Email A/B testing (also called split testing) involves sending two different versions of an email to separate portions of your subscriber list — with one variable changed — to determine which version performs better. You test version A against version B, measure the results on a primary metric (typically open rate or click-through rate), and use the winning version for the remaining list or future sends. Systematic A/B testing is the most reliable path to continuous email performance improvement, replacing guesswork with evidence.
The key discipline in email A/B testing is changing one variable at a time. Testing a different subject line AND different CTA button colour simultaneously makes it impossible to determine which change drove any performance difference. One variable per test, measured consistently, produces actionable and accumulating insight.
What to A/B test in email marketing
- Subject lines — the single highest-impact test; variations in length, tone, personalisation, curiosity vs. direct, questions vs. statements
- Sender name — 'Anika from Elite Digital' vs. 'Elite Digital' vs. 'Anika Patel'; personal sender names typically outperform brand names
- Preview text — the secondary subject line that appears in inbox view before opening
- Send time — morning vs. afternoon, weekday vs. weekend; meaningful for large lists, less statistically reliable for small ones
- CTA copy — 'Download now' vs. 'Get your free guide'; small wording changes can produce double-digit CTR differences
- Email length — long-form narrative vs. short punchy email; dependent on audience preference and email type
- Personalisation — emails addressing subscribers by first name vs. generic opening
- Plain text vs. HTML — particularly worth testing in welcome sequences and B2B nurture sequences
For A/B tests to be statistically valid, each variant needs a sufficient sample size — typically 500+ recipients per variant for open rate tests and 1,000+ for click rate tests. With small lists (under 1,000 total), A/B testing produces unreliable results because the margin of error is too large to draw confident conclusions. In this case, sequential testing — sending version A to one send, version B to the next, and comparing — gives directional insight even without statistical validity.
Most email A/B tests should be evaluated after 24–48 hours, which gives enough time for the majority of opens to occur (most email opens happen within 4 hours of delivery, with 90%+ within 24 hours). Setting a 4-hour winner determination window is too short — it over-indexes early openers, who may not be representative of your full audience. 24 hours is the standard evaluation window. If your list is large enough to test send time as a variable, the test window needs to account for a full business week to capture weekday vs. weekend behaviour differences.
Subject line testing delivers the highest impact for the least complexity and should be the first test any business runs. Write two subject lines for your next email: one factual and direct, one curiosity-driven or benefit-led. Send each to 25% of your list, measure open rates at 24 hours, and send the winner to the remaining 50%. After running 5–10 subject line tests, you will have clear patterns about what language and approaches your specific audience responds to — insight that transforms all future email writing.