Why tracking ChatGPT citations matters now
A growing share of the buying journey for Seattle service businesses now happens inside ChatGPT, Perplexity, Gemini, and Claude. When a homeowner asks ChatGPT to recommend a plumber in Ballard, the answer they get is the new shortlist. If your business is not in that answer, the lead never reaches your website, your phone, or your analytics. Traditional SEO dashboards are blind to it.
Tracking ChatGPT citations is how you make the invisible visible — and how you prove that your GEO and AEO investment is moving the only number that matters: how often AI tools recommend you to your customers.
The two metrics that actually matter
1. Citation share
Citation share is the percentage of relevant AI answers in your category that mention your business by name or link. For a Capitol Hill dentist, that means: of the answers ChatGPT, Perplexity, Gemini, and Claude give to the twenty highest-intent questions in your service area, how many cite you? This is the new ranking. It replaces keyword position for AI search.
2. Referral traffic from LLMs
Referral traffic from LLMs is the actual visit count, sessions, and conversions that originate from ChatGPT, Perplexity, Gemini, Claude, and the LLM-powered overviews inside Google and Bing. Citation share proves you are in the answer; referral traffic proves the answer is sending real people to your site.
How to track citation share
Build a prompt set that matches real buyer intent
Start with the twenty to fifty questions your best customers actually ask before they hire you. Mix transactional ("best emergency plumber in Seattle"), comparative ("X vs Y in Capitol Hill"), and informational ("how much does X cost in Seattle"). The prompt set is the measurement instrument — if it is weak, the data is meaningless.
Run the prompts weekly across all four engines
Each week, run the full prompt set through ChatGPT, Perplexity, Gemini, and Claude. Log the response, the citations, and whether your domain appears. For small businesses, this can be done manually in a spreadsheet. At scale we automate it through the OpenAI, Perplexity, and Anthropic APIs and store responses in a database for trend analysis.
Calculate citation share per engine and overall
Citation share = (prompts that cite you) / (total prompts). Track it per engine — ChatGPT and Perplexity behave very differently — and as a blended number across all four. The trend matters more than any single week.
How to track LLM referral traffic
LLM referrals show up in Google Analytics 4 and most other analytics tools, but they are scattered across referrer hostnames and easy to miss. The hostnames to filter for include chat.openai.com, chatgpt.com, perplexity.ai, gemini.google.com, claude.ai, and copilot.microsoft.com.
- Build a GA4 exploration filtered to those referrer hostnames and save it as a permanent report.
- Add a UTM-style breakdown so you can see which pages on your site the AI engines are sending traffic to. Those are your citation magnets — double down on the patterns that worked.
- Track conversions, not just sessions. A handful of LLM-sourced visits that book a job is worth more than thousands of low-intent Google clicks.
- Watch for the AI Overview referrer inside Google Search Console performance reports — it is reported separately from the standard organic listing.
Closing the loop with revenue
Citation share and referral sessions are leading indicators. Revenue is the only one that pays. Tie LLM referral conversions to actual booked jobs in your CRM, and you can finally answer the question every owner asks us: "Is this AI search stuff actually working?" When citation share rises, referral traffic rises, and revenue from that referral source rises, you have a defensible attribution chain — and a case study every competitor will spend the next year trying to copy.
If your agency cannot show you citation share and LLM referral revenue every month, they are not doing GEO. They are just hoping.
Where Seattle businesses should start
- Pick your top twenty buyer questions and run them through ChatGPT and Perplexity this week. Note who gets cited. That is your baseline.
- Set up a GA4 referrer filter for the six LLM hostnames above. You probably already have traffic you are not counting.
- Pick one underperforming question and rewrite the matching page to answer it cleanly, with local Seattle context and clear structure. Re-run the prompt in two weeks.
- If you want this productionized — weekly tracking, dashboarding, citation outreach — that is what we built Aliens Digital to do.
GEO and AEO without measurement is faith-based marketing. Tracking ChatGPT citations is how it becomes a real channel — one that compounds month over month and finally gives Seattle service businesses a way to quantify AI visibility against the competition.