# Proof Impact

What Proof Impact measures, how a conversion is credited to a widget view, what you need before it can measure, and which plans include it.

Proof Impact connects the conversions on your site, such as a sign-up or a purchase, to the widget views that came before them. You see how many conversions followed a widget view, and which widgets, testimonials and pages they followed. It's available on Premium and Business.

Open it from **Proof Impact** under **Analyze** in your project's sidebar. The **Overview** page above it has a **Proof Impact** card with an **Open Impact Dashboard** button too.

> **Note:** Proof Impact reports what visitors saw before they converted. It doesn't prove the widget caused the conversion. To test that, [run an experiment](/help/insights/run-a-proof-impact-experiment).

## How a conversion is credited to a widget

1. **A visitor sees your widget.** A published widget on your site reports a view once at least half of it has been on screen for a second.
2. **Your site reports a conversion.** On the page where the visitor converts, your site sends a conversion event you created, such as `purchase_completed`, with its amount if it has one.
3. **Proof Impact looks back.** It matches the conversion to the widget views that visitor had before it, inside the event's **Attribution window**: 4 hours unless you choose a longer one. By default, only views in the same browser tab count.

A conversion with at least one such view is attributed. The credit goes to the view's widget, the testimonials it showed, its page and its Proof Match rule. When the visitor saw several views, the attribution model decides which one gets the credit: see [Attribution models](/help/insights/read-the-proof-impact-report#attribution-models). A conversion that no view matches stays unattributed, and the **Data Quality** tab lists the latest of them with the reason for each.

## Attribution and experiments

Attribution is an observation. It tells you what a visitor saw before converting, not whether the widget made the difference. To measure a difference, Proof Impact splits your visitors at random, in two ways:

- **An experiment** shows one widget to half of its visitors and hides it from the other half, then compares how often each group converts. See [Run a Proof Impact experiment with a no-widget group](/help/insights/run-a-proof-impact-experiment).
- **Auto-Optimize** splits the visitors of a Proof Match widget between its rules and its default selection, and sends more of them to the option that converts better. See [Auto-Optimize a Proof Match widget](/help/insights/auto-optimize-a-proof-match-widget).

Both count conversions through the same tracking as the report.

## What you need before it can measure

- **A published widget on your site,** embedded with its script, or with the iframe code and its helper script.
- **A conversion event.** Proof Impact refuses events you haven't created. One event is your **Primary outcome**, the one the report opens on.
- **The conversion code on your site,** on the page where the conversion happens, or orders sent from your server.
- **Visitors it may measure.** See [Privacy](#privacy).

Until your project's first conversion arrives, the **Overview** tab shows these as a checklist, **Set up Proof Impact**. For the steps, see [Set up Proof Impact tracking](/help/insights/set-up-proof-impact-tracking).

## What each tab is for

| Tab               | What it's for                                                                                               |
| ----------------- | ----------------------------------------------------------------------------------------------------------- |
| **Overview**      | Whether tracking works, and the conversions and value that followed a widget view, for your primary event.  |
| **Attribution**   | The same numbers by widget, testimonial, Proof Match rule and page, under the attribution model you choose. |
| **Experiments**   | Testing one widget against no widget.                                                                       |
| **Data Quality**  | Conversions that weren't matched and why, how long data is kept, and your conversion events.                |
| **Auto-Optimize** | Finding the rule of a Proof Match widget that converts best.                                                |

[Read the Proof Impact report](/help/insights/read-the-proof-impact-report) covers **Overview**, **Attribution** and **Data Quality**.

## Who can use it and which plans include it

Proof Impact is on Premium and Business. The **Experiments** and **Auto-Optimize** tabs are on Business; a tab your plan doesn't include shows a lock with the name of the plan that does. On a plan without Proof Impact, the **Proof Impact** page shows an upgrade card with **Upgrade to unlock**.

Anyone who can open the project can read the report; the workspace owner and the project's Admins set it up. They also manage the conversion events, download the report from **Export** and change the tracking settings.

## Privacy

- **Same-tab session**, the default, matches a conversion only to widget views in the same browser tab, through a random ID kept in that tab. **Consented journey** also matches views the same visitor saw on other visits in that browser, and on the customer's other devices if your site passes their customer ID with `window.retestimonial.identify()`. Neither guesses who a visitor is.
- By default, visitors whose browser sends a Global Privacy Control or Do Not Track signal aren't measured.
- To measure only visitors who agree, set **Consent requirement** to **Require explicit consent** in the **Developer** tab of your project settings. Your consent banner then calls `window.retestimonial.setConsent("granted")` when a visitor agrees.
- Visitors who aren't measured still see your widgets.

## In this section

- [Set up Proof Impact tracking](/help/insights/set-up-proof-impact-tracking): create a conversion event, add the tracking code to your site and send a test.
- [Read the Proof Impact report](/help/insights/read-the-proof-impact-report): what each figure means, the attribution models, the **Data Quality** tab, changing an event and the export.
- [Run a Proof Impact experiment with a no-widget group](/help/insights/run-a-proof-impact-experiment): show a widget to half of its visitors and compare the two groups.
- [Auto-Optimize a Proof Match widget](/help/insights/auto-optimize-a-proof-match-widget): let a widget test its rules against each other.
- [Proof Impact shows no data or unmatched conversions](/help/insights/proof-impact-not-measuring): every reason nothing is measured or matched, in the order to check them.

## Next steps

- [Read your widget analytics](/help/insights/widget-analytics)
- [Set up Proof Match rules](/help/display/set-up-proof-match-rules)
- [Plans and pricing](/help/account/plans-upgrades-and-cancelling)