Branded vs non-branded split in Search Console: is a regex good enough?

Splitting branded and non-branded queries in Search Console sounds like a two minute job until you actually try it. Most people build a regex with the brand name and a few misspellings, then report the two buckets side by side. The problem is that it misses brand-adjacent queries like "acme pricing" or "acme vs x", which are often high intent, and it lumps in people who just typed your URL. So the non-branded number looks cleaner but ends up smaller than it should be.

Still, a rough split beats no split, as long as the report spells out what the regex catches. Otherwise the number looks precise when it isn't, and nobody questions it until a client asks why non-brand clicks dropped.

Search Console hides anonymized queries, so your branded and non-branded rows won't add up to the total clicks. That trips a lot of people up. Pull the unfiltered total and check both buckets against it before the split goes into a client report.