Skip to main content

How A/B Testing Works

A/B testing runs two versions of an email step against each other and automatically switches to whichever performs better.

Set up a test

  1. Open your campaign's Templates tab.

  2. Choose the step you want to test.

  3. Add a second variant to that step.

  4. Write a different subject line and/or body for Variant B.

  5. Save.

You can test the subject line and the body. Each step supports two variants — A and B.

How the winner is chosen

This is fully automatic — there's no manual step:

  1. Testing phase — the first 50% of your leads are split equally between Variant A and Variant B.

  2. Winner selection — once that half has been sent, the system compares the variants on your chosen metric.

  3. Remaining sends — the winning variant is used for the rest of the campaign.

The split is always even during testing, and the winner is picked and applied for you. You don't need to monitor the test or pick a winner yourself.

Choosing the winning metric

You can set which metric decides the winner:

  • Open rate — use when testing subject lines

  • Click rate — use when testing your call to action

  • Reply rate — use when testing the body and overall message

Reply rate is the most meaningful for cold outreach, but it needs more volume to be reliable.

Getting a trustworthy result

  • Test one thing at a time. If you change the subject line *and* the body, you won't know which caused the difference.

  • Give it enough volume. A campaign of 50 leads splits into two groups of 25 during testing — far too small to draw conclusions from. A few hundred leads per variant is a reasonable minimum.

  • Make the variants genuinely different. Two near-identical subject lines will produce a meaningless result.

What to test

  • Subject lines — short vs descriptive, question vs statement

  • Opening line — a personalized observation vs getting straight to the point

  • Call to action — asking for a call vs asking a simple question

Did this answer your question?