Non-brand terms bring new visitors. Branded terms capture people who already know you. If your tracked list mixes the two, you misread where growth can come from.
A branded keyword contains your brand name or a close variation of it and belongs to your domain alone. In Semrush's keyword guide that is the stated test, and Keyword.com's branded-versus-non-branded breakdown uses the same line. Their examples overlap: Semrush, Semrush tool for one; Adobe, Adobe Photoshop for the other. A non-brand keyword, by the same guides, covers the category alone, with no brand fragment or misspelling inside it.
Brand vs Non-Brand Keywords: The Difference
The test is whether one domain owns the phrase. A branded searcher already picked a destination and uses Google as a shortcut to reach it. A non-brand searcher describes a category, product, or problem and has no company in mind yet.
Keyword.com draws the contrast with one pair: best Samsung phones with high-end cameras carries a brand, while best Android phones with high-end cameras shops the same category on category terms alone. The same guide carries the split in table form, with branded as name-containing and bottom-of-funnel and non-brand as generic and top-and-middle, across audience and competition and conversion and content type.
Semrush notes in its own guide that studying both sides matters because they reach users at different stages, and its reports sort rows into Branded, Non-branded, and Branded-for-other-domains, the last covering terms that carry another domain's brand while the queried domain still ranks for them.
Non-Branded Keywords Examples and the Tricky Edge
Non-brand terms are generic descriptions of what someone wants. Shoe shop near me reads as non-brand against closest Nike shop near me as branded, alongside how to fix a leaky sink, best running shoes for beginners, and office cleaning company as further non-brand shapes. For Adobe, Keyword.com lists how to edit photos, best photo editing tools, how to create a reverse video, and converting video to GIFs as non-brand examples. Each describes the job early, before a company enters the choice.
Three edge cases decide whether a brand-containing phrase still counts as non-brand. Semrush works through them in its guide:
- Amazon discount looks branded but is not unique to Amazon's domain, because many shops target the phrase and use Amazon semantically rather than to navigate to Amazon.
- Dove is shared by the soap brand and Dove Chocolate, so neither domain can claim every dove query as uniquely its own.
- Chatham terms around a sandwich shop fail the uniqueness test because Chatham is also a city, so chatham deli or chatham menu point at a place as much as a business.
The rule from that guide: a phrase counts as branded only when it is unique to one domain. When many domains legitimately chase the same brand word, tooling files it as non-branded.
Which Terms Bring New Visitors
The two sides serve different stages. Keyword.com's breakdown and Semrush's guide agree on the shape:
| Branded | Non-brand | |
|---|---|---|
| Searcher | knows the business, wants that site or product | researching solutions, may not know your brand yet |
| Intent and funnel | navigational or transactional, bottom of funnel | informational or commercial research, top and middle of funnel |
| Traffic | narrower, fewer searches at decision stage | broader, higher volume from the wider category |
| Conversion | higher, intent is fixed on the brand | lower at first, searchers still compare options |
| Content that fits | homepage, core product pages, direct FAQs | blog posts, guides, category and landing pages |
Non-brand wins show up as milestones for exactly this reason. In one OpenSEO builder's thread the moment is named as a first big non-branded keyword, backlink checker, with click-through rate (the share of views that become clicks) as the next lever to lift, and the authority behind it described as branded queries plus backlinks plus popular YouTube videos and reels driving interest. The branded base helped the non-brand term move. A separate go-to-market analysis of five AI video platforms hunts from the other side: after measuring traffic scale and channel dependence, it asks which non-brand search demands hold SEO opportunity. One side celebrates the win; the other hunts where the next one sits.
Read your own list with that frontier in mind. Branded rows defend what you already earned. Non-brand rows decide whether anyone new finds you.
Check Brand vs Category Ownership Live
Expect different resistance on each side. The owner almost always holds first place for its own name, so branded competition stays very low, while generic terms draw stiffer contests from industry peers and large publishers across both unpaid listings (the regular results, not ads) and paid results. That asymmetry sets the content call: a homepage can hold a brand name with little fight, while a category term needs a guide or landing page strong enough to displace whoever holds the top now.
Semrush's guide says its reports sort a term as branded for a domain when the results page shows sitelinks carrying that site's address (the sub-links Google shows beneath a main result), a Knowledge Graph box linking to that site (the entity panel shown beside results), or the brand phrase placing that domain in the first unpaid listing, a method it illustrates with iPhone and Apple screenshots.
Run the ownership read on any term you are unsure about:
- Pull the live results page for the query, the ordered list Google returns, pinned to the country and language you care about.
- Ask whether sub-links, the entity box, and the first unpaid listing point at one domain. When all three line up, the term behaves as that owner's brand term.
- When the top mixes peers, publishers, and roundups with no domain repeating across those signals, the term behaves as open category ground and the pages at the top show the format to match.
- Stamp each read with its date. Ownership shifts when a new guide or category page takes the top.
Doing that read by hand works for a handful of terms. For a whole tracked list, pull the results pages programmatically: an agent working through qequals can fetch each query's current Google results and check which domain holds the first listing, row by row.
Then split the list for good. Keep the branded rows as the base you defend, and treat the non-brand rows as the ones that measure whether new people are finding you.
