A/B test feature

A/B Testing Your Bundles

The A/B Testing feature in Rapi Bundles lets you run controlled experiments on your bundle configurations. Test different prices, layouts, or discount strategies side by side, measure what actually drives revenue, and apply the winning setup with confidence.

In top of the preview, run the A/B test (screenshot)

Traffic Allocation

After clicking "Run A/B Test" and configuring your variant, click the settings icon ⚙️ to manage traffic allocation.

Even Split

Traffic is distributed equally across all variants. This is the recommended starting point, it ensures every variant gets a fair and comparable share of visitors, making the results easier to interpret and trust.

Custom Split

If you want to control the exact percentage of traffic each variant receives, custom allocation lets you do that. For example, you could redirect 30% of visitors to variant A and 70% to variant B. A visual slider makes it easy to set and adjust these percentages.

Ending the Test & Selecting a Winner

When your test has gathered enough data, it's time to close it and apply the winning configuration. There are three ways to do this:

1 — Manual Winner Selection

Review results anytime and declare a winner yourself. Once confirmed, the winning bundle variant goes live immediately on your storefront. A full record of the test, variants, results, and closing date, is saved automatically to your archive.

2 — Automatic Winner Selection

For merchants who prefer a hands-off approach, Rapi Bundles can monitor the test and select a winner automatically when the conditions are met. Here is how it works:

a- The test waits until enough orders have been recorded across all variants, a minimum of 200 orders by default to ensure the data is statistically meaningful.
b- The app evaluates revenue per visitor for each variant, factoring in the actual audience size each one received.
c- If one variant leads the others by a clear and significant margin, the test is closed automatically, the winner is applied to your live storefront and the other variants are archived.

⚠️ If the results are too close to call, the test continues running until a clear winner emerges.

📧 Winner Notifications via Mantle

When automatic winner selection is enabled and a test concludes, Rapi Bundles sends a structured event to Mantle, a signal that includes the bundle identity, the winning variant, and how decisive the result was.

From there, you or your team can build custom notification workflows in Mantle, such as:

• Sending an email to the store owner when a test closes (e.g. "Your A/B test finished — Variant B is now live.")

• Alerting an operations team or distribution list

• Triggering follow-up actions based on which variant won

3 — Suggested Winner

Prefer to stay in control while still getting data-driven guidance? When automatic selection is disabled, Rapi Bundles will still analyze the test and identify the likely winner but instead of acting on it, the platform sends you an email with the suggested winner and the recorded statistics. You then decide whether to close the test or let it keep running.

📁 Archive & Rerun

Rerun

Want to test a fresh hypothesis on the same bundle? You can launch a new test directly from the archive hub without having to start from scratch.

Updated on: 19/04/2026

Was this article helpful?

Thank you!