SEO branding is the work that makes a company findable and recognisable when people search for its name, and that increases the number of people who search for that name. They are two different activities: covering the brand searches that already exist, and growing the brand demand that isn't there yet.
Most articles on this subject stop at the first point and describe it without a single number. Here I try to do the opposite: start from what brand demand is really worth, using the UK market as the example, then get to how you measure it on your own site.
What brand demand is worth in the UK
I took eight companies that sell in the UK across a broad product category and compared the search volume for the company name with that of the generic keyword describing its category. Google Ads data via DataForSEO, United Kingdom, average of the twelve months to August 2026, collected in September 2026.
| Brand | Searches/month for the name | Generic category keyword | Searches/month | Ratio |
|---|---|---|---|---|
| Tesco | 6,120,000 | online supermarket | 590 | 10,373x |
| Currys | 5,000,000 | electronics shop online | 260 | 19,231x |
| Aldi | 5,000,000 | supermarket offers | 480 | 10,417x |
| B&Q | 2,740,000 | diy store | 8,100 | 338x |
| Ryanair | 2,240,000 | cheap flights | 450,000 | 5x |
| ASOS | 1,830,000 | clothes online | 1,000 | 1,830x |
| Decathlon | 1,000,000 | sports equipment | 2,900 | 345x |
| Boohoo | 823,000 | online clothing store | 1,600 | 514x |
The comparison needs a caveat that usually goes unstated: on the right there is a single generic keyword, while real generic demand is split across dozens of variants. Someone shopping for trainers doesn't just type "sports equipment", they type "mens running shoes", "trail shoes", "kids football boots". Add up all the variants and the ratio drops a fair bit.
Even so, the point holds. In none of these eight categories do the generic variants add up to the millions of monthly searches the company name collects. The most balanced case is Ryanair, where "cheap flights" is a strong generic and the name is still worth five times as much. In the other seven the gap is measured in hundreds or thousands of times.
This explains why keyword research done only on generic terms describes a much smaller market than the real one. It also explains why two companies with the same offline market share can have wildly different organic traffic: the one whose name is searched more starts with a pool the other doesn't have.
Google doesn't have a brand score, it has a list of questions
Google's public documentation contains no ranking factor called "brand". It does, however, contain an operational document that asks human raters to judge the reputation of a site and its author, and to do so using sources outside the site itself.
The Search Quality Rater Guidelines, version of 11 September 2025, tell raters in section 3.3.1 to look for "independent reviews, references, recommendations by experts, news articles", and add a sentence that works as a job description: "Your job is to independently evaluate the Page Quality of the website, not just accept information that appears on one or two pages of the website without further verification". Reputation, for Google, is what others say about you. Not what you say on your About page.
Google's guidance on creating helpful content comes from the same direction. It asks whether the author is identifiable, whether the content shows "first-hand expertise and a depth of knowledge", and whether someone researching the site that publishes it would come away with the impression that it is "well-trusted or widely-recognized as an authority" on its topic. This is the framework known as E-E-A-T, and none of its four letters can be earned by working inside your own site.
Be careful not to draw a conclusion bigger than the premise. The rater guidelines are used to evaluate the algorithms, they aren't the algorithm, and Google repeats this with every version. They are the clearest description we have of what Google considers a good result, but not a list of active signals.
In generated answers, mentions count, not links
The most interesting data here comes from outside the SERP. In May 2025 Ahrefs analysed 75,000 brands to find out what correlates with a brand appearing in Google's AI Overviews. Spearman correlation, global sample.
| Signal | Correlation with mentions in AI Overviews |
|---|---|
| Brand mentions on the web (including unlinked) | 0.664 |
| Anchor text containing the brand name | 0.527 |
| Search volume for the brand name | 0.392 |
| Domain Rating | 0.326 |
| Referring domains | 0.295 |
| Traffic from branded searches | 0.274 |
| Backlinks | 0.218 |
Text mentions of the name, including those that carry no link, correlate three times more strongly than backlinks. In the same study, brands in the top quartile for web mentions collect up to ten times as many AI Overview mentions as the quartile just below, and 26% of the brands analysed never appear at all.
It is worth saying what this number doesn't prove. It is a correlation on a global sample, so it doesn't show that earning mentions makes a brand appear in the answers, and it hasn't been replicated market by market. The cautious reading is that the two grow together, which still changes the order of priorities: if you are building a strategy for visibility in generative engines, a citation in a trade article without a link is worth the time you put into it, and for years SEO taught the opposite. On the signals that influence generated answers, we have gathered what is known in a guide to AI Overviews ranking factors.
How to measure brand within SEO
The metric is branded search share: how much of your organic traffic comes from queries containing the company name. You can calculate it in Search Console in five minutes.
- Open the Performance report and set a twelve-month period, so seasonality cancels out.
- Apply a "contains" query filter with the company name. Add the misspellings: if you are called Visilay, "visilay srl", "vislay" and "visilay agency" count too.
- Note the filtered clicks.
- Remove the filter and note the total clicks in the same query table.
- Divide. That is your branded to non-branded ratio.
Step four is the one almost everyone gets wrong. Search Console anonymises rare queries, so the total clicks you see in the Queries tab are lower than the site's total clicks. If you take the numerator from the query table and the denominator from the overall chart, your brand share comes out significantly underestimated. Numerator and denominator must come from the same table.
There is no public benchmark for this metric at site level, and anyone offering you one is probably quoting a figure for all of Google search. The most solid reference available is the Ahrefs study from May 2025 on around 150 million US keywords: 36.9% of unique queries contain a brand or product name, and these queries account for 45.7% of total search volume. A US sample, and it refers to all Google searches, not the share for a single site.
So don't use it as a target. A healthy brand share depends on the business model: a general ecommerce site is often above 60%, a publisher can sit below 5% and be in excellent health. What matters is the direction over time and what drives it. If brand share grows while non-branded clicks stay flat, you are harvesting demand someone else generated. If non-branded grows and brand stays flat, you are bringing in people who don't remember whose site they visited. It is worth keeping alongside the other SEO KPIs in a monthly report.
The four levers that move brand share
Own the SERP for your name. When someone searches for your company, the first page shows your site and then eight results you don't control: LinkedIn, Glassdoor, company directories, old press coverage, a competitor bidding on your name in Ads. Look at the SERP for your brand in an incognito window and ask whether those nine results describe the company you are today. If you have physical premises, your Google Business Profile listing and reviews weigh as much as the site, and that takes you into local SEO.
Get others to name you. In the light of the Ahrefs numbers, an unlinked mention in a trade piece has a value nobody would have given it ten years ago. It changes how digital PR is done: the request is no longer "add the link", it is "use the full name". Links remain useful and link building isn't retiring, but it stops being the only goal of the same activity.
Publish something people can cite. The cheapest way to get named is to publish a figure that doesn't exist anywhere else. The table at the start of this article is built that way, and anyone with access to Keyword Planner can reproduce it: company names, category generics, a ratio. Original data, even on a small scale, works for years. An article that rewrites what is already on page one gets cited by nobody, not even by language models, which in their answers draw on sources that say something of their own.
Appear where search isn't Google. People search inside YouTube, Reddit, marketplaces, ChatGPT. Brand share on Google grows weeks behind exposure elsewhere, because the sequence is nearly always the same: someone sees you somewhere, then goes to Google and types your name. If you want to know where your brand appears and where it doesn't, there are AI visibility monitoring tools that track mentions across the various assistants.
In the UK the sequence is shorter than people think. According to Ofcom's Online Nation 2025 report, adults spend an average of four and a half hours online a day, most of it on a smartphone. The distance between seeing a brand and searching for it is two seconds and a thumb.
When brand isn't the answer
There is one case where all this reasoning has to be turned on its head, and that is industrial B2B. A buyer who needs to source a manufacturing process doesn't search for a supplier's name, because they don't know any. They search for the process, the material, the size.
For Macropix, a Milan-based manufacturer of custom LED screens and one of our clients, none of the nine keywords that bring the most qualified contacts contains the company name. The biggest is "ledwall", with 4,400 searches a month in Italy. The other eight describe products or applications. In that project, working on brand share would have been time badly spent: brand demand didn't exist and couldn't exist, because the market doesn't think in brands. The method we used there is described in the piece on SEO for manufacturers, and the broader principle in the guide to B2B SEO.
The practical rule I take from it: brand share is a goal when your market buys by brand, and a consequence when it buys by specification. In the first case you build it, in the second you collect it later, when the people who found you by specification start searching for you by name. That step comes for Macropix too, only it comes at the end rather than the beginning.
How to set up one route or the other for a specific company is what we work on in our SEO projects, and the results, with numbers and timescales, are among the case studies we have published.
Frequently asked questions
One thing the industry rarely says: branded search share is also the most honest way to notice that a company is buying traffic instead of building a market. A site that grows only on non-branded queries for three years running, with its own name flat throughout, has a problem no SEO fix will solve, because it is capturing demand without leaving a trace in the minds of the people passing through. The relationship between the two curves is the diagnosis. SEO is just one of the places where it gets treated.
