We ran 100,000 distinct prompts across both ChatGPT and Perplexity and compared the citation sources. The result was not what we expected.
The citation overlap is smaller than you'd think
For the same prompt, ChatGPT and Perplexity cite the same source only 31% of the time. That means for any given query, there's a 69% chance they're drawing from completely different pages. The era of one-size-fits-all content strategy is over.
Platform-specific citation preferences
- ChatGPT favors: large news publishers, Wikipedia, established brand .com pages, and government/institutional sources
- Perplexity favors: specialized industry publications, recent news (< 30 days), expert blog content, and research reports
- Claude favors: long-form analysis, academic-adjacent content, and detailed technical explanations
- Google AI Overviews favors: highly-ranked Google search results, structured data pages, and Google-integrated sources
Strategic implications
This data argues for a platform-differentiated content strategy. Brands that create a single piece of content and hope it works across all platforms are leaving significant visibility on the table. Building platform-specific content variations — same topic, different style and structure — consistently outperforms single-version approaches in our testing.