Prompt
I want to know whether our brand is getting mentioned more often in ChatGPT than in Perplexity, and whether the cited sources are our own content or competitors'. I don't need polished marketing language, just a practical way to measure it and report it monthly. What should I look for?
Latest observation
Here’s a practical way to measure this monthly without overcomplicating it.
1) Define exactly what you want to count
You want two things:
- Mention frequency
- How often your brand appears in answers from ChatGPT vs Perplexity
- Source mix
- When your brand is mentioned, are the cited sources:
- your own content
- competitors’ content
- third-party sources
- no citation / uncited
- When your brand is mentioned, are the cited sources:
You should track both, because a brand can be mentioned a lot but mainly through competitor or third-party pages.
2) Use a fixed prompt set
Create a stable set of prompts you’ll ask both tools every month.
Include prompt types like:
- Brand discovery
- “What are the best [category] tools for [use case]?”
- Comparison
- “How does [your brand] compare to [competitor]?”
- Problem/solution
- “What’s the best way to solve [pain point]?”
- Alternative search
- “What are alternatives to [competitor]?”
- Intent-specific
- “Which [category] platform is best for [industry/team size/budget]?”
Keep it consistent:
- same prompts every month
- same wording
- same language
- same region/settings if possible
- same number of runs per prompt
If you change prompts every month, the trend becomes noisy.
3) Decide what counts as a “mention”
A mention should be a clear reference to your brand in the response, such as:
- your brand name
- product names
- parent company name
- common misspellings or abbreviations
Track mentions at the response level and optionally at the sentence level.
Basic monthly metrics:
- % of prompts that mention your brand
- total brand mentions
- average mentions per response
- share of responses mentioning competitors instead
- brand mention rate in ChatGPT
- brand mention rate in Perplexity
4) Capture citations separately from mentions
For each response, record:
- Platform: ChatGPT or Perplexity
- Prompt ID
- Date
- Response text
- Brand mentioned? yes/no
- Competitor mentioned? yes/no
- Citations present? yes/no
- Citation URLs/domains
- Citation type:
- own domain
- competitor domain
- independent third-party
- forum/community
- no citation
This is important because:
- ChatGPT may mention your brand without citations depending on mode/settings
- Perplexity is citation-heavy, so you can analyze source mix more directly
5) Classify source ownership
For every cited URL, label it as one of these:
- Owned
- your website, blog, docs, help center, YouTube, LinkedIn, etc.
- Competitor-owned
- competitor websites and their content
- Neutral third-party
- review sites, news, analyst sites, communities, directories
- Unknown / needs review
Then aggregate monthly:
- % of citations that are owned
- % competitor-owned
- % third-party
- % uncited mentions
This tells you whether the models are “learning” from your content or mainly from others.
6) Compare ChatGPT vs Perplexity using simple ratios
Useful comparisons:
Mention share
- Brand mention rate in ChatGPT
- Brand mention rate in Perplexity
- Difference in percentage points
Example:
- ChatGPT: 42% of prompts mention your brand
- Perplexity: 28%
- Gap: +14 points in ChatGPT
Source ownership share
Among all cited sources in brand-mention responses:
- Owned content share
- Competitor content share
- Third-party share
Example:
- ChatGPT brand mentions: mostly uncited or cited to third-party pages
- Perplexity brand mentions: mostly cited to competitor articles That would suggest different visibility patterns.
7) Watch for “presence” and “position” separately
A brand can be:
- mentioned at all
- mentioned first
- recommended positively
- mentioned in a comparison
- mentioned as an alternative
Track a few simple fields:
- Appears in answer?
- Appears in top 3 recommendations?
- Position in list: first, second, etc.
- Sentiment/tone: positive, neutral, negative
- Context: recommended / compared / mentioned as option / not recommended
This helps you avoid confusing “mentioned once in passing” with “actually surfaced as a recommended brand.”
8) Keep an eye on query intent
Separate results by intent category, because brand visibility may differ a lot:
- informational
- commercial investigation
- comparison
- alternative-seeking
- how-to / educational
- brand-specific
Report monthly by intent, not just overall.
Example:
- strong in comparison prompts
- weak in “best tools” prompts
That’s more actionable than one blended number.
9) Use a human-coded sample if volume is small
If you only have, say, 20–50 prompts per month, manually reviewing the responses is enough.
If you have more, use a lightweight review sheet:
- prompt
- model
- response
- brand mentioned?
- sources owned/competitor/third-party?
- notes
Even a spreadsheet works.
If volume gets large, automate extraction of:
- brand mentions
- cited domains
- classification by domain
But still sample-check manually so you don’t trust bad parsing.
10) Suggested monthly report format
Keep it simple:
Executive summary
- ChatGPT mention rate: X%
- Perplexity mention rate: Y%
- Owned-source share: X%
- Competitor-source share: Y%
- Main change vs last month: up/down
Breakdown table
| Metric | ChatGPT | Perplexity | MoM change |
|---|---|---|---|
| Prompts with brand mention | |||
| Average mentions per response | |||
| % responses with citations | |||
| % citations from own content | |||
| % citations from competitors | |||
| % citations from third parties | |||
| % responses where brand is top recommendation |
By intent
- comparison prompts
- best-of prompts
- alternatives prompts
- problem/solution prompts
Notes
- any major citation source shifts
- new competitor pages showing up
- your content appearing/disappearing
- unusual prompts or model behavior
11) What “good” looks like
You’re probably looking for:
- your brand mentioned more often in one tool than the other
- your own content cited more often than competitors’
- your brand appearing in higher-intent prompts
- decreasing reliance on competitor pages over time
A practical target is:
- more brand mentions month over month
- rising share of owned citations
- fewer competitor citations for the same topics
12) Common pitfalls
Avoid:
- changing prompts every month
- mixing ChatGPT and Perplexity settings
- counting generic mentions of your category as brand mentions
- ignoring citation ownership
- assuming one result means broad visibility
- comparing different prompt sets across tools
Also note:
- ChatGPT and Perplexity may behave differently depending on browsing/search mode
- citation behavior can change over time
- one tool may cite more sources without being “better” for your brand
13) If you want the simplest workable setup
Use a spreadsheet with columns:
- Date
- Prompt ID
- Prompt text
- Platform
- Response text
- Brand mention? (Y/N)
- Competitor mention? (Y/N)
- Cited URLs
- Citation domain
- Citation owner type (owned / competitor / third-party / unknown)
- Placement (top / middle / bottom / not ranked)
- Tone (positive / neutral / negative)
Then each month:
- count mention rate by platform
- count citation ownership by platform
- compare month-over-month
- highlight the biggest source/domain changes
If you want, I can also give you:
- a ready-to-use spreadsheet template, or
- a monthly reporting framework with formulas/KPIs you can hand to an analyst.
Brands