If we haven't said it enough in this ebook, then we'll say it one more time! The best marketers test their campaigns regularly, and email marketing presents a great opportunity to test all facets of your marketing strategy. In this section, we'll discuss all of the various email tests that you may want to run and what you'll be trying to determine from those tests.
Why Use Email Tests?
The first question that you may be asking yourself is, "Why would I bother to run different types of email tests against each other? I know what works!" You may be right. You may know what has worked best for you in the past or even what works best in your industry based on what your competitors do. But there may just be something that you don't know! The marketing landscape changes constantly. It's possible that a promotion that you ran brought in an entire database of customers who will respond to the color blue more favorably than they will respond to the color red. You won't know this, however, unless you regularly structure email tests to check different elements against each other.
In short, an email test gives you a competitive advantage. While your competitors and your customers may be assuming that things are all the same in the world, you're learning ways to take your email marketing program from good to great. The change may end up being exponential instead of incremental.
Why is Email a Great Testing Ground?
More than any other marketing channel, email presents you with an almost ideal channel for testing various marketing ideas, offers, concepts and more. Why is that?
You Can Easily Split into Equal Lists:Whereas you can't control the demographic breakdown or frequency of visitors to your website,you can easily create an evenly demographically split version of your email list. Because you collect additional information about your email list, you can ensure that all your email testing segments include different types of customers. Even if you don't collect additional information about your email list, if you simply split your list down the middle, you'll almost always end up with an evenly demographically split list to test to.
Real Time Results:Email tracking is almost entirely real-time. You can see how many people open, click, and respond to your email almost as quickly as it is sent. You won't need to wait for long-tail reports to see if the difference in the response to two email tests is significant. You make changes to your other marketing channels based on the response to your email tests quickly and effectively.
Low Effort:Unlike a print campaign or a telemarketing campaign, you can test marketing creative collateral and promotional concepts with a fairly low effort. All you'll need is some graphics updates and some copy updates. You won't need to design and produce print pieces or to take up vast man hours and human resources for a telephone campaign.
Ability to Isolate Components:Finally, the amount of data that is tracked with an email marketing campaign is significant and allows you to isolate components. Did women respond to one type of email often while men responded to an entirely different type? Did one email service provider give you the best response rates? Because email marketing data is so detailed, you can craft very specific email marketing campaigns for the future based on what yields the best results.
How to Run an A/B Test
Before you begin to run your email test, you'll want to make sure that you understand the best way to split your email list for a clean A/B test. Because you want the resulting lists to be as even as possible in terms of valid email addresses and demographic information, consider taking the following steps before splitting your email list into two parts:
- Remove Inactive Users: Begin by removing all of the most inactive users from your main list. This may simply be users who have never opened an email, or it may be users who haven't opened an email in a very long time. You'll need to decide for yourself what the criteria will be.
- Remove Highly Active Users: Also remove your most active users. This most likely will mean people who open the majority of the emails that you send, but it may also simply mean anybody who has opened an email within the last month or the last two email sends.
- Sort Alphabetically:Often, the best way to sort a list and then split it down the middle is to sort the list alphabetically. It will often give you the most randomized data.
- Do Not Sort by Join Date: However you choose to sort your list before splitting it,be sure that the list is not sorted by the sign-up or join date of the users. Your most recent sign-ups will be more active and likely to open than your older ones. If you sort by join date, one list is likely to respond to the email better than the other one based simply on the fact that they have engaged with your company more recently.
- Split Your Highly Active or Highly Inactive Users Separately:After you have sorted and separated your main list, use the same technique to sort and separate your highly active or highly inactive users. Then split the results back into the two new segments of your main list. This will ensure that each of your A/B segments will have a sampling of average, highly active, and highly inactive users.
- Check the Percentage of Email Service Provider Addresses:Do a quick check of each of your newly halved email lists. Make sure that you have roughly the same percentage of Hotmail, Gmail, Yahoo! and any other large email service providers on each list. If you do not, you may risk having your results altered if one of your email service providers sends your message to the junk or spam folder.