5 Simple Steps to Improve Your SMS Marketing with A/B Testing

Previously on this blog, I’ve shared information on how to run successful SMS marketing campaigns. Hopefully, you’ve put that information to work and are seeing results. Do you ever wonder if you could be doing better, though? Are there ways to squeeze out even more business from the same list?

Sometimes, a minor change in wording, timing, or even word placement could affect response rates. This is the sort of thing you want to discover by A/B testing. If you haven’t heard of it before, or aren’t sure you understand it, you’ll want to read my previous blog, How A/B Testing Works in SMS Messaging. It will teach you the basics and the terminology you’ll need to understand this blog on applying it to your business.

The speed of this A/B testing is different than using it for websites. There, you may have to wait weeks or even months to get enough visitors to make your test valid. But SMS campaigns usually have quick turnaround times because of how fast the messages are delivered and viewed.

While SMS messages are usually short, there’s still lots of things you can change. Here’s a list of ideas:

Timing – You already know the importance of timing in your marketing. But could a small change make a big difference? What if you sent the message a half hour earlier, or later? Or maybe in the morning instead of the afternoon? Figuring out the optimal time of day for your messages is probably one of the best uses of A/B testing there is.

Call to action – There’s a whole science to selecting the right words for a call to action. In SMS messaging, the options are more limited than on a website or printed collateral. But there are still options. Is “Click here” better than “Get it here”? Or is “Stop by” better than “Visit”? Is a link to a mobile website better than a coupon code? Time to test and find out.

Offers – These are the heart and soul of SMS marketing. The people on your list are not your average customer. They really adore your brand and products (or they wouldn’t have signed up). So, the offers they are willing to take may be different than your email list, or printed adverts. Try some variations and find out what they respond to the most.

Coupon codes – It may sound silly, but redemption rates can be influenced by coupon codes. This is especially true if the customer has to type them in to redeem an offer. It’s a simple change that you can make easily, but result in a much higher return.

Keywords – Keywords function much like calls to action in SMS marketing. They’re not always, but often action words like JOIN, SAVE, or WIN. Sometimes they are related to the topic of the SMS message, like TICKETS to win tickets to a show. It’s worth testing which keywords work best with a campaign, especially if you’re renting them on a shortcode. Finding the best one could save you, and make you, more money.

Now that you have an idea of what to test, let’s get into the steps you’ll need to follow.

A/B testing involves statistics. That’s the math that helps you figure out if your test gave you good results or not. But for it to work properly, you need a large enough sample size. A sample is just a small portion of your larger list.

Generally, the recommended minimum list size is 1000. This amount lets you perform a test using about half your list (more on that in a moment). This means that you’ll be able to send the optimised message to the other half and get the benefit of the improvement.

You can do testing with smaller lists, but you’ll have to accept more error or less certainty in your results.

To help you find the right sample size for your list, you can visit this page or SurveyMonkey. Both have simple online tools for calculating sample size.

Now that you have your sample size determined you’ll need to create two lists with that number of people on it. For example, if your sample size is 248, then make two distribution lists each with 248 random people on them.

Next chose the message you’ve been sending that’s doing well, but you want to test to see if it can be better. This is the control message. Then pick one thing about the message and make your change (see the list above for examples of what you can change).

Send the control to one of your distribution lists and the one you changed to the second list (just label them so you know which is which – why not A and B?). It’s important you schedule them to go at the same time, unless the thing you are testing is the timing.

Once you send out your messages, all there is to do is wait and see how they performed. Of course, with SMS messaging you don’t have to wait very long. Usually you’ll see how messages did within the first few hours. Depending on your offer, you may need to wait until the redemption period is over to get complete results.

Gather all the data on your two messages. What was the performance difference between the two? Also remember, when you found your sample size you entered an acceptable amount of error (assuming you used the websites above to determine it). When you look at the performance, be sure to include that error.

For example, let’s say message B did 10% better than message A. But when you picked your sample size you used an error (confidence interval) of 10. This means you can confidently say message B would do between 0% and 20% better than A if you sent it to your whole list (10% improvement ±10% error). So, message B should do at least as well as A, but as much as 20% better.

Based on your test results, you now can make some decisions about the test you just ran. Pick the message you found did the best and send it out to the rest of the list (not the ones you used in testing). Make sure to check the results of this message too. Did it fall within the expected performance your testing gave you? If not you’ll want to investigate what else may have happened to change the results, or if the original test was valid.

But if everything went correctly, you’ll be enjoying the improved response and better ROI.

This process may give you ideas on other things you’ll want to test in the future.  Testing can be done on a regular basis or for specific campaigns, but you’ll want to plan for them in advance.

If you have questions about the process above, you can refer to my previous blog on testing. That will help you understand the terms and give you an overview of the online tools that make it easier. Once you get started, you’ll learn a lot about your business, your customers, and enjoy better results from your SMS marketing.