Measurement, not promises · ChatGPT + Perplexity · tracked monthly on our own products since May 2026
Every AI answer has ~10 slots for sources.
Whoever fills them gets the buyer. Do you hold any of yours?
When your buyers ask ChatGPT what to use, its answer cites a shelf of sources. That shelf is the new page one. We measure who holds every slot in your niche, then map the ones you can actually take. We ran the same playbook on our own two products. One reached 9% of its niche's AI answers — the most-cited domain in it; the other stayed at 0%. The difference was measurable before we wrote a single article.
best screen recorder for freelancers
THE ANSWER CITES 10 SOURCES
These are the answer slots. Whoever fills them gets the buyer — and we measure who fills yours.
FIG. 1 — SHARE OF ANSWER SLOTS · 5 MONTHLY RUNS · 20 BUYER QUERIES · 2 ENGINES
Two products. Same playbook. Opposite endings.
We ran this on ourselves before selling it to anyone. The winner was predictable on day one.
THE DELIVERABLE — WHAT $300 BUYS
Not a dashboard. A verdict, a map, and a route.
Delivered in 48 hours as a written report plus a 30-minute readout. See a real sample report →
- 01Your answer-slot share Across ~20 buyer-intent questions in ChatGPT and Perplexity — the number almost nobody knows about themselves.
- 02Competitor shares The same measurement for your top 5 rivals. Who is actually winning the answers — often not who you think.
- 03Slot-holder analysis, query by query Who holds every slot you're missing: indie blogs, Reddit threads, review sites, or entrenched incumbents.
- 04The where-you-can-win map Head terms may be closed. Narrower intents almost never are — even our losing niche had open doors. You get the map of yours.
- 05Prioritized intent hit-list The specific queries to go after first, in order of winnability.
- 06Why your pages aren't cited A teardown of 1–2 of your pages against the pages engines actually cite for the same question.
- 07The recommended play Content, community, or listings — dictated by who holds your slots, not by what we'd like to sell you.
THE READING — WHO HOLDS YOUR SLOTS DECIDES THE STRATEGY
Three doors. Your audit says which one is open.
The content play
Outpublish them with an accurate, interlinked topic cluster. This is the route we took ourselves.
→ our 3% → 9% caseThe community play
Engines keep citing the same threads. Be genuinely present in them instead of publishing into a void.
→ blogging can't win here — we'll say soThe listings play
When G2 and Capterra own the answers, the profile is the battleground — win the listing, not the blog.
→ cheaper than a content retainerCLAIMS DISCIPLINE — WHAT WE WILL NEVER TELL YOU
Struck from our vocabulary.
Nobody can. The two engines disagree with each other 36% of the time on the same page. Anyone promising placements is lying to you.
In half of our own citation events, engines picked a different page of ours than the one built for that query. You grow a cluster's share, not a page's rank.
We own a 0% result and publish it. Some niches are closed to content — if yours is, we'll prescribe the cheaper play instead of selling you articles.
PRICING — START WITH THE MEASUREMENT
Aim the real money for $300.
- Full 7-part report (above)
- ~20 queries × 2 engines, raw data attached
- 30-min readout call
- Agencies: white-label per client domain
- Same measurement, re-run monthly
- Trend report — the signal is the trajectory
- Alerts when you gain or lose slots
- Publish-ready, fact-verified articles
- Targeting your audited open intents
- Every competitor claim verified by hand
FAIR QUESTIONS — ASKED BY PEOPLE WHO ALMOST DIDN'T BUY
Can't I just ask ChatGPT about my product myself?
You can, once. The measurement is 20 buyer-intent queries (chosen to match how buyers actually ask), across two engines, repeated monthly, with the ~36% engine disagreement filtered out. And a number without a benchmark is trivia: our before/after dataset is what tells you whether your 4% means "early and climbing" or "stuck." You're buying calibrated interpretation, not API calls.
How exactly do you measure it?
10–20 buyer-intent questions per niche, put to ChatGPT (gpt-5.5 with web search) and Perplexity (Sonar) through their APIs. Every source URL an answer cites counts as one slot; your share is how many of those slots are your domain. Because the same question can return different sources from one call to the next, we run every query 3× and average — a single sample is noise. We use the current models on purpose (older ones cite spam and miss obvious leaders). Today that's two engines; Google's AI Overviews and Gemini aren't in the measurement yet, and we'd rather say so than imply we cover everything.
We already use an AI-visibility dashboard.
Dashboards report numbers. This is a verdict and a route: which specific intents are open, who holds every slot you're missing, and which of the three plays fits — benchmarked against the only public same-operator win-and-loss dataset we know of. Keep the dashboard; it'll be useful for watching the plan work.
Is $300 worth it?
It's less than 1% of a typical quarterly GEO retainer, delivered before you sign one instead of after. One of our own two products would have burned five figures on an unwinnable niche — the measurement caught it. That's the purchase: spending $300 to aim the real money.
Who's behind this?
A solo operator — the same person who ran the case study on this page, on their own products, with their own money. Every number here comes from real tracker runs; the raw JSON is available on request.
Know your share before you spend.
One measurement now beats a retainer's worth of guessing. If your niche is closed to content, you'll know that too — and what to do instead.
Buy your audit — $300 Reply within 24h · report within 48h of kickoff