SpyderBot Documentation
Metrics Reference
Definitions below describe eligible conditions and display intent. Report-level rates use the production aggregation contract unless a metric is explicitly labeled as a raw population rate. Tooltips distinguish aggregated results from raw support counts.
Report scope metrics
| Metric | Unit | Data class | Definition |
|---|---|---|---|
| Active LLM Agents | Count | Observed operational metadata | Distinct planned agents with at least one terminal valid or invalid outcome versus planned agents |
| Prompt Set | Prompts | Observed configuration | Unique canonical prompts frozen for the report run |
| AI Systems | Systems | Observed configuration | Eligible AI-system policies included in the report run |
| Executions | Attempts | Observed | Completed or assigned prompt-to-system attempts under the run policy |
| Valid Responses | Responses | Observed | Completed responses that passed validation and are eligible for metrics |
| Prompt-system Coverage | Cells | Derived | Eligible prompt-system pairs containing at least one valid response |
| Run Quality | State | Derived | Fitness of the report run for validation and publication |
Quick Overview metrics
Quick Overview contains four fixed metric slots. When a connected source changes the unit or data class, the card label becomes unit-specific. An estimated visit and an observed session must not appear under the same unlabeled definition.
| Metric | Data class | Definition |
|---|---|---|
| Total Visits / Observed Website Sessions | Estimated or Observed | Website visits under the displayed source and period, or eligible sessions from a connected first-party source |
| Category Rank | Derived | Position by raw AI Share of Voice within the declared category comparison set |
| AI Referrals (Estimated) / Observed AI Referral Sessions | Estimated or Observed | Estimated AI-attributed visits, or eligible sessions attributed to an AI source by a connected first-party source |
| Bot Traffic / AI Crawler Requests | Estimated or Observed | Estimated automated visits, or eligible HTTP requests from a connected server, CDN, or infrastructure source |
AI Referrals (Estimated) on the Overview Tab is not AI Traffic product measurement.
Intelligence metrics (selected)
| Metric | Public definition |
|---|---|
| Mention Coverage | Eligible valid responses mentioning the target brand divided by eligible valid responses |
| AI Share of Voice | Target-brand mention units divided by mention units across the declared comparison set |
| Prompt Reach | Eligible prompts with at least one valid target-brand mention divided by eligible prompts |
| Citation Coverage | Citation-eligible responses containing an AI-supplied citation divided by citation-eligible responses |
| Inclusion Rate | Eligible responses mentioning the brand divided by eligible responses |
| Recommendation Rate | Recommendation-eligible responses recommending the brand divided by recommendation-eligible responses |
| Top-3 Rate | Rank-eligible responses placing the brand 1â3 divided by rank-eligible responses |
| Commerce Mention Coverage | Eligible commerce responses mentioning the brand divided by eligible commerce responses |
| Leadership Coverage | Leadership-eligible responses mentioning an expected leadership entity divided by leadership-eligible responses |
Display rules
- Every rate exposes count and denominator.
- Low-sample results prioritize count and denominator over a precise percentage.
- Unavailable data must not be interpreted as zero.