- 1.Why Testing Your ChatGPT Visibility Matters
- 2.Before You Start: Understanding What You Are Measuring
- 3.Step 1: Build Your Query Test Set
- 4.Step 2: Set Up Your Testing Environment
- 5.Step 3: Run Your Queries and Record Results
- 6.Step 4: Run the Same Query Set for Your Top Competitors
- 7.Step 5: Test Across Perplexity and Google AI Overviews
- 8.Step 6: Analyze Your Results
- 9.Step 7: Build Your Action Plan from the Data
- 10.Limitations of Manual Testing
- 11.Get a Professional Visibility Test
Why Testing Your ChatGPT Visibility Matters
If you don't know your current ChatGPT citation rate, you have no baseline to improve from and no way to know whether your GEO efforts are working. Testing your AI visibility is not complicated — but it requires a systematic approach that produces reliable data instead of anecdotal impressions. This guide walks you through a practical, no-cost visibility test that any contractor can run in an afternoon, using nothing more than a browser and a spreadsheet.
Before You Start: Understanding What You Are Measuring
You are measuring citation rate — the percentage of times ChatGPT mentions your business by name in response to relevant queries. A 0% citation rate means ChatGPT does not know you exist. A 50% citation rate means you appear roughly half the time. A 90%+ citation rate means you are the dominant recommendation in your query cluster. The goal is to establish your current baseline, identify what your top competitors' rates are, and use the gap between those numbers to build a prioritized action plan.
Step 1: Build Your Query Test Set
Create a list of 10 to 20 queries that represent how your ideal customers would ask about your services. Vary the phrasing, specificity, and conversational style. For a plumbing contractor in Dallas, your query set might include: 'best plumber in Dallas,' 'who should I call for a burst pipe in Dallas,' 'recommended plumbing company in North Dallas,' 'licensed plumber for water heater replacement Dallas,' 'how do I find a reliable plumber in Frisco TX,' and 'plumbing contractor reviews Dallas.' The variety is important — you want to test across the range of how real customers actually phrase their searches.
Step 2: Set Up Your Testing Environment
Each query must be run in a fresh ChatGPT session to avoid context contamination from previous queries. Open a new Incognito or Private Browsing window for each test. If you have a ChatGPT account, log out or use a session where no prior conversation history is present. For the initial test, run each query three to five times in fresh sessions, recording whether your business was mentioned each time. This gives you a preliminary citation rate estimate before committing to a full 50-sample run.
Step 3: Run Your Queries and Record Results
Set up a simple spreadsheet with your queries as rows and each test run as a column. For each session, paste your query into a fresh ChatGPT window (use ChatGPT 4o with web browsing enabled for the most current results), and record: whether your business was mentioned (yes/no), the position in which you appeared if mentioned (1st, 2nd, 3rd, etc.), which competitors were mentioned and in what position. After 5 runs per query, you have enough data to identify patterns even if the sample size is too small for statistical significance.
Step 4: Run the Same Query Set for Your Top Competitors
Identify your top 3 local competitors — the businesses you compete with most directly for the same customer segment. Run the same query set and ask ChatGPT specifically about them: 'what do customers say about [Competitor Name],' 'is [Competitor Name] a good roofing contractor in [city],' and 'how does [Competitor Name] compare to other roofers in [city].' Record how frequently each competitor appears in general query responses and how ChatGPT describes them when specifically asked. This competitive data is often more revealing than your own citation data because it shows you the standard you need to match.
Step 5: Test Across Perplexity and Google AI Overviews
ChatGPT is one platform — but AI-driven discovery happens across multiple systems. Run your core query set through Perplexity (perplexity.ai) and Google AI Overviews (search for your queries on Google and note when AI Overviews appear). Some contractors appear consistently on one platform and rarely on others. This platform-specific variation reveals whether your gaps are universal (you need more fundamental GEO work) or platform-specific (you are strong in one ecosystem but absent from another, which requires targeted signals for that platform's specific data sources).
Step 6: Analyze Your Results
Once you have your baseline data, calculate your overall citation rate across all queries and all platforms. Note the queries where you appear most consistently — these are your current strengths. Note the queries where you appear least consistently or not at all — these are your priority gaps. Compare your rates to your competitors' rates to calculate your competitive citation gap. If you are appearing in 20% of queries and your top competitor appears in 70%, that 50-point gap represents the magnitude of the GEO investment needed to reach parity.
Most contractors discover they are invisible on 60-80% of their relevant query variations. This is not a reflection of your actual quality — it is a reflection of how your online presence is currently structured for AI systems. It is fixable.
Step 7: Build Your Action Plan from the Data
Your test results will point to specific GEO actions. If you appear on zero queries, the foundational issues are likely entity normalization and schema markup. If you appear on some queries but not location-specific ones, your geographic content coverage is insufficient. If you appear on broad queries but not specific service queries, your content answer capsules need to be written for those specific service intents. Use the gap patterns in your data to prioritize your GEO investment — fix the highest-frequency gaps first.
Limitations of Manual Testing
Manual testing with 3 to 5 runs per query provides directional data but limited statistical confidence. To get reliable citation rates (margin of error under 10 percentage points), you need 50 or more fresh-session runs per query. Manual testing at that scale is time-intensive — a 20-query set at 50 runs each represents 1,000 total test sessions. Professional GEO services automate this sampling using API access to AI systems, which produces statistically reliable data at a fraction of the time cost.
Get a Professional Visibility Test
If you want reliable, statistically valid citation rate data across your full query set and all major AI platforms — benchmarked against your specific competitors — Market Disruptors can run a professional visibility audit for your contracting business. Book a free strategy call and we will show you exactly what your current AI visibility looks like, where your biggest gaps are, and what the fastest path to improved citation rates looks like for your market.
Kristina Shrider
National Growth Architect & Behavioral CMO
Kristina is the founder of Market Disruptors Agency and an independent AI marketing researcher. Her published work includes From Automation to Judgment (18 independent citations) and the MAD-M™ governance framework. The GEO methodology and CitationIQ™ measurement platform used across this research library are based on her original work.
View research profile →