How much of your organic traffic is just your brand name

165 clicks from people who did not know the name

I got read access to the search data and pulled 28 days on our main site, split by whether the query contained our brand name. It took half a day, because the data had been sitting there the whole time.

The 165 is the number that describes the marketing. Brand search is people who already know the company and are navigating to it, and a search engine is doing the work of a bookmark. Four years of our content production sat behind that, and almost the entire organic result was people typing our name into a search box.

The subdomain that makes the number four times too big

The first time I ran it I got 637 non-brand clicks, because I had not filtered the report, and the difference is not marketing traffic.

The domain property contained a marketplace subdomain, so unfiltered it mixes somebody else's pages in with our own. Almost all of the difference sits on customer PDFs, product manuals that people who already bought are looking up. That is traffic belonging to our customers rather than to our content team.

This is the kind of error that never announces itself. Both reports run, both look reasonable, and the unfiltered one is nearly four times the real figure, 637 against 165.

The pages that rank and get skipped

Several of our pages had real reach and returned almost nothing, and I could see the cause in the click rate rather than in the position.

A click rate under 0.2 percent at position 8. Those pages rank, they get seen, and nobody clicks.

That is a title and snippet problem on a page that already exists, and it is the cheapest of the fixes available.

From six months inside a B2B GTM organization with 36,000 accounts in its CRM.

Our category's core keyword sat at position 30, which is effectively invisible. So the library ranked. It ranked on the wrong queries, and on the right ones it was seen and skipped.

Why this went unseen for four years

None of it was hidden and none of it needed a tool anybody had to buy.

We had never connected the search data, so every ranking statement in our content process stood on somebody's impression. That is the same shape as a campaign nobody evaluated after it went live, which is how a whitepaper counted downloads and produced nothing else. One consequence shows the cost cleanly, and I found it in the same afternoon: the same query was ranking on two of our own pages, one at position 3.6 and one at position 46. Cannibalization inside the index, running for years, invisible without the one report nobody had opened.

What to do instead

1. Make non-brand clicks the first number in any content review. Read it before anything gets written, and read it split from brand every time.
2. Filter to your own pages before you believe any figure. A domain property can contain subdomains that belong to somebody else, and the unfiltered number looks just as plausible.
3. Sort by impressions and low click rate, not by position. A page that already ranks and gets ignored needs a new title, and the click rate is the case for rewriting one instead of writing a new page.
4. Carry the date with every ranking figure. That data runs about three days behind and reaches 16 months back, so a process that treats it as today's number gives a confident wrong answer.
5. Check whether two of your own pages compete for one query. Nothing in the tool tells you, and it costs both of them.

The thresholds for when a page is worth updating had to come out of our own numbers. There is no industry benchmark for it, and anybody quoting one has invented it, which is worth carrying into every measurement conversation. The measurement discipline underneath it is the same one that decides whether a market size number is worth anything.