Ahrefs Highlights Limits of Referral Data for AI Search

AI ranked second only to Google as a source of signups for Ahrefs, placing it ahead of YouTube, according to the company’s blog. The finding offers a different view of AI’s contribution than referral traffic alone would have provided.

Ahrefs says relying only on referral data would have substantially understated AI’s role. Its broader point is that some influence occurs outside the activity captured in referral reports. For marketers assessing a channel, the distinction matters: a source that contributes to a signup is not necessarily represented fully by the traffic attributed to it.

The comparison is specific to Ahrefs. The available source summary does not give signup counts, a measurement period or details of the method used to identify signup sources. It therefore does not establish how AI compares with Google or YouTube for other businesses, or how much of Ahrefs’ signup activity each source represented.

The finding also does not mean referral reporting should be discarded. Rather, Ahrefs’ account highlights the risk of treating one measure as the complete answer to a broader question. A traffic report and a signup-source comparison can describe different aspects of performance, so marketers should be clear about which question each measure is intended to answer.

When reviewing AI search visibility, teams can separate the question of whether a brand appears in AI-related experiences from the question of how signups are attributed. They can also distinguish what their reports directly show from what those reports may not capture. That helps prevent a low referral figure from being treated, on its own, as a complete assessment of AI’s contribution.

What this means for marketers

Review referral traffic alongside any available information about signup sources, rather than using it as the sole measure of AI’s value. When sharing results, specify the metric and its limitations. Avoid applying Ahrefs’ ranking to your own business without comparable evidence, and be cautious about drawing conclusions where the available data does not show how a signup source was identified.