# Experiments (A/B tests)

> A/B testing is not a screen in PersonaCart today. Compare a change before and after with Analytics instead, and see which numbers to watch.

Source: https://docs.personacart.com/growth/experiments

In short: {"A/B testing is not a screen in PersonaCart today. Compare a change before and after with Analytics instead, and see which numbers to watch."}

PersonaCart does not have an Experiments screen today. There is no place in the app to split your visitors between two versions of a page or an offer and have a winner picked for you. An earlier version has been removed, and nothing about it appears in your sidebar.

This page tells you what to use in the meantime. You can still test a change with a plain before-and-after comparison that is built into [Analytics](/growth/analytics).

1. Pick one number to move, such as **Conversion rate**.
2. Make one change and wait one to two weeks.
3. In **Analytics**, pick a **Custom** period from the day you changed it.
4. Set **Compare** to **Previous period** and read the arrows.

## Test a change by comparing two periods

1. Write down the number you want to move: **Conversion rate**, **Orders** or **Avg Order Value**, all on the Analytics Overview.
2. Make one change, for example a new headline on your landing page, a new price, or a new product photo. Change one thing at a time so you know what caused the result.
3. Wait one to two weeks, so each period has enough visitors.
4. Open Analytics, choose a **Custom** period that starts the day you made the change, and set **Compare** to **Previous period**. Each tile shows the change in percent with an arrow.
5. Save the setup as a view (**Save this view**) so you can return to it.

## Watch the right numbers

- **Conversion rate** and the **Conversion Funnel** card on the Overview: how many visitors reach checkout and buy.
- **Conversion** on [one product's analytics](/growth/analytics-product-detail): the share of its viewers who bought it. It only shows once at least 20 people viewed the product.
- **Where the money came from** on the Overview: split by product, source, country or new versus returning.
- Each sent [campaign](/growth/campaigns) has its own results page, so you can compare two campaigns you sent to similar groups.

:::note[Not as clean as a true split test]
A before-and-after comparison is not as clean as a true split test: seasons, holidays and sales also move your numbers. Compare **Same period last year** as well when your business has a strong seasonal pattern, and avoid running two changes together.
:::

## Questions
