How each engine decides what to cite
In our previous article on GEO we explained the general logic: live search vs. trained knowledge. What the latest data shows is that, within live search, each engine has very different preferences when choosing what to cite.
| Engine | What it prioritizes | Average citations per response |
|---|---|---|
| ChatGPT | Domain authority; Wikipedia is ~48% of its sources | ~10 citations |
| Perplexity | Freshness (content <30 days gets cited 3.2x more) and Reddit (~47%) | ~22 citations |
| Gemini / AI Overviews | Classic SEO signals integrated with Google's index | Variable, tied to organic results |
The most important data point for your strategy: only 11% of domains cited by ChatGPT are also cited by Perplexity. These are almost entirely different source audiences.
What this means in practice: there's no single "GEO strategy." If your content is optimized only for domain authority (which favors ChatGPT), Perplexity is likely to barely cite you — and vice versa if you publish a lot but without backlinks or institutional presence.
How to win on ChatGPT
- Domain authority. Backlinks from recognised sites still carry a lot of weight — unlike what happens on Perplexity.
- Neutral, encyclopedic tone. The bias toward Wikipedia isn't a coincidence: content that explains rather than sells is more likely to get cited.
- Structured markup (Schema.org). Helps the model extract concrete data without ambiguity.
- Presence in reference sources: recognised industry directories, press, institutional pages.
How to win on Perplexity
- Freshness above everything. Perplexity searches live on every query; content published or updated in the last 30 days is 3.2x more likely to be cited.
- Presence on Reddit and forums. Almost half of its citations come from there — genuinely participating in relevant industry communities carries real weight.
- Publish on a consistent cadence. A blog updated weekly (like this one) has a structural advantage over one that publishes once a quarter.
- Long, in-depth content. Unlike ChatGPT, here length correlates positively with citation likelihood.
How to win on Gemini and Google AI Overviews
- Classic technical SEO still matters — these systems are integrated with Google's traditional search index, unlike ChatGPT and Perplexity.
- "Featured snippet" format: direct, short answers right below each subheading, easy to extract.
- Structured data (Schema.org) and good Core Web Vitals — the same technical SEO fundamentals you already know still apply here, more than on the other two engines.
Conclusion: prioritise based on where your audience searches
You don't need to win on all three engines at once. If your typical customer uses Google to find suppliers, prioritise classic technical SEO. If your industry lives in technical or niche communities, Reddit and content freshness on Perplexity carry more weight. If you want to be seen as a neutral, authoritative reference, bet on backlinks and encyclopedic-style content aimed at ChatGPT.
What is universal: publishing frequently, having structured markup, and being present in the sources each engine consults. At Dataverse Solutions we applied exactly this checklist on our own site — we cover it in detail in our own GEO audit case study.