Data Definitions
Precise definitions of every Outercite metric, including how each one is calculated in plain terms.
Every number in your Outercite dashboard is calculated using a consistent, documented method. This page defines each metric precisely so you can interpret results with confidence and compare them over time.
All figures on this page reflect the current verification logic.
Metric reference
| Metric | What it measures | How it is calculated |
|---|---|---|
| Citation Rate | The proportion of checks where your business was cited | (Citations confirmed / Total checks) x 100. A check is one prompt sent to one engine. The aggregate citation rate across all tracked organisations is currently 48%. |
| Visibility Score | A 0-100 composite of your overall AI search presence | A weighted combination of citation rate, average position, share of voice, and engine coverage. Higher is better. Aggregate score is currently 74/100. |
| Share of Voice | Your citations as a proportion of all tracked citations on shared prompts | (Your citations / Total citations for all tracked businesses on the same prompts) x 100. Requires at least one competitor to be tracked. See Competitors and Share of Voice. |
| Confidence Score | Certainty that a citation is genuine, per the verification pipeline | 0-100. Derived from the agreement strength between the deep-analysis model and the cross-verification model. Both models must agree for a citation to be counted; the confidence score reflects the degree of that agreement. Not a performance metric. |
| Threat Score | How aggressively a competitor is appearing on your prompts | A composite of the competitor's citation rate on your shared prompts, the intent weight of those prompts (buying-intent prompts score higher), and whether the competitor is displacing your mentions. Range: 0-100. |
| Threat Level | A bucketed label derived from the threat score | Three indicative bands: Low, Medium, and High. Exact cut-offs are subject to change. |
| Average Position | Where in the AI response your business typically appears | The mean ordinal position of your business across all cited responses. Position 1 means you were the first business named. Lower numbers are better. Unranked (not cited) checks are excluded from this average. |
| Lost Citations | Prompts previously cited that are no longer cited | Counted per verification window. A citation is "lost" if the same prompt-engine pair produced a confirmed citation in the previous window but did not in the current one. |
| Per-Platform Coverage | Your citation rate broken down by AI engine | Calculated separately for each of the six engines: ChatGPT, Claude, Perplexity, Google AI, Grok, and DeepSeek. Uses the same citation rate formula applied to that engine's checks only. |
Per-engine citation rates
These are aggregate rates across all tracked organisations.
| Engine | Citation Rate |
|---|---|
| Perplexity | 58% |
| Grok | 53% |
| ChatGPT | 51% |
| Claude | 47% |
| DeepSeek | 42% |
| Google AI | 38% |
| All engines | 48% |
Perplexity is the strongest citer in the current window. Google AI is the hardest engine to win citations from.
What counts as a "check"
A check is one prompt sent to one engine and processed through the full verification pipeline. If you have 10 prompts and track 6 engines, one full verification run produces 60 checks. Across all tracked organisations, roughly half of all checks produce a confirmed citation, reflecting the 48% aggregate citation rate.
What counts as a "citation"
A citation is a confirmed mention of your business in an AI search response. To be counted, the mention must pass the dual-model verification step: the deep-analysis model and the cross-verification model must both independently confirm the mention. Partial agreements are recorded but not counted as citations; they contribute to a lower confidence score instead.
Verification window
Outercite runs checks on a rolling schedule. The "current verification window" refers to the most recent complete cycle of checks across all your prompts and engines. The window length depends on your plan and the number of prompts you track.
Portfolio health buckets (agency accounts only)
Agency accounts see each client workspace assigned to one of three buckets based on citation activity over the last two weeks:
| Bucket | Meaning |
|---|---|
| Needs Attention | Citation rate has dropped or lost citations are rising significantly |
| Monitoring | Activity is stable but below target thresholds |
| On Track | Citation rate is stable or improving, lost citations are low |
Try this in Outercite
Open your dashboard and compare your citation rate against the aggregate benchmark (48%). Then check the per-engine breakdown to see which of the six engines is your strongest and weakest performer.