Source Analysis
Source analysis is the practice of collecting every page and domain AI engines cite when answering a set of buying questions, then ranking them per engine to find which directories, review sites, and websites drive who gets recommended.
Also known as: Citation analysis, Cited-source analysis
Source analysis takes the raw list of cited sources from hundreds or thousands of AI answers and turns it into something you can act on. Which domains show up most? Which specific pages? Does ChatGPT lean on Yelp while Perplexity leans on Reddit? Is your own site anywhere in the list? It's the difference between noticing one answer linked Angi and knowing that Angi shows up in most of Copilot's answers for your trade in your city.
It has to be done per engine, because the engines don't read the same web. Profound found ChatGPT and Perplexity overlap on only 11% of cited domains, with 37.4% of citations ChatGPT-exclusive and 51.6% Perplexity-exclusive. Google AI Overviews and Copilot overlap just 6%. A source that dominates one engine can be nearly absent from another, so a blended source list would tell you to fix the wrong things.
Source analysis also catches shifts over time. Profound measured Reddit at about 46.7% of Perplexity's citations on commercial queries in 2025, as a share among its top cited domains, and that figure reportedly dropped to around 24% by January 2026. Yelp disclosed a data-licensing deal with OpenAI in February 2026. If you did your source analysis once in 2025 and built a plan around it, parts of that plan are already stale.
For a dentist or contractor, the output is a short, specific to-do list. If Healthgrades is cited in most answers about your specialty and your profile there is incomplete, complete it. If competitors are cited from their own pricing pages and you don't have one, write one. If Reddit threads about your city keep appearing and none mention you, that's a reputation gap, not a website gap. Source analysis is what tells you which kind of problem you have.
Tibly runs source analysis continuously on the answers it tracks across seven engines, ranking the cited domains and pages for your prompts and showing which ones cite you versus your competitors. Where the fix is a page on your own site, Tibly drafts it and publishes to your Tibly-hosted site; where it's a listing, it tells you which one.
Keep exploring the glossary
- Cited SourceA cited source is a specific web page or domain, such as a Yelp listing, a Reddit thread, or a business's own service page, that an AI engine references as the basis for its answer, showing where it learned it.
- Citation (AI answer)In an AI answer, a citation is a source the engine references or links to while generating its response, and the number of sources cited per answer varies widely by engine, from about 2.5 for Copilot up to about 7.7 for Google AI Overviews.
- Competitor Benchmark (AI answers)A competitor benchmark in AI answers is a side-by-side comparison of how often each business in a market gets named for the same buying questions on the same engines, revealing who holds the slots you don't and why.
- Aggregator (directory site)An aggregator is a third-party directory or review site such as Yelp, Angi, Thumbtack, Healthgrades, or Avvo that collects listings and reviews for many businesses in a category, and AI engines frequently cite them when recommending local businesses.
Common questions about Source Analysis
- Why do I need source analysis per engine instead of one combined list?
- Because the engines cite different sources. Profound found ChatGPT and Perplexity overlap on only 11% of cited domains, and AI Overviews and Copilot on just 6%. A combined list averages away the differences and points you at sources that matter for an engine you may not care about.
- How often should source analysis be refreshed?
- Continuously, or at least monthly. Reddit's share among Perplexity's top cited domains reportedly went from about 46.7% to around 24% within a year, per Profound, and licensing deals like Yelp's with OpenAI in 2026 shift what engines read. A one-time audit ages fast.
- What should I do once I know which sources AI engines cite?
- Sort them into two piles: sources you control and sources you don't. Complete and correct your own listings and pages in the first pile. For the second, like Reddit threads or news mentions, focus on the reputation work that earns mentions there. Then keep tracking to see whether the engines' citations move toward you.