Methodology

How VisibilityIndex works

VisibilityIndex measures two things: whether your site is structurally ready to be understood by AI systems, and whether supported AI systems actually include your business in recommendation-style outputs.

Run free scan Open app

1) What we measure

The free scan and the full audit are related, but they answer different questions.

Free scan

Structural readiness

The free scan checks identity clarity, service clarity, contact and location signals, structured data, trust signals, and basic extractability across a bounded set of pages.

Full AVI audit

Measured AI visibility

The full audit runs standardized prompt packs across supported AI systems and records whether your business appears, how well it appears, and how broad that visibility is across enabled surfaces.

2) Attribution

A result only counts when it can be tied back to the selected business.

Hard match

Domain or website evidence

The strongest positive case is a direct website, domain, or citation match to the selected business.

Soft match

Name + location + context

If a direct domain match is missing, VisibilityIndex can still count a result when the surrounding identity context strongly aligns and does not conflict with a competitor. Ambiguous mentions do not count.

3) Score logic

The score starts from measured evidence, then adjusts for breadth across enabled surfaces.

Formula

Step 1 — Per platform

channel score = frequency × 0.5 + rank × 0.5

Frequency: how often you appeared across 15 searches.
Rank: where you placed when you did appear.

Step 2 — Weighted average

primary = avg(channel score × market weight) × 100

ChatGPT 55% · Gemini 25% · Perplexity 12% · Claude 8%

Step 3 — Final score

final = (primary × coverage) - site penalty

Coverage: 1/3 platforms = ×0.50 · 2/3 = ×0.77 · 3/3 = ×1.00
Site penalty: up to 12 points for missing trust signals.

Interpretation

Broad inclusion beats a one-channel hit

A single isolated mention can help, but visibility across multiple enabled surfaces is treated as stronger than a one-surface spike. Readiness diagnostics inform fixes, but they do not replace actual inclusion.

4) Confidence

Confidence reflects measurement quality, not a promise of outcome.

What it means

Run quality

Confidence reflects attribution quality, channel measurement completeness, and how clean the evidence was inside the run.

What it does not mean

Not a business guarantee

Confidence is about measurement quality. It is not a guarantee of rankings, traffic, calls, leads, or revenue.

5) What the free scan checks

The readiness scan is deterministic and focused on a small set of high-signal pages.

Core checks

Business name, service category, and service area clarity
Phone, address, hours, and contact signals where applicable
Services, FAQ, pricing, reviews, and credentials signals
Organization or LocalBusiness structured data where applicable
Basic heading, list, and content-structure diagnostics

Trust checks

VisibilityIndex also checks a small set of trust and corroboration signals such as obvious crawler blocking, review-platform presence, and optional helper files like /llms.txt. These are supporting signals, not guarantees of recommendation.

6) Limits

Clear boundaries keep the methodology honest.

Important limits

VisibilityIndex does not guarantee traffic, rankings, calls, or revenue.
AI outputs can vary over time, by platform, by account state, and by prompt wording.
A strong readiness scan does not guarantee recommendation inclusion.
Optional files like /llms.txt are helper signals, not an AI passport.

Plain-English summary

The free scan checks whether your site is structurally ready to be understood. The full audit checks whether supported AI systems are actually choosing to include your business.