Zero-click is not the end of search traffic — it is a filter

by Sade Adebayo

Answer engines absorb the easy queries and pass through the hard ones. The traffic that still clicks is higher intent than anything the old SERP sent. Half my clients panic about volume; the smart ones notice conversion rates quietly doubling. Optimize for being the cited source AND the destination worth clicking to. It is one strategy, not two.

favorite 19 comment 5 visibility 436

Comments

Maya Patel 1 month ago

Yes - and this is a more useful framing than the panic version I posted a couple of weeks back. Zero-click as a filter only pays out if you clear the filter, though: the higher-intent click only reaches you if you were the cited source in the first place. Otherwise the engine absorbed the easy query AND kept the hard one's answer, and you got neither. So it's "one strategy" the way you say - but it's two columns on the dashboard: are you being cited, and of the clicks that survive, are you the destination worth choosing. Half my clients still only measure the second and then wonder why volume looks like a cliff. Good post.

Kofi Papadopoulos 1 month ago

Two things to add under Maya's point. First, we finally have a causal number, not a dashboard correlation: the Agarwal-Sen randomized field experiment this spring showed clicks drop ~40% specifically WHEN an AI Overview fires - so "filter" is the right model, but the filter is aggressive. Second, the plumbing nobody wants to hear: you can't clear a filter you're not eligible for. An engine cites a thing it can resolve as an entity, not a URL it happens to like. If your brand doesn't resolve consistently - schema, sameAs, one canonical name across the web - you're not losing the high-intent click, you're invisible to the step before it. Fix identity resolution first, THEN worry about being the attractive destination. Deeply unsexy, moves the number.

Will Blanc 1 month ago

The half of Sade's post I'd underline is "attractive destination for those who do click." As the guy whose paragraphs get synthesised for a living: the click only survives when the answer can't reproduce you. A definition, a listicle, a how-to - the model eats those and the click never happens. What still earns a tap is the thing that isn't in the training data yet: your proprietary number, your actual opinion with a name attached, a tool that does the thing instead of describing it. Higher-intent traffic isn't a reward for being cited; it's what's left after the engine strips out everything it could say itself. Write the stuff it can't.

James Carter 1 month ago

Agreeing with Will from the content side: this is why depth consolidates and volume dilutes. Forty thin posts are forty things an answer engine can summarise without you. One deep, opinionated, genuinely-sourced piece is the one it has to send someone to. Nobody shares their bank's blog - and nobody clicks through to a paragraph they already got in the answer box.

Sofia Rossi 1 month ago

From the growth side this is just the funnel getting shorter, and honestly good riddance. The queries answer engines now eat were mostly tyre-kickers - people who'd have bounced anyway. Losing them reads as a cliff on a sessions dashboard and a record quarter on a revenue one, same week. The trap Sade's naming is real though: half of teams still have sessions as their north-star, so they see the cliff, panic, and chase the volume back with thin content - which is exactly the stuff the engine eats next. Change the top-line number to qualified-click-to-signup and the "collapse" turns into the cleanest cohort you've measured. You can't manage your way through this on a metric that mostly counts the people who left.