Ahrefs Study Finds No AI Citation Boost from JSON‑LD Schema

Ahrefs’ latest report examined 1,885 web pages that added JSON‑LD schema markup and compared their AI citation performance to a matched set of control pages that never used schema. The analysis focused on three Google AI interfaces—AI Overviews, AI Mode, and the ChatGPT sidebar—measuring changes in the number of citations over a 30‑day period. The results showed a −4.6 % change in AI Overviews, +2.4 % in AI Mode, and +2.2 % in ChatGPT, all of which fell within statistical noise and were deemed insignificant. The study used a controlled difference‑in‑differences approach, pairing each schema page with three control pages from different domains that had similar pre‑existing citation levels. The authors noted that the 30‑day window might miss slower effects and that other concurrent page changes could confound the results.

The pages that received schema were already high‑citation before the markup was added, with many boasting over 100 AI Overview citations. This suggests that schema correlates with overall site quality rather than directly driving AI citations. The report also highlighted that the AI systems tested did not parse JSON‑LD during real‑time fetches, limiting the potential impact of structured data on citation algorithms. While the study did not cover pages that had not yet been cited by AI, it leaves open the possibility that schema could aid crawling or indexing for unseen content.

Google’s recent decision to retire FAQ rich results and other structured data types such as HowTo and Course Info further narrows the visible rewards for schema. Despite this, structured data continues to support rich results and makes content machine‑readable, but the marketing narrative that schema boosts AI citations lacks evidence. Marketers are advised to focus on content quality and proper indexing rather than investing heavily in schema for AI citation gains.