How do you test your campaigns?

Testing is so important to any successful email marketing program. Your subscribers will surprise you time and time again. I can't speak for everyone, but I think that's one of the things that I find most appealing about being in marketing. The psychology behind human thinking and purchasing behavior is really quite fascinating.

Testing gives you all sorts of data that you can't really otherwise get. It's not like you're going to throw questions like "what color of button do you prefer?" into the user sign-up process. Those kinds of nuances are insights that can be gleaned only from testing.

In terms of working with your team and bringing it into your process, I think it's got to be baked into the earliest conversations about new emails. Otherwise, it becomes an afterthought that can typically get pushed to the side. If you wait 'til send time to discuss testing, the conversation tends to become "well, the email is ready, let's just send it. We'll test next time." Much better to develop your email all along with test concepts in mind.

My favorite thing to test is definitely preheader text. I think it is easily the most overlooked, underconsidered part of email marketing at this point. I've seen preheader text tests yield CTR increases upwards of 30%. Subject line tests are also good for quick, easy wins. Subject lines affect not only open rates, but also end up affecting everything. They can have a dramatic impact on subscribers' overall perceptions of every aspect of an email.

Whew! I think that's enough test rambling from me. Excited to hear everyone else's thoughts and approaches!

The way we decide what to test is a pretty simple 2 step process.

1. Have a good idea
2. Convince the rest of the team it's a good idea
These ideas can come from anywhere, analysing stats, reading an article, looking at other peoples emails, occasionally even gut instinct.

We try not to do too many full redesign tests, we want to get some continuity in the emails. Plus we know the current design works, so don't want to miss out on too much traffic if the test goes bad. The smaller (low impact) tweaks like link colour or image alignment we do more often.

A while ago we tested pulling our landing page and email design together, this gave us great results so we do it with all our campaigns now. However the big draw back (for me) is when we split test a new email design I need to build separate landing pages for each variation.

The most important with testing (I think) is to use a control. There are so many variables, time of send, seasonality, even something like the weather or the football scores can affect your results.

Cheers