Should I stop writing listicles entirely?

No — context matters. If you're optimizing for Google AIO (still 41% listicle share) or Gemini (27%), listicles remain the dominant format. If you're optimizing for Perplexity or ChatGPT Search, comparison content is now the higher-leverage bet. Most teams should be writing both: a strong comparison page targeting commercial intent, and a complementary listicle that surfaces it across all your category-adjacent queries.

How long until a new comparison article starts earning citations?

From real engagement data: Perplexity typically begins citing a well-structured comparison page within 36–72 hours of publication; ChatGPT Search takes 3–5 days; Claude takes 2–3 weeks. The lag is mostly indexing, not content quality.

Is this dataset large enough to act on?

Caveat first: 320 prompts is a useful directional signal, not a statistical proof. The deltas we report (4–6pp) are large enough to be meaningful given the sample size, but a single weekly snapshot can be noisy. We recommend treating four-week trends as the reliable signal — which is exactly what this monthly report covers.

State of AI Citations

Q: What makes a comparison article 'well-structured' for AI citation?

Four things: (1) explicit X-vs-Y framing in the H1 and meta description; (2) a feature-comparison table with clear column headers; (3) a one-sentence verdict in the opening paragraph (LLMs lift the first sentence disproportionately); (4) FAQPage schema covering the head-to-head questions buyers actually ask.

May was the cleanest signal we’ve measured since the benchmark started running. Two coordinated movements (comparison content up on ChatGPT and Perplexity by 4pp each) and one outlier (Google AIO’s continued listicle dominance). Below is the data, our interpretation, and the actionable implication for what to ship in June.

The headline movement: comparisons gain on ChatGPT and Perplexity#

Two surfaces, same direction, same magnitude. ChatGPT moved comparison citation share from 24% to 28% (+4pp). Perplexity moved from 30% to 34% (+4pp). Neither moved by chance — the deltas hold across most categories, with B2B SaaS being the most extreme: Perplexity now gives 42% of its B2B SaaS citations to comparison-shaped content, up from 37%. Fintech mirrors this almost exactly (35% → 39%).

Why now? Best guess: both surfaces have refined their re-rankers to prefer structurally-explicit answers for commercial-intent queries. “What’s the best X for Y” rewards pages with explicit X-vs-Y framing in the H1, a feature table, and a one-sentence verdict in the opening paragraph more than it rewards a 3,500-word monolithic guide covering the same products. The structural penalty for monolithic content is meaningful and growing.

The listicle story: declining on most surfaces, still dominant on AIO#

Listicles lost share on three of five surfaces this week: Perplexity (-6pp), Google AIO (-5pp), and Gemini (-3pp). But listicles still own 41% of Google AIO citations — by a 13pp margin over the next surface (Gemini at 27%). This isn’t a death-of-listicles story. It’s an AIO-loves-listicles-more-than-anyone-else story.

The implication: if your category’s buyer journey goes through AIO (educational content, top-of-funnel research, head-term queries with informational intent), keep investing in listicle content. If your buyers are in commercial-intent surfaces (Perplexity for product research, ChatGPT for tool selection), shift the investment to comparison content.

What didn’t move: Claude#

Worth flagging because it’s surprising. Claude’s matrix shifted by less than 2pp on any cell this week. Claude remains the surface most rewarding to long-form guides (24% share) and explainers (21% share) — combined, that’s 45% of Claude’s citations going to depth-led content.

If you’re optimizing for Claude specifically (worth doing for high-consideration B2B buyers; Claude’s users skew technical and senior), the bet is on cluster authority and depth, not format-of-the-moment. Two long, well-cross-linked explainers will outrank ten thin pieces.

What’s coming next month#

Our four-week running trend (which becomes available in next month’s report once we have the full window) will tell us whether comparison content’s gain on ChatGPT and Perplexity is the start of a structural shift or a single-week wobble. If the May → June trajectory holds, expect comparison content’s combined share across those two surfaces to cross 35% by the end of Q2.

The data behind this report#

Pulled from The Visibly Index — 640 prompts queried weekly across ChatGPT, Perplexity, Claude, Gemini, and Google AI Overviews, with cited URLs classified into one of seven content types. Sample sizes per category and full methodology at /benchmark/methodology. To get the diagnosis specific to your domain, run the free audit.

State of AI Citations — May 2026