Q1 2026 AEO test results: 200 crypto sites, 60 prompts, 3 engines
Our quarterly AEO test ran in Q1 2026 across 200 crypto sites, 60 prompts, and 3 engines (ChatGPT, Perplexity, Claude). Median citation rate was 4.2%. The top 10 sites all share three traits. The bottom 50 share one. Here are the patterns and the underlying data.
Dataset: 200 sites, 60 prompts, 3 engines
200 sites split across categories: 40 DEXes, 30 CEXes, 30 wallets, 30 lending protocols, 25 L1s, 25 L2s, 20 NFT marketplaces. Sample chosen by TVL (DeFi), volume (exchanges), or user count (wallets).
60 prompts split: 20 informational, 20 commercial, 20 comparative. Identical prompt set to our Q4 2025 test for direct comparison.
Engines: ChatGPT (GPT-4o), Perplexity (Sonar Pro), Claude (Sonnet 4.6). Each prompt run on each engine, two runs per engine spaced 14 days apart. Citations counted only if they appeared in both runs.
Total citations recorded: 4,847 confirmed (out of 11,200 total mentions across all runs and engines).
Citation rate distribution
Citation rate = (prompts where the site appears in any engine's citation list) / (total prompts in the site's category).
Distribution across 200 sites: median 4.2%, mean 7.8%, top quartile starts at 9.6%, top decile starts at 17.4%. The top site (Defillama) had a 38% citation rate. The bottom 50 sites had under 1% (effectively zero).
The skew matters more than the median. AEO is winner-take-most: 20% of sites get 80% of citations in their categories. The 'long tail' of crypto sites with under 1% citation rate represents 56% of our dataset. These sites are functionally invisible to AI engines.
Variance by category: wallet pages had the highest median citation rate (6.8%), L1 pages had the lowest (2.1%). Wallets win because the queries are commercial and concrete. L1s lose because category queries are abstract and AI engines fall back to Wikipedia and Coingecko.
Top 10 cited crypto sites
Ranked by overall citation rate across all 60 prompts.
1. Defillama (38%). TVL data is the most-quoted crypto data on Perplexity and ChatGPT.
2. Etherscan (29%). Cited any time a query involves a contract address or transaction.
3. Uniswap (24%). The protocol docs at docs.uniswap.org drive most of the citations, not the marketing site.
4. MetaMask (22%). The MetaMask 'Learn' section gets cited on broad informational queries.
5. Coinbase (20%). The 'Learn' content at coinbase.com/learn drives most of the citation share, not the exchange product pages.
6. Aave (18%). Similar to Uniswap: docs drive citations.
7. Solana Foundation (17%). The official Solana docs at solana.com are heavily cited on L1 queries.
8. Lido (15%). Restaking and staking queries route through Lido more than any other staking site.
9. Optimism (14%). Documentation and the developer hub drive citations.
10. Arbitrum (13%). Same pattern: docs.arbitrum.io is the cited path, not the marketing site.
Pattern: 8 of 10 are cited primarily for their documentation, not their marketing pages. Crypto docs are the most underrated AEO surface in the space.
Patterns shared by the top 10
Three traits show up in all 10 top-cited sites.
Trait 1: complete schema across the entire site. Every page has at minimum FinancialProduct or FinancialService schema with author/organization metadata. FAQ schema on relevant pages. No missing canonical tags. No invalid JSON-LD.
Trait 2: dated, authored content. Articles and docs have visible publication dates, author names, and links to author pages with credentials. Generic 'team' bylines or no bylines are absent in this set.
Trait 3: documentation as a primary SEO surface. Treated docs as a content investment, not a developer-only resource. Cross-linked from marketing pages, indexed, in sitemap, schema-tagged.
What the bottom 50 share
The bottom 50 sites (those with under 1% citation rate) share one trait above all others: they have client-side rendered hero content with no SSR fallback.
How we tested: view-source on the homepage. If the H1 and primary product description were missing from the raw HTML and only appeared after JS execution, the site was flagged as CSR. 41 of 50 bottom sites had this issue.
The other 9 bottom sites had SSR but suffered from one of two issues: no schema at all, or schema that contradicted the visible content. Either kills citation eligibility.
Secondary issue: 31 of 50 bottom sites had no docs or developer hub. Their entire content footprint was a marketing site. This caps citation potential because AI engines route technical queries to docs first.
What changed Q4 2025 to Q1 2026
Five shifts in 12 weeks.
1. Reddit citations up 8%. Perplexity's reliance on Reddit kept growing. Now 23% of all Perplexity citations in our test (was 21%). ChatGPT also increased Reddit citations to 11% (from 8%).
2. Wikipedia citations down 12%. AI engines are routing fewer queries through Wikipedia, especially for tokens and protocols. Direct crypto sources are taking the share.
3. Documentation citations up 19%. The single largest movement. Protocol docs are eating into both Wikipedia and crypto news for technical queries.
4. Coingecko vs CoinMarketCap shifting. CG up 4%, CMC down 3%. The DeFi-specific data CG publishes (TVL, yield rates) is winning over CMC's broader market data.
5. News sites down across the board. Cointelegraph, Coindesk, Decrypt all lost share. AI engines increasingly skip news for category queries and only cite news on event-specific queries.
Common questions
Where can I find the raw data?
We publish the prompt list and aggregated citation counts at /resources/aeo-q1-2026-data/. Individual site-level results are available to enterprise audit customers.
Will the methodology change in Q2 2026?
Same 60 prompts, same 3 engines, same 14-day repeat window. Only the site sample updates to reflect new launches and TVL shifts.
Why is Defillama #1 by such a wide margin?
Numerical data with date stamps. AI engines treat freshly-stamped data as authoritative. Defillama's TVL pages get cited on every DeFi-related commercial query.
Can a new site reach the top decile in one quarter?
Rare but possible. Two sites in our test moved from under 1% to over 10% in Q1 by adding docs, schema, and consistent author bylines. Took roughly 80 days of consistent work.
Does this data apply outside crypto?
Citation rate distribution shape (winner-take-most) is universal. Specific source rankings are crypto-specific.
Audit your crypto site in 60 seconds
8-module deep scan. AI visibility, schema, technical SEO, backlinks. One domain free forever.