the highest click-through rates on the SERP — so a small rise in branded volume moves the top-line clicks number hard. Blend the two, report one total, and the honest half of the work is hidden behind the flattering half.

It gets worse. Branded traffic tends to grow on its own as a business does any marketing at all. Run a radio spot, sponsor a little league team, wrap three trucks, ask for reviews — brand searches go up. If your SEO report counts those clicks, your SEO invoice is billing you for your truck wrap.

Nobody has to be a crook for this to happen. The default view of most reporting dashboards is total organic clicks. Laziness produces the same chart as fraud. Your job is to refuse to accept the blended number, whatever the intent behind it.

How do you split the two in Search Console in five minutes?

Use the Performance report's Custom (regex) query filter — it takes one expression and one toggle, and it works on the free version of the tool. Google's advanced filtering documentation confirms the filter uses RE2 syntax and that default regex matching is not case-sensitive, so you do not need to write every capitalization.

• Open Search Console > Performance > Search results. Set the date range to 16 months so you have a real baseline.
• Click + Add filter, choose Queries, then choose Custom (regex).
• Enter your brand pattern with pipes between variants: (acme|acme plumbing|acmeplumbing|acme plumbin|acme plumming). Include misspellings — people type badly.
• Leave it on Matches regex. Write down the clicks. That is your BRANDED number.
• Now edit the same filter and switch it to Doesn't match regex. Write down the clicks. That is your NON-BRANDED number.
• Repeat both for the same period last year, or for the month before the agency started. The delta in the non-branded number is the only delta that belongs to SEO.

Two traps. First, regex matching is a partial match by default — your expression matches anywhere in the query unless you anchor it with ^ or $. That is usually what you want ('acme' catches 'acme plumbing near me'), but it bites when your brand is a common word. If you are Summit Roofing, a bare 'summit' will also capture 'summit county roofing permit,' and you will under-report the SEO you actually earned. Anchor it: (summit roofing|summitroofing|summit roof co).

Second, the two halves will not add up to your total, and that is not a bug. Google states plainly that the query table caps at 1,000 rows and that anonymized rare queries are omitted from the table but counted in the chart total, so 'match' and 'does not match' totals may not sum to the unfiltered total. Compare branded to branded and non-branded to non-branded across periods. Never try to reconcile them to the headline number. If Search Console itself is new to you, start with our Search Console guide for service businesses.

Which number actually proves your SEO is working?

Non-branded clicks to pages that sell — and the booked calls that follow them. That is one number, and it is the only one that survives a hostile audit.

Non-branded clicks alone are still soft. A blog post about 'how much does a new furnace cost' can pull hundreds of non-branded clicks from people three states away who will never buy. Layer the page filter on top of the non-branded query filter and look at what lands on your service and location pages. That is intent that can pay you.

Then close the loop to revenue. Traffic is not a business outcome. We cover the full chain — click to call to close — in how to measure SEO ROI for a service business, and this is the exact reason we build reporting around booked calls instead of sessions.

The verdict: if an agency will only show you the first row, you are being managed, not served. Ask for the last two.

What if your brand searches came from a truck wrap, not SEO?

Then SEO is taking credit for someone else's budget, and you can prove it in one chart. Pull monthly branded clicks from Search Console and lay them next to your monthly spend on everything that creates brand recall — paid ads, direct mail, sponsorships, vehicle wraps, a review-request push.

Brand search is a lagging indicator of every other marketing dollar you spend. If branded queries jumped 60% in the same month you doubled Google Ads spend, the ads made those searches. The SEO retainer was in the car, not driving it.

Analytics muddies this further. In GA4, Google defines Organic Search as users arriving 'via non-ad links in organic-search results, including Google's AI Overviews and AI Mode,' while Direct is defined as arriving 'via a saved link or by entering your URL' (Google Analytics Help). A person who remembers your name will scatter across both buckets depending on whether they typed your name into Google or into the address bar. Search Console's query filter is cleaner than any GA4 channel report for this one job — use it.

Should an agency ever be paid on branded traffic growth?

Almost never — with one narrow exception: a genuine rebrand or a new-name launch where building brand recall is the written mandate. Outside of that, paying an SEO on branded growth is paying twice for demand you generated somewhere else.

It also creates a lousy incentive. An agency compensated on total organic clicks will happily chase brand-adjacent, low-effort queries and let the hard non-branded work rot. The metric you reward is the work you get.

Watch for the softer version of this too: a report that leads with impressions, average position, keyword counts, or 'pages indexed.' Those are the cousins of branded traffic — technically true, structurally useless. Cheap retainers lean on them hardest, which is a pattern we break down in what $500-a-month SEO actually buys.

What should a monthly SEO report show instead?

Six lines, in this order: non-branded clicks, non-branded clicks to money pages, booked calls attributed to organic, the pages that moved, what was shipped this month, and what ships next. If a report is longer than one screen, it is padding.

Branded traffic still belongs on the report — just in its own row, clearly labeled, and never counted as an SEO win. It is a useful health signal. A branded line that is falling while you are still spending on ads means something is wrong with your reputation, not your rankings.

Two more habits worth building. Pull the data yourself at least once a quarter — you own the Search Console property, so nobody can filter what you see. And remember Search Console reports on a 2-to-3-day lag in Pacific Time, so a report that shows yesterday's numbers is coming from somewhere else.

This is also why we do not sign 12-month contracts. If the non-branded line is flat after a fair runway, you should be able to walk, and our SEO service is built month-to-month so you can. An agency that needs a lock-in is telling you what it expects that line to do.

Want a second set of eyes on the report you have been handed? We will run your Search Console data through this exact split and tell you what share of your 'growth' was people who already knew your name. Get my free audit.