> For the complete documentation index, see [llms.txt](https://docs.elevateab.com/elevate-helpcenter/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.elevateab.com/elevate-helpcenter/developer-reference/external-api/statistics.md).

# Statistics

How to read Elevate A/B Testing External API statistics: units, the significance status, the winner, probability to be best, and goal codes.

Every results endpoint reports the same verdict on an experiment. This page explains how to read it; the [API reference](/elevate-helpcenter/developer-reference/external-api/api-reference.md) covers the endpoints themselves.

## Units

Rates and probabilities are percentages from 0 to 100 (`3.4` means 3.4%). `lift` is the change against the control, also in percent. Money is in the store currency, given in `currency`.

## Status

Elevate A/B Testing uses Bayesian statistics: instead of p-values, each variation gets a **probability to be best** on the experiment's goal. `statistics.status` summarizes it:

| Status              | Meaning                                                      |
| ------------------- | ------------------------------------------------------------ |
| `Collecting Data`   | Not enough visitors, conversions or days yet to judge        |
| `Trending Positive` | A variation leads the control, not yet with enough certainty |
| `Trending Negative` | The control leads, not yet with enough certainty             |
| `Near Significance` | The leader is close to the significance threshold            |
| `Significant`       | The leader has reached the significance threshold            |

The threshold (90% by default) and the data minimums are your store's own, from **Settings → Statistical Configurations**, so the status here is the status Elevate A/B Testing shows for the same experiment.

## Winner and probabilities

`/results` adds the detail behind the status:

```json
"statistics": {
  "status": "Significant",
  "winner": { "variationId": 1002, "name": "Variation 1", "probability": 99.2 },
  "probabilities": {
    "CONVERSION_RATE": [
      { "variationId": 1001, "probability": 0.8 },
      { "variationId": 1002, "probability": 99.2 }
    ]
  }
}
```

`winner` is the variation most likely to be best on the goal; it can be the control. Both fields are `null` while collecting data.

## Goals

`goal` is the metric an experiment is judged on: `CONVERSION_RATE`, `REVENUE_PER_VISITOR`, `AVERAGE_ORDER_VALUE`, `PROFIT_PER_VISITOR`, `ADD_TO_CART_RATE`, `CHECKOUT_START_RATE`, `CHECKOUT_COMPLETION_RATE` or `CUSTOM_EVENT`. For `CUSTOM_EVENT`, the experiment's `customEventTrackingId` names the event.