Search Console branded queries: split them out in reports or keep one total?

Search Console totals have the same problem almost everywhere: branded queries carry a big share of the clicks, so the non-brand story gets buried. Splitting them out with a regex on the query dimension is easy, but then you have to agree on what counts as branded. Misspellings, the product name without the company name, and founder names all leak in. My view is that you should split them, but only for trend reporting, not for judging the content team. Branded clicks mostly follow demand created by ads, PR and word of mouth, so SEO shouldn't get credit or blame for them. How do others handle the regex, and does anyone keep a maintained list of brand variants?

Agreed on the split, but branded traffic isn't purely "not SEO's" problem. Whether it's worth separating really depends on brand volume. If branded queries are most of your clicks, one total hides the non-brand trend, so splitting is essential. If they're a small slice, the split just adds maintenance for little gain. For the list itself, I'd keep the regex in one shared file and review it every quarter, since new product names and typos creep in fastest right after a campaign launch.

Quarterly is a fine cadence, but the regex has a blind spot. Search Console hides very low-volume queries for privacy, so those clicks show up as an anonymized row instead of a query. Your regex only sees the queries that cleared the threshold.

Say a clinic gets 1,200 clicks a month and the query table accounts for 900 of them. The other 300 can't be sorted into branded or non-brand, so your non-brand trend rests on a partial sample.

Compare the sum of the query rows with the property total and note the gap before you trust any split. If that gap grows, your branded share and non-brand trend are both drifting for reasons unrelated to demand. I'd track the gap right alongside the list review.

Atlas, I don't think the anonymized gap is a reason to distrust the split. It mostly hits the long tail, and that tail is almost all non-brand. Branded queries are high volume, so they clear the threshold and get counted. That means the gap usually drags your non-brand trend down a little rather than touching your branded share.

"Branded always clears the threshold" only holds for the main name. Misspellings and a brand-new product name can sit at a handful of clicks for months, and that's exactly the post-launch window you flagged. So some branded clicks may end up in the anonymized bucket and inflate the non-brand trend. Which way the bias goes depends on which queries are rare, not just on whether they're branded. Do your common misspellings clear the threshold at all?