Email Marketing›Extra· 30 min read

A/B Testing Your Emails

Stop guessing what works. Send two versions to small groups, see which wins, then send the winner to everyone.

What you will learn

Explain what an A/B test is and why it beats guessing
Choose one variable to test (subject line, send time, content)
Read an A/B test result and pick the winner correctly

Test, do not guess

Two marketers argue about which subject line is better. Instead of guessing, you let your subscribers decide. An A/B test (also called a split test) sends version A to one small group and version B to another, then measures which performed better. The winner goes to everyone else.

The golden rule: change only one thing

If version A and version B differ in three ways, and B wins, you will never know which change caused it. So you test one variable at a time. The most useful things to test in email are:

Subject line — the biggest driver of your open rate.
Send time — morning vs evening, weekday vs weekend.
Content / CTA — different offer wording or button text.

How an A/B test actually runs

Say a bakery has 1,000 subscribers and wants to test two subject lines. Here is exactly how the tool splits and sends:

An A/B subject-line test: test on a small slice, then send the winner to the rest

Subject A: "Weekend cakes are 30% off"
Subject B: "Psst... your weekend treat is waiting"

Test on a small slice (say 20% = 200 people):
  100 get Subject A  ->  28 opens  ->  open rate 28%
  100 get Subject B  ->  41 opens  ->  open rate 41%

Winner: Subject B (41% beats 28%)
Send Subject B to the remaining 800 subscribers.

Note: Only the subject line changed between A and B, so we can trust the result: B earned a 41% open rate versus A’s 28%. The tool then automatically sends the winning subject (B) to the 800 people who were held back. More opens, no guessing.

Running your own test, step by step

Here is the simple process you follow inside any email tool:

Pick one variable to test (start with the subject line).
Write two versions that differ only in that one thing.
Choose a test slice of your list (often 10–30%).
Let the tool send A to half the slice and B to the other half.
Wait a few hours and read which version won (higher opens or clicks).
The tool sends the winner to the rest of your list automatically.

Note: The reason you test on a slice first is safety: if a subject line flops, only a small group saw it, and the majority still get the proven winner. This is how you keep improving without risking the whole list.

Which metric decides the winner?

Match the metric to what you tested, or you will read the result wrong:

You tested	Judge the winner by
Subject line	Open rate (did more people open it?)
Send time	Open rate (when do people open most?)
Content or CTA	Click rate (did more people act?)

Tip: Give your test enough people to trust the result. A difference of 2 opens out of 10 means nothing — it could be luck. Test on at least a few hundred people when you can.

Watch out: Never change two things at once “to save time”. If you test a new subject line and a new send time together, a win tells you nothing about which change actually worked.

Q. You run an A/B test on two subject lines. Which metric tells you the winner?

Answer: The subject line’s job is to earn the open, so you judge a subject-line test by the open rate. (You would judge a content/CTA test by the click rate instead.)

✍️ Practice

Write two subject lines for a gym’s “free trial” email that differ in only one way, and say which metric decides the winner.
A test slice shows Version A at 22% opens and Version B at 19% opens. Which goes to the rest of the list, and why?

🏠 Homework

Design a full A/B test for one real email: the variable, the two versions, the test-slice size, and the metric you will use to pick the winner.