LoudPixel Docs
Reference

Reference — Citation Share Score formula

Exact formula, inputs, edge cases, and dollar interpretation for the LoudPixel Citation Share Score (0-100) — and how to read each band.

The Citation Share Score is LoudPixel's single headline metric for AI search visibility. It expresses, as a number from 0 to 100, the share of audited buyer-intent queries in which an AI engine cited your domain at least once.

Definition

Formal: Citation Share Score is the proportion of audited queries that produced at least one citation of the tracked domain across the configured set of AI engines, scaled to a 0-100 range.

Informal: Out of the questions buyers ask AI in your category, how many name-drop you.

Formula

citation_share_score = (queries_citing_you / total_queries_audited) * 100
InputTypeSourceNotes
queries_citing_youintper-engine answer parsecounted once per query even if cited by multiple engines
total_queries_auditedintaudit run configdefault 20 on free tier, up to 200 on paid
(output) scorefloatcomputedrounded to 1 decimal in UI, full precision in API

A query counts as "citing you" if your root domain appears in the answer text, the source link list, or both — see edge cases below for the precise matching rules.

Score bands

BandInterpretation
0-20AI engines almost never mention the tracked domain; competitors hold the category in AI answers.
21-50Intermittent presence, typically on branded or long-tail queries; most category queries miss the tracked domain.
51-80Strong presence on category queries; the tracked domain is a frequent but non-default answer.
81-100The tracked domain is the default answer in its category across most audited engines.

Edge cases

  • All engines time out. If every configured engine returns an error or timeout for the audit, queries_citing_you and total_queries_audited are both undefined; the score is reported as null, not 0. Do not interpret null as zero presence — re-run the audit before acting on it.
  • Mixed-case domain matching. Domain matching is case-insensitive. loudpixel.ai, LoudPixel.ai, and LOUDPIXEL.AI are counted as the same domain.
  • Subdomain handling. By default, blog.acme.com rolls up to acme.com and counts as a citation of the root. Enable strict subdomain in the audit settings to count each subdomain separately.
  • Citation in source link only. If the domain appears only in the engine's source/citation list and not in the answer prose, it still counts toward queries_citing_you.
  • Negative-sentiment citation. A citation that recommends against the tracked domain (e.g. "avoid acme.com") still counts toward queries_citing_you. Sentiment is tracked separately by the Citation Sentiment Score and is out of scope for this metric.

Dollar interpretation

The score is a leading indicator of AI-referred pipeline. To translate a score change between two audits into recovered monthly pipeline, use:

recovered_pipeline = (score_delta / 100) * category_avg_ai_referred_session_value * monthly_query_volume

Where score_delta is the percentage-point change between two audits (for example, 30 − 10 = 20), category_avg_ai_referred_session_value is the typical revenue per AI-referred session in your category, and monthly_query_volume is the monthly volume of cited buyer-intent queries surfacing the tracked domain.

Worked example: a score that moves from 10 to 30 (Δ20), with a category session value of $80 and 120 monthly cited buyer-intent queries, yields (20 / 100) * 80 * 120 ≈ $1,920/mo in recovered pipeline. Treat this as directional sizing, not a forecast — actual conversion depends on landing-page quality, intent match, and downstream funnel performance.

On this page