One ChatGPT check missed a brand it names in at least half its answers 23.4% of the time in our tracking. So how to track brand mentions in ChatGPT starts with checking more than once.
Most guides tell you to re-run ChatGPT checks "multiple times" or "3 times minimum", with no data behind the number.
We tested how many checks it takes. We tracked 23 questions (prompts) in ChatGPT, Gemini and Google AI Overviews over 37 run dates, from July 9 to October 18, 2026, and recorded 2,221 answers.
A single check is a sample, not a verdict.
The short version:
One ChatGPT check missed a brand it names in at least half its answers 23.4% of the time (911 of 3,897 checks).
Three checks in a row all missed it only 5.5% of the time (200 of 3,657); five checks, 2.2% (74 of 3,417).
53.0% of ChatGPT's brand mentions (3,361 of 6,347) went to brands it named in fewer than half its answers.
Brand Mentions vs Citations in ChatGPT: What You Are Tracking
A mention is when the answer text names your brand. A citation is when the answer links one of your websites as a source.
Our tracking logs the two in separate fields, because either can happen without the other. Gemini cited 3.54 websites per answer and Google AI Overviews 6.54 (the 2026-10-15 check excluded).
ChatGPT cited 4.39, but its citation volume shifted in October, so treat that average as a rough guide. Here is how many websites each AI cited per answer:
Our tracking, 23 questions. ChatGPT's figure is a rough guide because its citation volume shifted in October.
What this means: a mentions-only log tells you half the story.
Before You Start: Will One ChatGPT Check Be Enough for Your Brand?
One check works only for brands an AI names nearly every time. Those are rare.
Taking each brand on each question as one case, ChatGPT named 1,182 cases at least once and only 6 every time. Across all three AIs, it was 7 of 3,659.
For everyone else, one check is a weak read. ChatGPT named 553 cases on just one of 37 dates.
The typical case showed up in 6.1% of ChatGPT's answers to its question.
Here is how rare the steady brands were next to the one-off ones:
ChatGPT, across our 23 questions: steady brands against one-date brands.
If your brand is outside that steady handful, plan on repeated checks.
One ChatGPT Check Missed Brands It Usually Names 23.4% of the Time
We call a brand a regular on a question when ChatGPT named it in at least half its answers to that question. Of ChatGPT's 1,182 cases, 120 were regulars.
a) Misses: it leaves out brands it names most of the time
For those 120 regulars, a single ChatGPT check left the brand out 23.4% of the time (911 of 3,897 checks). A one-check report would log each of those as a loss.
b) One-off mentions: it shows brands it rarely names
For brands ChatGPT names in fewer than half its answers, one check still showed the brand 9.7% of the time (3,361 of 34,650 checks). Those occasional brands supplied 53.0% of all ChatGPT mentions (3,361 of 6,347).
Both errors side by side:
Misses hit the brands ChatGPT names most; one-off mentions come from brands it rarely names.
Every brand came and went at least once
All 457 brands ChatGPT named at least once were named on one check and missing on the next, or the reverse, on at least one question.
In our tracking, a brand ChatGPT had named before appeared or disappeared from one check to the next 14.7% of the time. Across all three AIs, it was 15.7% (17,974 of 114,294).
Gemini and Google AI Overviews missed even more
Get Your Clients Cited Where Competitors Already Are
Our tracking, 23 questions, 37 dates. Misses fall fast with each extra check.
How Many ChatGPT Checks You Need Before You Call a Drop
The rule our data supports: wait for 3 checks in a row before you call a brand missing, and 5 if the call goes to a client or a boss.
For regular brands, all 3 checks in a row missed the brand 5.5% of the time on ChatGPT, 6.0% on Gemini and 7.4% on Google AI Overviews. At 5 checks, that fell to 2.2%, 1.8% and 2.0%:
Each extra check cuts the miss rate on all three AIs.
The likely reason is that each AI looks up sources and writes a fresh answer each time, so brands at the edge of its picks come and go between checks. Our data cannot test that cause: we see which brands appear in each answer, not how the AI assembled it.
How to Track Brand Mentions in ChatGPT in a Spreadsheet, Step by Step
A spreadsheet is enough, using the same columns as our own tracking. The six steps as one checklist:
The manual method, with the columns our tracking uses.
#1 Pick your questions and keep them the same
Write category questions the way your buyers ask them, then keep the list fixed. Ours used 23, such as "Which AEO tool is best for agencies" and "Best software to track AI answer rankings".
#2 Log one row for each brand in each answer
Use these columns:
date
question
AI
answered (yes or no)
brand
mentioned (1 or 0)
websites cited
One answer becomes many rows, since one ChatGPT answer names many brands at once.
#3 Re-check on a set schedule
Ours ran mostly every 3 days. Pick a schedule your team can hold.
#4 Mark "no answer" apart from "not named"
When an AI returns no answer, log the check as missing, not as "not named." We skip those checks; scoring them as misses would inflate every miss rate above.
#5 Treat different spellings of a brand as one brand
Brand names vary in case, spacing and suffixes such as "AI" or ".io". We matched them with a hand-checked list of 106 extra spellings.
(Our own first brand list had one vendor twice, because one check wrote its name with an "AI" suffix and the next left it off.)
#6 Wait for 3 misses in a row before you report a drop
Flag a drop only after 3 misses in a row on the same question and AI, and take it to a client only after 5. One blank row is not a loss.
What to Ask a ChatGPT Tracking Tool Before You Pay for One
A tracker saves the logging if it checks the way the numbers require. Ask any tool three things:
How often does it re-check each question? One ChatGPT check missed a regular brand 23.4% of the time.
Does it keep history for each AI? Even 3 checks in a row all missed on ChatGPT 5.5% of the time.
Does it separate mentions from citations? A mentions-only view leaves out the source side.
Get Your Clients Cited Where Competitors Already Are
SurgeGraph is one option here: it's an AEO and AI-visibility platform whose AI Visibility dashboard tracks mentions and citations over time, with the number of engines tracked set by plan. Its Prompt Performance page shows how many questions sat at low, medium or high visibility on each check date, with brands mentioned per question:
SurgeGraph Prompt Performance chart: how many questions sat in each visibility band (Low, Medium, High) on each check date.
What to Do This Week to Track ChatGPT Mentions
Five moves, each tied to a number above:
Don't report a drop from one ChatGPT check. One check missed a regular brand 23.4% of the time.
Wait for 3 checks in a row before calling a loss, and 5 for client calls. Three checks still all missed 5.5% of the time; five, 2.2%.
Treat a single "named" result as a lead. 53.0% of ChatGPT mentions went to brands it names in fewer than half its answers.
Don't trust one check on Gemini or Google AI Overviews. Their one-check miss rates were 29.1% and 30.7%.
Log citations apart from mentions. Google AI Overviews cited 6.54 websites per answer and Gemini 3.54.
The five moves on one card:
The five moves and the numbers behind them.
When this doesn't apply: brands ChatGPT names on almost every check (6 of 1,182 in our tracking), and branded questions, which our 23 category questions did not test.
What These ChatGPT Tracking Numbers Can't Tell You
What our tracking covered, at a glance:
The scope behind every number in this post.
Sample: 23 category questions about AEO tools, 37 dates, 3 AIs, 2,221 answers, one project, US English. With no branded questions, we can't say how brand-name questions behave.
Spacing: checks were mostly 3 days apart and up to 9, with no same-day repeats, so a string of checks mixes check-to-check swings with change over time.
Labels: we set the "regular" label and tested it on the same 37 dates.
Collection: answers from our tracking, not a logged-in user's chat. Google AI Overviews checks with no answer are skipped.
Position: we scored whether a brand was named, not where.
Citations: the 2026-10-15 Google AI Overviews check was a tracking gap and is excluded from its citation figures. ChatGPT's citation volume shifted in October, so read its 4.39 as a rough guide.
Wrapping Up
Pick 5 to 10 questions your buyers ask, check them in ChatGPT on a fixed schedule, and log every brand in each answer. Report nothing as lost until 3 checks in a row agree.
Q: How many checks do you need to know how often ChatGPT names your brand?
A: More than 5: for regular brands on ChatGPT, 5 checks in a row still showed them in fewer than half the answers 16.4% of the time (562 of 3,417). At 11 checks, it was 10.8% (290 of 2,697), against 17.2% on Gemini (463 of 2,688) and 15.8% on Google AI Overviews (303 of 1,921).
Q: How long does it take to confirm a drop in ChatGPT brand mentions?
A: About 6 days for 3 checks and about 12 days for 5, if you check every 3 days. In our tracking, 1,738 of 2,152 gaps between checks (80.8%) were 3 days; those spans are our arithmetic, not a measured result.
Q: Does every ChatGPT answer cite a source?
A: No. 102 of 750 ChatGPT answers in our tracking cited no website at all, against 70 of 750 on Gemini and 6 of 698 on Google AI Overviews.
Q: How many brands does one ChatGPT answer name?
A: ChatGPT named 8.46 brands per answer on average across 750 answers in our tracking. Gemini named 9.61 and Google AI Overviews 7.16.
Q: Do these rules hold for Perplexity or Google AI Mode?
A: We can't say from this data, because our tracking covered ChatGPT, Gemini and Google AI Overviews only. Test the 3-check habit on your own questions first.
NOTE:
This article was written by SurgeGraph's AI writer using a custom Author Synthesis voice profile, then reviewed by a human editor. Author Synthesis learns a brand's voice from real writing samples so generated content matches its tone, cadence, and phrasing.
Eli Taylor
Digital Marketer at SurgeGraph
Eli lives and breathes digital marketing and AI. He always seeks new ways to combine AI with marketing strategies for more effective and efficient campaign executions. When he’s not tinkering with AI tools, Eli spends his free time playing games on his computer.
