> ## Documentation Index
> Fetch the complete documentation index at: https://docs.tryxlr8.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Analytics

> Compare experiment executions over time with trends, baselines, model compare, sources, and insights.

Analytics is the multi-execution reporting view for an experiment. Use it when you want to see how visibility or sentiment changed across runs — not just the latest snapshot.

Navigate to **Reporting → Analytics** in the sidebar, then select an experiment.

***

## Analytics vs experiment results

|          | Analytics                           | Experiment results                         |
| -------- | ----------------------------------- | ------------------------------------------ |
| Scope    | All successful executions over time | One selected execution                     |
| Best for | Trends, baselines, movers           | Deep dive into a single run                |
| Location | **Reporting → Analytics**           | **Strategy → Experiments → \[experiment]** |

***

## Dashboard tabs

After you open an experiment, the analytics dashboard includes:

* **Overview** — KPIs, top movers, source concentration, category performance
* **Trends** — charts across executions
* **Model compare** — visibility experiments only
* **Sources** — citation mix, domains, tracked URLs
* **Themes** — sentiment experiments only
* **Insights** — auto-generated insight cards

***

## Filters and baselines

Use the filter bar to narrow by:

* **Models**
* **Categories**

You can also toggle execution sanitization to include only executions where every active model returned at least one query result.

On **Overview**, choose a baseline comparison:

* **Default** — compare to the previous execution
* **Previous** — explicit previous-execution baseline
* **First** — compare against the first successful execution

***

## KPI bands

**Visibility experiments** surface metrics such as overall visibility, visibility trend, winning categories, and at-risk categories.

**Sentiment experiments** surface metrics such as mean sentiment, sentiment band, negative categories, and sentiment trend.

***

## Empty states

* **No Execution Data** — run at least one successful execution before analytics can load
* **Unable to Load Analytics** — retry or confirm the experiment still exists

***

## Related

<CardGroup cols={2}>
  <Card title="Reading results" icon="chart-bar" href="/experiments/results">
    Interpret a single execution's visibility or sentiment results.
  </Card>

  <Card title="Insights Agent" icon="comments" href="/intelligence/insights-agent">
    Ask natural language questions about experiment data.
  </Card>

  <Card title="Scheduling" icon="clock" href="/experiments/scheduling">
    Keep analytics fresh with recurring executions.
  </Card>

  <Card title="Experiments" icon="flask" href="/experiments/overview">
    Create and manage visibility and sentiment experiments.
  </Card>
</CardGroup>