How GeoReputation calculates visibility scores

A GeoReputation visibility score is a number between 0 and 1000 that summarizes how well an AI assistant represents a brand across a set of prompts. This doc walks through the formula, the weights, and the decisions we made about edge cases.

If the math here ever drifts from the live implementation, the doc is wrong. The Last Reviewed date at the top of the page is the freshness signal.

The shape of the underlying answer

A scoring run executes a configured set of prompts against a configured set of AI models. Each prompt + model combination produces one answer record with three signals:

• mentioned — was the brand named in the response?
• rank — if mentioned, what position (1st, 2nd, 3rd, etc.) did the brand appear at in the recommended list?
• quality — a 0-1 score reflecting how prominently and favorably the brand was discussed. Four things go into it: how early in the answer the brand appears, whether it was presented as the primary recommendation or one option among several, how many competitors were named alongside it, and a fixed credit for being mentioned at all. That result is then adjusted for the tone of the mention (see Sentiment, below).

Note that the position of a brand within a ranked list is not currently part of the quality score. We measure where the mention falls in the text of the answer, which is a close but not identical thing. Ranked-list position is implemented and switched off while we validate it.

Each answer is actually a roll-up of three samples (we run every prompt against every model three times in a row) so transient model variability does not dominate the score. The sample-level normalization produces a mention rate between 0 and 1 instead of a single bit.

Run-level formula

A single run score combines two components: coverage and quality.

Coverage is weighted by how consistently the three samples of an answer agreed with each other. An answer whose samples all said the same thing counts slightly more than one where they disagreed. The weight never drops below 0.75, so disagreement dampens a signal rather than erasing it.

A separate per-prompt reliability factor, ranging from 0.5 to 1.5, exists in the system and is applied to an internal run-level measure. It is not applied to the per-prompt scores or the brand score shown in the product.

Why 70/30

Coverage is the larger weight because being mentioned at all is the threshold question: if an AI assistant does not name you, no amount of quality on the mentions that do happen will fix that. Quality is the secondary signal that distinguishes a brand mentioned in passing from one given a confident recommendation.

We landed on 70/30 after testing 80/20 and 60/40 against the brands we had data for. 70/30 produced the rank order that matched our internal taste for "this brand has stronger visibility than that one" in the cases we hand-checked.

Stability weighting

Each answer carries a stability score from 0 to 1 based on whether the three samples agreed with each other. The weight applied to that answer is:

weight = 0.75 + (0.25 × stability_score)

A perfectly stable answer (same outcome across all three samples) gets weight 1.0. A perfectly volatile answer (every sample different) gets weight 0.75. The discount is intentionally gentle: volatile prompts still count, just less.

Brand score (across runs)

A brand has its own score, which is the mean of its per-prompt scores. If you ran 12 prompts and got per-prompt scores of [800, 750, 600, 900, 700, 850, 500, 600, 700, 800, 750, 900], the brand score is round(8200 / 12) = 683.

Each prompt contributes the score from the last run that scored that prompt, which is not always the same run for every prompt. If a prompt is monitored weekly and another daily, the brand score blends a fresh number with an older one. This keeps a prompt that was not part of the most recent run from silently dropping out of the average, at the cost of the brand score being a blend across scan dates rather than a single snapshot.

A few important properties:

• Only prompts in your active scoring set contribute. Prompts you have archived or marked exploratory are excluded.
• Only models in the scoring set (your configured AI models) count. Adding a new model does not retroactively change historical scores; we snapshot which models were in the scoring set at the time an answer was produced and filter on that.
• Era-guarded: when the scoring methodology changes meaningfully (which has happened twice since launch), historical runs are tagged with the era they ran under. The current brand score only averages runs from the current era so the number means the same thing across the chart.

From score to tier

The 0-1000 score maps to one of six tier labels. These labels are the canonical vocabulary used across the UI, exports, and AI-generated narratives.

The bands are intentionally wide. We chose round 200-point intervals over finer-grained bands so a 5-point swing on a noisy run does not flip a brand between tiers. The cost is that two brands inside the same band can have meaningfully different scores; tier is for narrative, not for fine comparison.

How sentiment affects the score

Coverage does not care about tone. Being named is being named, whether the mention is glowing or lukewarm.

Quality does care. Once the four quality factors are combined, the result is adjusted for the classified tone of the mention: a positive mention is credited more, a neutral one is unchanged, and a negative one is roughly halved. When the tone cannot be classified with confidence, we apply a small discount rather than assuming the mention was favorable.

Because quality is capped at 1.0, a positive tone cannot lift a mention above the ceiling that prominence already set. In practice sentiment protects against a brand looking strong on the strength of mentions that were actually unfavorable.

What the score does NOT capture

Worth naming the limits in the same place as the formula:

• Citations. Whether an AI assistant cites your website as a source does not affect this score. Citations are tracked and reported separately. A page of yours can start being cited without the visibility score moving, and that is expected rather than a contradiction.
• Reasoning correctness. If an AI mentions the brand but says something factually wrong about it, that does not lower the score. Detecting hallucinated claims about a brand is a separate pipeline.
• Brand-mention context outside the answer. The score only reflects what the AI said in response to the prompt. Off-platform brand health (review sites, social, news) is tracked separately.

Edge cases worth knowing

Zero coverage

If no prompt in a run produced a mention, coverage is 0 and quality is undefined (we cannot average quality across zero mentions). The score is 0, which maps to the "Not Visible" tier.

Single-model runs

A run with only one model produces a valid score, but the score is much noisier because there is no cross-model averaging. We surface this in the UI as a confidence indicator on the run-detail page.

Fallback answers

When the live API call to an AI model fails, we sometimes fall back to a cached or retried answer. Runs where more than 10% of answers came from fallback get a "low confidence" badge in the dashboard; the score is still produced but should be interpreted with the disclosure.

Questions about the math

If you are evaluating GeoReputation and want to dig deeper than this doc on any specific decision, reach out and we will walk you through it. The formula is not secret. The same goes for customers who want to understand a specific score on their own dashboard: ping us and we will explain how that number was produced.