Run A/B tests in cadences

Test one variable at a time and choose winners from enough real engagement data.

5 min readUpdated June 2026

Choose one thing to test

A/B tests are most useful when the audience is similar and only one variable changes.

Video walkthrough placeholder showing an A/B test on a cadence email step and the analytics comparison after sends.
Video: Set up a cadence A/B test
Video walkthrough placeholder showing an A/B test on a cadence email step and the analytics comparison after sends.
Cadence step editor showing variant A and variant B with subject line or message body comparison.
Cadence step editor showing variant A and variant B with subject line or message body comparison.

Good test variables

  • Subject line.
  • First sentence.
  • Call to action.
  • Proof point or offer.

Avoid false winners

  • Do not change subject, offer, and CTA in the same test.
  • Do not call a winner after a tiny sample.
  • Compare reply or meeting rate, not only open rate.
  • Archive the losing variant with notes so the team learns over time.
Cadence analytics comparison showing two variants with send count, reply rate, meeting rate, and selected winner.
Cadence analytics comparison showing two variants with send count, reply rate, meeting rate, and selected winner.

Ready to try this in FlowForce?

Start free — set up your workspace and follow along with this guide.

Start freeEmail support