Email split testing (A/B testing) is a tool that helps you understand which elements of an email perform better. Its goal is to increase open rate, click rate, conversions, and reduce unsubscribes. In this article, we'll cover what to test, how to run A/B tests, and which tools to use.

What split testing is

Split testing is comparing two or more versions of an email on a portion of your audience. The winning variant is then sent to the rest of the list.Example:

  • Variant A → subject: "30% off today"
  • Variant B → subject: "Your personal 30% discount"The metrics compared are: open rate, clicks, conversions.

What elements to test

  1. Subject line Affects open rate.Example: "Free shipping today" vs "Today only: free shipping".
  2. Preheader The text after the subject line.Example: "Order now with free shipping" vs "Shipping's on us".
  3. Sender Example: "Maria from Letteros" vs "The Letteros Team".
  4. Send time and day For example, Tuesday 10:00 vs Thursday 14:00.
  5. Email design Images, button color, structure.
  6. CTA button (call to action) Example: "Buy now" vs "Learn more".
  7. Personalization Example: "Ivan, a discount just for you" vs "Your discount".

How to prepare a split test

  1. Define your goal: increase open rate, boost click rate, reduce unsubscribes.
  2. Test one element at a time, otherwise you won't know what worked.
  3. Determine your sample: at least 1,000 addresses, ideally — 25% to A, 25% to B, 50% for the final send.
  4. Set your KPIs: open rate, clicks, purchases.

How to launch it

  1. Randomly segment your list.
  2. Send the test variants simultaneously.
  3. Collect data (24-72 hours, depending on the audience).
  4. Pick the winner based on your KPIs.

How to analyze the results

Common split testing mistakes

❌ Testing several elements at once❌ A sample that's too small❌ Comparing only one metric❌ Running tests at different times❌ Drawing conclusions without statistics

Which tools to use

  • Letteros → create multiple versions of an email for manual testing (subject lines, CTAs, design).
  • Unisender → automatic testing of subject lines, content, senders.
  • Mindbox → A/B testing of campaigns, automated chains, and product recommendations.
  • Retail Rocket → testing personalized emails and recommendation blocks.

Best practices

  • Test short, specific subject lines.
  • Don't overcomplicate the design.
  • Test hypotheses, not guesses.
  • Use automation wherever possible.
  • Listen to your audience — analyze feedback.

Checklist for launching an email A/B test

  1. Define your goal → What do you want to improve? (open rate, click rate, conversions, unsubscribes)
  2. Formulate a hypothesis → Example: "Change the email's subject line → open rate will increase"
  3. Choose one element to test → Subject line, preheader, button, image, send time
  4. Prepare variants A and B → Create 2 email templates in Letteros
  5. Split your audience into segments → Randomly, at least 1,000 addresses
  6. Set your KPIs and test duration → For example, open rate over 48 hours
  7. Launch the test simultaneously
  8. Collect the data → Open rate, CTR, unsubscribes, complaints
  9. Determine the winner based on KPIs
  10. Record the results and conclusions

Example A/B test report

Subject line: A → "30% off today only"B → "Your personal 30% discount"

Metric: Open rate

Results: A → 25% open rateB → 32% open rate

Winner: Variant B

Conclusion: A personal touch increases open rate. We're rolling it out to future campaigns.

How to make testing a regular habit

  1. Build a testing calendar.
  2. Use checklists (preparation → segmentation → launch → analysis).
  3. Document your results.
  4. Test regularly, not just when performance drops.