Why I don’t test subject lines that much

Published 17 th April 2020

Why I don’t test subject lines that much

Over the course of the next few posts we want to start challenging some of the conventional wisdom that makes up email marketing best practice. These are not in any particular order, but are the most common areas we have to push back against with clients.

Today we look at the myths of subject line testing.

This is a classic reaction I remember getting from someone:

“You don’t test subject lines? That’s shocking!”

You might even be thinking well this person is right, it is shocking. Immediately they thought we were some bunch of frauds. Who sends emails without a bit of A/B testing right?

Indeed, it’s a pretty reasonable and sensible assumption testing subject lines will improve performance. That’s the problem with conventional wisdom – it stops people from engaging in any deep thinking about their specific problem, and therefore all the treatments they use are prescribed without this full understanding.

So yes, we do test subject lines – but not that often on normal campaigns. Some of the reasons are:

  • Fundamentally subject line tests very rarely make that much of a difference. Maybe you might see 1-2% difference in open rate between subject lines. It’s not going to ever make or break the success of a campaign.

  • Subject line isn’t even the most significant factor as to whether someone will open the email. We can get a much greater uplift (20-25% increase in clicks) over a sustained time period by testing the ‘From’ name. This always has more impact than the subject line, while if your database is fatigued and fed up with what you have been sending them previously obsessing over subject lines won’t make much difference.

  • If we send to a small sample of my list as a test group of 2-3 subject lines I need to give it long enough before I send the remainder. The trouble is we always need to send emails quickly. Giving 1-2 hours is usually not enough to be reliable. We would often need 12-24 hours to find a winner unless we purposely write a stupid subject line. No-one has the time for that, and rarely do marketers want to eat up a significant part of their database for test purposes

  • What we put in our subject line this time cannot be repeated so what use is the test? As the content of my emails change it means I can’t re-use the same subject line

  • Someone has to write the subject lines, get them signed off, have someone else come back with their pointless change just because they can, and then hope whoever is setting the campaign up doesn’t mess it up. I’d much rather focus resources on things we know are important like segmentation and crafting more valuable content.

We started this series of posts with the obsessive nature of testing of the subject lines as I see so many marketers not just wasting time but also causing unnecessary complexity into their email production process. It is also the fact that the focus is on the wrong part of the email marketing strategy. Making your emails more relevant and valuable is where major uplift occurs.

Is there ever a time to test subject lines?

Yes of course there is. It’s rare we do on a one-off ‘business as usual’ mailing. Generally, our rules of thumb are around if it’s a repeatable test. Triggered based messaging or where we will repeat the campaign to other parts of the database are typical cases (although chances are we’ve got something else in mind we’d rather test first).

And what about AI powered subject lines?

So, you may have heard of platforms like Phrasee that claim to be able to write better subject lines than you with the aid of AI? It’s a concept we like as you are not wasting time on sub-optimal subject lines. You don’t have the issue of not learning from tests as the platform is interpreting the right words to use on any occasion making a constant improvement to the language you use.

Where we believe the big win here is not the huge uplifts on single campaigns that are claimed, but the constant micro uplift on every single campaign sent. For brands sending 3 emails a week getting the benefits of around 150 improved subject lines every year really does add up to significant wins for minor effort compared to the copywriting and sign-off required for traditional split testing of subject lines.

If I’m not testing subject lines what should we be doing?

This article has given you lots of reasons not to test subject lines but how do you overcome the battle with that senior person that heard subject line testing is what you should be doing?

For this you need to find some alternative means of boosting opens to demonstrate you know best and to get them off your case.

  • Test the from name: Use a personal name such as ‘Name – Brand Name’ will always get a 20-25% boost in opens and clicks.

  • Segment your database: If subject lines are tailored based upon one of several segments rather than generic across all will almost certainly boost opens – typically between 10% and 30% increase based upon how tailored you can make it

  • Personalise: And no we do not mean first name! Can you use a past purchase or travel destination in your copy? Anything personalised will stand out far more, and generate more opens

  • Send more emails: Why obsess with open rate when it’s open reach we are after? Simply sending the email to non-openers will boost the number of opens overall by around half the amount again on the next send. The open rate for this re-send might look crummy but let’s say you had 50,000 opens of your first email on a 300,000 send size (16.7% open rate). If the resend gets another 25,000 different recipients opening you now have a total of 75,000 unique openers. The overall reach with this 75,000 is now 25% of your overall 300,000 database, a 50% increase!

Adding these types of practical tools to your armoury can help avoid wasting your time fighting incorrectly perceived conventional wisdom.