A/B test your flows

Why A/B test?

A/B testing can take your marketing from good to great, as you’ll get to know your audience better one test at a time. A/B testing allows you to:

• Prove a hypothesis: Confirm what actually matters to customers in your flows. Test new, creative ideas to uncover which ones increase engagement and revenue.
• Craft the right flow message: Use flow message engagement as a source of truth. Test copy, incentives, images, and more to find out what matters most to your recipients when timed right.
• Improve flow pathways: Determine the time delays, number of messages, and content type (e.g., email vs. text message) that drives conversions; thus, optimizing the flow of your flow.

How to A/B test your flows

It’s time to test; log into Klaviyo to start experimenting. In the dropdowns below, follow along to test an individual flow message or a full flow pathway.

Note: Only test 1 element of your flow at a time.

A/B test a flow messageA/B test a flow pathway

A/B test a flow message

What to test

• Email subject line or preview text
• Written copy
• Images, GIFs, or emojis
• Call to action
• Discount amounts or incentives
• Colors, fonts, or styling

How to test

1. Select the flow message you would like to test.
2. Click Create A/B test in the sidebar that appears.
3. Adjust one component of each variation; for instance, the subject line of each email.
4. Determine the percent of recipients who should receive each variation (e.g., 50% for an even split of 2 messages).
5. Set your A/B test to either: select a winning variation automatically once statistical significance is reached (i.e., when Klaviyo mathematically determines that a variation produced improved performance), or manually return to choose a winner once you feel you have enough data to assess performance.

A/B test a flow pathway

What to test

• Time delay
• Number of messages
• Type of messages (e.g., email vs. SMS)

How to test

1. Navigate to the flow you want to A/B test.
2. Drag in a conditional split.
3. Select Random sample as the condition.
4. Select a percentage of your audience to go down the YES path (e.g., 50% for an even split) and edit each pathway. In the example below, you’ll see one pathway will send 2 messages and the other will send 1.
5. Set each path live.
6. Once you feel as though you have enough data to determine a winning variation, return to your flow and click Show Analytics in the top right of the flow builder.
7. Determine which branch leads to higher opens, clicks, and conversions.
8. Click on the conditional split.
9. Set the percentage of recipients going down the winning path to 100%.
10. Set the branch's emails into draft mode for whichever branch is no longer in use.

Deciding what you want to learn tells you which of the two methods to reach for. Sort each question into the test that answers it.