Are AI citations and brand recommendations telling different stories?
A Semrush analysis found that the most-cited domain in a ChatGPT category matched the most-mentioned brand only 20.8% of the time.
Think about someone asking ChatGPT to recommend project management software. The answer might reference an industry comparison article while recommending several software vendors.
The publication earns a citation, while the software companies earn brand mentions and potentially a place on the buyer's shortlist.
These are two different outcomes worth measuring. Citation share indicates which websites ChatGPT uses as sources, while brand recommendation frequency shows which companies appear when users ask for solutions.
A company could have strong citation visibility through its educational content while appearing less frequently in product recommendations.
For teams measuring AI search performance, are you tracking citations and brand recommendations as separate metrics?
Replies
Yes, and I think they need different work too. Citations come from being the page that answers the question well. Recommendations seem to come from how often you're named on pages you don't own: Reddit threads, directories, comparison posts. We're just starting to track a fixed set of buyer questions and log both separately, plus which third-party sources show up when we're missing. Early days, but the gaps already look different for each. Curious whether you've seen one reliably lead the other?