Searchmetrics Visibility: what it measured and what replaced it
Searchmetrics' Visibility Score was an index of organic search presence. The modern equivalent for AI answers measures mentions and citations instead of ranked keywords.
Searchmetrics Visibility was, for many years, one of the standard ways to express a site's overall organic search presence as a single number. This article explains what the score measured, why index-style visibility scores work the way they do, and what the equivalent looks like now that answers are generated rather than ranked.
What the score measured
Searchmetrics' Visibility Score was an index rather than a real-world quantity. It combined a site's ranking positions across a large fixed keyword set, weighted by the estimated search volume of each keyword and by the expected click-through rate at each position. The output was a single number that rose when a site ranked better for more valuable keywords and fell when it did not.
Indexes of this kind — and several competitors published similar ones — had two useful properties and one significant limitation. They were excellent for spotting sudden movements, because an algorithm update or a technical failure showed up as a visible cliff. They were good for comparing sites within the same market, since everything was measured on the same keyword universe. But the absolute value meant nothing on its own: a score of 1,200 was only interpretable against the site's own history or a direct competitor's score on the same index.
Searchmetrics itself was acquired and its products were folded into other offerings, so the specific score is no longer the reference point it once was. The concept, however, is exactly the one the AI search era needs.
The modern equivalent
Generated answers have no positions and no click-through curve, so a position-weighted index cannot be computed. The equivalent construction for AI search replaces ranked keywords with tracked prompts, and position weighting with mention and citation weighting:
- The fixed keyword universe becomes a fixed prompt set, chosen to represent real buying questions.
- Ranking position becomes whether and where the brand is mentioned in the generated answer.
- Click-through weighting becomes prominence weighting, since a brand named first carries more influence than one listed last.
- Competitor comparison stays identical in spirit, and becomes more important, because absolute AI visibility scores are even less interpretable in isolation than search visibility indexes were.
What carries over from the old model
The lessons from a decade of visibility indexes apply directly.
- Keep the measured universe fixed, or the series becomes meaningless.
- Treat the absolute number as arbitrary and the trend as the signal.
- Always report against a competitor set measured the same way.
- Investigate cliffs immediately, since sudden drops usually indicate a technical failure rather than a market shift.
- Never present the index alone to a leadership audience without the underlying detail, because a single composite invites misinterpretation.
What does not carry over
The biggest change is variance. Rankings were stable enough that a daily index made sense. Generated answers differ between runs, so any AI visibility index has to be built on repeated executions and smoothed, or it will show movement that is purely noise. Anyone presenting an AI visibility index without disclosing the repetition count is publishing a noisier number than they realise.
Common questions about visibility indexes
- Is Searchmetrics still available? Searchmetrics was acquired and its products were folded into other offerings, so the specific Visibility Score is no longer the industry reference point it once was. The concept, however, survives in every visibility index still published.
- What replaced the Searchmetrics Visibility Score? For organic search, the equivalent indexes published by other SEO suites. For AI answers, the closest construction is a mention-rate index computed over a fixed prompt set with competitor overlay.
- Can I compare an old visibility score to a new one? Only within the same index and the same measured universe. Indexes are internally consistent and mutually meaningless, which was true of search visibility scores and is even more true of AI visibility scores.
- Why did index scores have no units? Because they were composites of position, volume, and expected click-through, deliberately scaled to be comparable over time rather than to represent a real-world quantity. AI visibility percentages are at least interpretable as a rate, which is an improvement.
- What is the main difference in the AI era? Variance. Rankings were stable enough for a daily index. Generated answers differ between runs, so any AI visibility index must be built on repeated executions and smoothed, or it reports noise.
- Should I still use a single composite score? For an executive summary, with the detail behind it. Never as the only number, and never without the prompt set and competitor set alongside.
Tip: if your team is used to reading a search visibility index, introduce AI visibility using the same chart shape and the same rules — fixed universe, competitor overlay, trend not level. The familiarity does more for adoption than any amount of explanation about how the underlying engines differ.
Still stuck? We typically reply within 1 business day.
Contact support