Stop optimizing for Perplexity. Seriously, look at your numbers

This is going to be unpopular with people who have built a practice around it, and it comes with a genuine caveat I will get to. But look at the traffic split before you allocate another quarter of effort: ChatGPT sends roughly ten times the referral traffic Perplexity does, and a great many AEO programmes still split their attention evenly across engines because a dashboard has four columns.

The split

SE Ranking's analysis of AI referral traffic puts ChatGPT at roughly 75% of referrals sent to websites, with Perplexity around 7%. Other measurements differ on the exact figures, and Similarweb's panel data shows ChatGPT's share of generative AI usage declining as competitors grow, but no credible dataset puts Perplexity anywhere near parity on traffic sent.

Meanwhile the two engines want almost opposite things. The cross-engine citation data shows Perplexity leaning heavily on Reddit and community sources while ChatGPT leans on Wikipedia and editorial publishers, and only about 11% of domains are cited by both. Splitting effort evenly across two engines that share a tenth of their source lists is not diversification. It is doing two jobs at half strength.

The caveat, which is real

Three things complicate the simple version. Perplexity is the only major engine currently paying publishers for citations, which matters enormously if you are a publisher rather than a brand. Traffic share is not influence share, and as our zero-click piece argues, referral volume is the least reliable measure in this entire channel. And audience concentration beats market share every time: if your buyers are engineers or analysts who live in Perplexity, the global average is irrelevant to you.

"Splitting effort evenly across two engines that share a tenth of their source lists is not diversification. It is doing two jobs at half strength."

What to do this week

Stop treating the four columns in your dashboard as four equal priorities. Find out which engine your actual buyers use, by asking them, because nobody has ever regretted asking ten customers where they research. Then weight effort to that answer, not to the tool's layout. For most B2C and general B2B, that means ChatGPT gets the majority of the work and the community layer serves Perplexity as a by-product. If you publish, invert it, because the engine paying for citations deserves attention disproportionate to its traffic. What nobody should be doing is splitting evenly by default because the interface implies it.