DerivateX Research

Reports

Original benchmark studies on how Google AI Overviews and ChatGPT choose which B2B software to recommend and cite. Every report publishes its methodology.

GEO Statistics 2026 Benchmark Report
Statistics reportJune 18–24, 2026 Google AIO panel

Generative Engine Optimization Statistics 2026

12.2% first-party citation rate

Only 12.2% of AI Overview citations in this panel pointed at a recommended product’s own site (154 of 1,259).

  • 12.2%first-party citation rate (154 of 1,259)
  • 67.4%Citation Gap, sources not also in organic (848 of 1,259)
  • 63.4%listicle / “best X” share (798 of 1,259)
  • 30 of 100ghost queries, products recommended, zero own-site cites
Read report : Generative Engine Optimization Statistics 2026
The Agreement Gap
AI search benchmarkJune 30, 2026

The Agreement Gap: ChatGPT and Google AI Overviews recommend the same tools, then cite almost none of the same sources

4% same sources cited

Across 15 buyer-intent queries and 402 citations, ChatGPT and Google AI Overviews agreed on 32% of the tools they recommended but only 4% of the sources they cited.

Read report : The Agreement Gap: ChatGPT and Google AI Overviews recommend the same tools, then cite almost none of the same sources

How these reports are built

  • Real buyer queries100 buyer-intent queries across 20 B2B software categories, run through Google AI Overviews, with 1,259 citations logged.
  • Repeated runs40 B2B SaaS categories in ChatGPT with web search, each buyer question run ten times (233 recommendations).
  • Cross-engine scoring50 B2B SaaS companies scored out of 100 across ChatGPT, Perplexity, Claude and Gemini, using 1,400 buyer prompts.
  • Method on the pageEvery report includes a methodology section explaining how the study was run.