A/B testing email subject lines

A/B testing sounds a lot more fancy than it is. In reality, several tools make testing your marketing techniques very simple and very, very useful these days.

One of the things we highly recommend testing as soon as possible are your email marketing subject lines. Email marketing is one of our favorite (and most effective) digital marketing channels for both B2B and B2C organizations.

So why not constantly be improving it?

Subject lines

For email marketing, we want to initially be testing subject lines over the content of the email because if someone doesn’t open your email – the content is moot. They will never see that content. Optimizing the insides of the email before optimizing the single biggest thing people look at before even deciding to open an email is crazy talk.

Don’t waste your time. Focus on subject lines first, then think on more advanced tests for things like the content later on.

Test for one variable

When you’re testing subject lines, test for one variable at a time. Test very, very small things and test them for specific segments.

When I say one variable, I mean you want to isolate the two subject lines you’re testing down to a single primary difference. For example, testing “How are you doing today?” versus “How are you doing today, Megan?” by including a personalization field in the subject line. In this test, you would be judging whether adding a customized personalization field makes a difference.

Or testing, “Improve your garden today!” versus “Improve your garden today.” In this test, you would be determining if the punctuation mark at the end has any sort of a difference.

You could test if emoji versus emoji-less impacts open rates. You could test a statement format versus a question format. You could test length – shorter subject lines versus longer subject lines. You could test for all of the things.

Track it and assess

We are in complete and total support of tracking your marketing. It’s important to improving your marketing efforts and keeping your progress organized.

A/B testing is no different. Make sure that you’re tracking what you’re testing and the results so that you can look back and compare these trends. It may be that one type of subject line works very well for one segment and continues to do so. However, it may be that it does well for a short period of time and then you may need to learn from your other results how to keep those open rates doing well.

You won’t know unless you can go back and look at your tests.

No seriously – just one variable

But please, for the love of all that’s good – listen to me here. Isolate that variable when you’re just getting started (and even consider sticking with that approach as you go).

I’ve worked with too many companies that don’t understand this basic premise. They will test wildly different subject lines week over week or month over month and be like, “We can’t gain any conclusions from these test results. Our target audience is just wildly unpredictable!”


Just, no.

Don’t try testing subject lines with extremely different structures or words or formats. Focus on one variable. One company I worked with tested “[Their name]:” versus “[Their name]” (with no semicolon) in the beginning of the subject line and had 20% more opens with the semicolon version.

Note – that’s not normal. Usually you won’t see such a large discrepancy. It also depends on the size of your tests. The smaller your tests, the less statistically significant the results may be. Here’s a great calculator on A/B significance.

But you still want to keep plugging along. Just…with one variable.

One exception

If your open rates remain very, very low even as you’re testing subject lines, consider testing the Sender information. info@ email addresses or other generic email addresses are typically caught up in spam.

Try testing whether sending it from one person’s email address is better than another person’s. It may be that you have improved deliverability rates if you’re no longer using a generic address and it may be that people are more likely to open from a more well-known person at your company.