Split test – the easy way to find out what works

By APSIS

2011-11-14

Do you find it hard to decide which image to choose for your email message? Or is it difficult not to use your favorite headline?

Perhaps it is time to stop trusting your instincts and find a better way to decide what works. A split test lets your recipients decide which version they prefer – and you know for a fact that a majority of your readers will get the best alternative. In this article, we will guide you through the basics of split testing!

Are you curious about what it is that makes your email messages successful – and why it works?

You have probably reviewed the reports from your email campaigns and made a few assumptions about the success rate, but it might be hard to pinpoint the exact reason why your subscribers preferred one message to another. Was it the content, the offer or simply the send time that was right? How will you know for sure?

A split test lets you look into the future. Create two or more slightly different messages and send them to randomly selected recipients – two equally large parts of a list. You will learn something about your recipients, and you will have information that you can act upon immediately by sending the best performing message to the rest of the list. And, the best part – there is no need to speculate or to rely on a hunch; you have the numbers to prove your case in advance.

Careful preparations and planning increase your chance of success. Here’s how to do it:

Identify your goals

Start by thinking about what you want to know and what you want to improve. Do you want a higher open rate? More clicks? Conversions? Would you like to lower the number of people who unsubscribe, or perhaps increase the number of people who share your content on Twitter or Facebook?

If your answer is “all of the above”, you may have to pick a primary goal – but test something that affects several areas. For instance, it is hard to increase the number of clicks and conversions without simultaneously increasing the open rate, which makes the open rate a good starting point in most cases.

What to test

Anything in your email campaign can be tested. But be sure to test only one thing at the time, so you can be sure about what it was that made the difference. Try to make the two test versions radically different from each other. If the difference is too insignificant, there is a risk that the results will not be conclusive.

A few ideas

• Subject line. A good starting point. It is relatively easy to come up with alternative subject lines and you get conclusive results in the open rate.

• Design, color and style. You may need to make major changes in appearance to see a difference in the interaction. If you normally use a lot of images, perhaps a text-based communication may interest those who are not usually so active in your letter. Or vice versa – if you tend to have text-based messges, perhaps more images may improve the readability and increase the number of clicks?

• Offer. You can get a good response when you offer a 25 percent discount. But what if you got the same results with a 10 percent discount? Or no discount at all? Start with a test!

• Call to action. Do sales increase with links that says “buy here” or do your recipients prefer to go through a “read more”-link?

• Landing Page. What happens when someone clicks? A good landing page can be crucial to the outcome. To test landing pages, send the same message with different landing pages.

• Scheduling. What time of day or what day of the week? Send an identical message at different times to know for sure.

The size of the test list

How many recipients do you need for a reliable result? It is difficult to give a definite answer. More recipients give you a better result. But you do not want too many to receive the message that lose – then you might miss many opportunities for business. Try sending your split-test to 10% of the list – then you know that 95% of the list will get the winning option.

How long does it take?

You do not necessarily need to get all the answers before you decide which message won the test. You will hopefully see the trend early on, even if your messages generally continues to be opened up to a week after sending. 
Split Tests removes much of the uncertainty associated with your mailings, so you can be bolder and more creative. When you try something new on a small portion of your list, you have the opportunity to take risks. And if the risk pays off in the test – then the reward will be great when you send to the entire list!

© APSIS International AB. All rights reserved.