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

# Control group

> Hold a share of the people your campaigns reach out of the messages, and see how many more buyers your campaigns really produced.

Attributed revenue tells you who bought after a message. It doesn't tell you who would have bought anyway. The control group answers that. During each round, a share of the people your campaigns reach gets no messages. RevBridge then compares purchases between the people who received messages and the people who were left out. The difference is what your campaigns generated.

> [!NOTE]
>
> Rolling out: this feature is being enabled account by account. If you don't see it, contact [support@revbridge.ai](mailto:support@revbridge.ai).

Go to **Settings → Control group**. Only organization admins can change it.

## Turn it on

**1. Choose the held share**

Under **Held share and purchase window**, move the **Held share** slider. That's the share of the people your campaigns reach who will get no messages during the round. The **Sensitivity** section updates as you move it, and nothing changes until you apply.

**2. Choose the purchase window**

The **Purchase window** sets how long each person's purchases count after they enter, from 14 to 60 days. RevBridge shows how many days your customers usually take to buy again. If you pick a shorter window, you get a reading sooner, but purchases the message only brought forward may count as gain.

**3. Turn on the switch**

Turn on **Control group** and confirm. It applies to every campaign in the account.

> [!WARNING]
>
> Changing the purchase window restarts the round today and discards the reading in progress.

## How rounds work

A round runs for a set period. People held in a round get no messages until it ends. Then a new round starts with a new draw. The page shows when the current round ends and the dates of the first and final readings.

If you turn the control group off, people already held stay held until their purchase window is complete.

## Read the result

| Section | What it shows |
| - | - |
| **The group now** | People held in this round, messages not sent because of it, and the history of rounds. |
| **Sensitivity** | The smallest difference in buyers the test can detect with your share and window. |
| **Answer** | What your campaigns generated in buyers, with a range, for example "The CRM generated +X% in buyers within 30 days of each person entering". The number only appears once a reading has a clear answer. |

If the volume is too small to answer, the page tells you to raise the share or wait for the next round. If the round shows no clear difference, the page says so.

## Where else you'll see it

* In the monitor of a [triggered campaign](https://docs.revbridge.ai/campaigns/event-based.md), held people show the status **Control group**, with the reason: held by the account's control group.
* Within a single campaign, **Creative Optimization** in Analytics shows a different comparison: RevBridge's picks against a small share of random picks. See [How RevBridge decides](https://docs.revbridge.ai/concepts.md#how-you-see-the-gain).

> [!TIP]
>
> **Ask the assistant.** With the [RevBridge MCP](https://docs.revbridge.ai/assistant/overview.md) connected, ask: *"Did my campaigns lift purchases against the control group, and how confident is that?"*


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.