Configure an A/B test
Test one change, pick a meaningful metric, and record what you learn.
Purpose and outcome
Why it matters. Small changes like subject line or CTA placement can change response; testing tells you which.
Expected outcome. An A/B test with one variable, a chosen metric, and a recorded result.
When to use it
- Larger sends where there’s time to wait for a winner.
Prerequisites and access
- A/B testing available in the subscription.
Suggested roles (not yet approved): Marketing Coordinator
Instructions
- 1
Write the question, e.g. “Does a CTA higher in the email get more clicks?”
Recommended practice - 2
Create the A/B test from the email and build version B.
From training· ~1:07:30 - 3
Change one main variable only.
Recommended practice - 4
Choose the success metric: open rate for subject tests, click rate for content tests.
Recommended practice - 5
Set the test audience split and test duration.
Training used 10% (5% / 5%) and four hours as an illustration, not a default.
From training· ~1:08:30 - 6
Confirm what happens to the remaining audience.
Recommended practice - 7
After the send, record the result and what you’ll do next time.
Recommended practice
Caivan example
Version A has “Explore Floorplans” at the bottom; Version B moves it under the headline. Metric: click rate.
Completion checklist
Common mistakes and troubleshooting
Results are very close.
Small audiences can’t show a reliable winner. Treat it as inconclusive.
Results are very close. Small audiences can’t show a reliable winner. Treat it as inconclusive.
Sources and items needing confirmation
- From training· ~1:07:30–1:09:00Caivan | HubSpot Marketing Training, Sep 28, 2026 (recording transcript)
- Recommended practiceHubSpot Knowledge Base: Run A/B tests for marketing emails
- A/B availability and minimum requirements.
Account configuration described in training reflects September 28, 2026 and may have changed.