Somewhere right now, a GEO agency deck is promising a client "+60–90% AI traffic in six months." Ask a simple question of that number: measured how?
There is no public benchmark linking an audit-score improvement to citation frequency, traffic or revenue. AI engines do not publish their selection logic. Independent evaluations of content-side GEO tactics find that most produce no statistically significant effect — the academic C-SEO Bench work tested dozens of method-and-domain combinations and found significant positive results in only a handful. An industry this young, measuring a target this unstable, cannot honestly promise percentage outcomes. So the projections you see in sales decks are not conservative estimates. They are decoration.
Our rule
When we rebuilt Citeify's methodology in August 2026, we made it a written, enforced rule in the report pipeline:
- No projected traffic percentages.
- No monetary values.
- No guaranteed citations or rankings.
Every impact statement in a Citeify report has to trace to something the audit observed. Not "fixing this is worth £4,000 a month" — but "OAI-SearchBot cannot currently retrieve your pricing page, so ChatGPT cannot cite it when buyers ask about products like yours." Cause and effect you can verify yourself, today, for free.
Doesn't that make the reports weaker?
Commercially, invented numbers are seductive — they close deals. But they also detonate on contact with reality: three months later the promised lift has not materialised, and the client correctly concludes the whole audit was theatre. The one asset an audit product has is being believed.
We would rather tell you the uncomfortable version: what is measurably broken, what is measurably true of AI engines' behaviour, and what we measured about your visibility — we publish our own score, which is currently a bracing 41/100, so we take our own medicine.
The full list of claims we refuse to make — and everything we do measure instead — is on our methodology page, version-stamped on every report we produce.