Glossary
LLM Visibility
Not a single mathematical standard — always ask which prompts and which models.
Closest Getllmspy equivalents: LLM-Score, Share of Voice, GPI.
Definition
“LLM Visibility” is industry shorthand for measuring whether ChatGPT-class assistants recommend, mention, or cite your brand when buyers ask category questions. Agencies may show dashboards, heatmaps, or share-of-voice bars. Because there is no universal formula, the useful question is always operational: Which prompt pack? Which geography? Which model endpoints? Without those, two vendors can report wildly different “visibility”.
Metric translation example
| Vendor term | Practical equivalent |
|---|---|
| Visibility index | blended score (like LLM-Score + SoV mix) |
| Recommendation share | mention share under recommendation prompts |
| Trust score | correctness + citation quality |
When translated to explicit components, stakeholder confusion drops sharply.
Translation graphic:
How it's computed
Common ingredients: (1) a fixed or organic prompt set, (2) multi-model execution, (3) mention detection + sentiment + hallucination flags, (4) aggregation into a headline score or rank. Some tools emphasise citations, others emphasise raw mentions — clarify before comparing numbers.
Vendor-comparison checklist
- Same prompt denominator?
- Same model mix?
- Same handling of missing model runs?
- Same quality criteria for mention correctness?
How it works in practice
Translate marketing speak into checks
- Ask for the exact prompt list or sampling method.
- Ask whether competitors are in the same denominator (SoV style) or not.
- Demand dated snapshots so you know if ChatGPT updated overnight.
Decision table
| Claim in sales call | Validation question |
|---|---|
| "Visibility up 30%" | Up against which fixed baseline? |
| "Best-in-class coverage" | Which exact model list and regions? |
| "AI-ready score" | Formula and weighting disclosed? |
How to read it
Treat “visibility up 20%” as suspicious unless the vendor discloses denominator changes, new models added to the bundle, or new prompts inserted into the pack.
How to operationalize "LLM visibility"
- Pick one internal KPI stack (e.g., LLM-Score + SoV + citation share).
- Lock prompt and model denominator for one quarter.
- Require evidence snapshots for every large delta.
- Separate methodology changes from performance changes in reporting.
Reporting maturity
| Level | What teams report |
|---|---|
| Basic | one blended number |
| Good | blended number + component metrics |
| Advanced | components + evidence quotes + methodology notes |
Advanced reporting is what makes visibility metrics board-safe.
When to use
- When evaluating a new monitoring vendor.
- When PR asks for an external buzzword mapped to internal KPIs.
- When you need a single slide title but still document the underlying metrics.