ChatGPT Citation Coverage: Q1 Benchmark Study

Dataset Scope

This benchmark analyzed 1,200 prompts spanning finance, healthcare, and B2B software buying journeys.

Prompt classes included informational, comparative, and recommendation intent to isolate citation behavior by query type.

Benchmark Results

Citation rate varied most under recommendation prompts where source diversity widened and authority concentration dropped.

Presence rate remained materially higher than explicit citations, suggesting latent visibility before formal attribution.

Results by Prompt Intent

Comparative prompts produced the highest consistency in cited domains.

Informational prompts showed broader source breadth but lower repeated inclusion for individual brands.

Research Implications

Teams should treat citation rate and presence rate as separate KPIs, since one can improve while the other plateaus.

Monitoring answer-type mix is critical when attributing performance changes to content or authority initiatives.