How to Measure Share of Voice in ChatGPT and Perplexity
Measure share of voice in ChatGPT and Perplexity by scoring the same buyer prompts each month: count mention slots, split by brand, and track citation rate. No fabricated benchmarks required.
How to measure share of voice in ChatGPT and Perplexity: choose five to ten named competitors, lock twenty to forty buyer prompts, run each prompt once per product per month in fresh sessions, score whether each brand is mentioned or cited, then divide your brand's wins by the total scored slots to get mention share and citation share separately.
When the fix belongs on a live site, run the AI brand visibility checker and read our AI SEO for marketing agencies guide for the page types that vertical needs.
Classic SEO share of voice tools read rank trackers. AI assistants do not expose the same API. Peec AI and similar products describe share-of-voice reporting inside their own dashboards; this guide is the manual method behind that line on a slide. Use it to validate whether a paid tracker matches what your buyers actually see.
What AI share of voice means
Pick one definition and keep it:
- Mention share: prompts where the assistant names your brand in the answer body.
- Citation share: prompts where the assistant links or footnotes a URL on your domain.
Report ChatGPT and Perplexity separately. Combine them only after each column is stable, or you will smooth over a Perplexity footnote win with a ChatGPT miss.
This page does not replace the GEO tracking tools comparison when you need automated exports. It does replace guessing from one anecdotal chat.
1. Define the competitor set
List five to ten brands buyers actually shortlist, not every logo on a market map. Include:
- Direct substitutes at your price band.
- The incumbent your sales team loses to most often.
- The "good enough" free option that appears in evaluation threads.
Exclude vendors you never see in deals, even if they rank on Google. AI share of voice is a sales mirror, not a vanity leaderboard.
2. Build the prompt panel
Each row is one prompt ID and exact wording. Cover:
- Category picks ("best [category] for [audience]").
- Alternatives ("[your brand] vs [competitor]").
- Problem-led jobs ("how to [job] without [pain]").
Do not add prompts mid-month unless the panel was clearly wrong. Changing the denominator makes month-over-month charts lie.
3. Score each answer
For each prompt and product, record:
| Field | What to capture |
|---|---|
| Date | Day of the run |
| Product | ChatGPT or Perplexity |
| Brands mentioned | Ordered list as written |
| Your citation? | Yes or no for your domain |
| Winning URL | Path cited when you or a competitor footnotes |
When Perplexity footnotes a competitor blog post, copy that URL. That path is the page to beat. When ChatGPT names brands without links, mention share still counts; citation share stays zero.
4. Calculate share of voice
For each product and month:
- Mention share (you) = prompts where you are named ÷ total prompts.
- Mention share (competitor X) = prompts where X is named ÷ total prompts.
- Citation share (you) = prompts footnoting your domain ÷ total prompts.
Example with twenty prompts: you are named on six and cited on two. Mention share is 30%. Citation share is 10%. Report both numbers; do not average them into one hero metric.
Plot month over month for the panel only. A spike on one prompt is a clue, not a trend, until the next full run confirms it.
5. Act on the gaps
Low citation share with decent mention share means assistants know your name but quote someone else's page. Ship or rewrite the keeper URL for that prompt with a citable opening block. The comparison brief guide covers structure for evaluation prompts.
Low mention share on evaluation prompts means you are not in the shortlist at all. Check whether a comparison page exists, whether it matches SERP intent, and whether internal links point to it. The keyword gap analysis guide helps list missing page types.
Run the free AI brand visibility checker when you want a packaged read on visibility before you rewrite the panel. Use the free preview to see sample pages in your design. Orchory builds cite-ready pages on the prompts you approve, starting at $990/mo for ongoing production. Send your site when you want a first page on a call, or start with the AI brand visibility checker for a quick baseline on ChatGPT, Perplexity, and Google AI Overviews. Browse the free SEO tools hub for on-page checks while you ship.
