Create an Email Marketing Testing Plan

If you manage an email marketing list for your company or on behalf of a client, this information is for you. Here I will cover the basics of testing with email marketing, and those just getting started will most likely find this beneficial. This article does not take the benefits of list segmentation, customer database integration or other important lifecycle marketing considerations into account. Nor is this article about “E-Blasts”. In fact this blog will never discuss the “blasting” of anything. Bombs, missiles, irrepressible flatulence – these things are blasted. Email marketing messages are sent. Sorry if I am coming off like an email marketing snob.

If you have built a credible list (i.e. no purchased or rented lists) and you are ready to learn more about the audience, testing is a great option. If you do not test, and there are far too many who do not, you are simply leaving valuable knowledge and money on the table.

Why Test Emails?

  • Better Results Now
    More often than not, simply experimenting with different testing elements can yield a better result. More info on those specific elements soon.>

  • Better Results Later
    The primary reason for testing is to provide and derive the utmost value. Testing paves the way to understanding what your customers want, what appeals to them, what they are likely to respond to in the future.

  • Better Results Elsewhere
    Don’t limit the impact of results to just your email marketing campaigns. Testing messages, pricing, creative layouts, etc. can also impact other marketing programs. Utilizing your current customer base as a source of information can help you make important decisions about communicating with new and potential customers.


Testing Methods

A/B Split Tests

  • All elements are the same except for the one being tested (e.g. subject line, graphic, call to action)
  • Administer this test by experimenting with one variable each time you send an email

Multivariate Testing

  • Test multiple elements in a single send
  • Requires a larger list to garner a statistically significant result
  • More about the combination of elements that produces results rather than one or two winners from the bunch.
  • An upcoming post will cover Taguchi-style testing in detail.

Elements You Can Test

Message

  • Subject Line
  • From Name
  • Headlines & Subheads
  • Content – Message vs. Message, Product vs. Product, Graphic vs. Graphic
  • Promotion – $ Off vs. % Off, “Limited Time” vs. Ends Wednesday, Something vs. Nothing
  • Post Click Experience – Landing Page A vs. Landing Page B (this actually leads to even more testing opportunities)

Appearance

  • Graphics vs. Text (this one is admittedly a bit old school)
  • Look & Feel – Colors, Styles, Fonts, Image-Heavy vs. No Images
  • Layout – Email Structure, Element Placement, 1-Column vs. 2-Column

Delivery

  • Time of Day – Many believe that lunch time is a great time to send. Consider your audience. Try early morning and late night messages too.
  • Day of Week – Don’t neglect the weekends. Depending on your product. Sending on a Saturday or Sunday may work very well.
  • Tools – Not sure what tool you like best? Consider running a trial with two email marketing tools and find out which performs better.

How to Measure

  • The Almighty Dollar
    Ultimately, the most important measurement is the amount of revenue you can generate from one version or combination compared to another.
  • Clicks
    The best measure of engagement and interest for an email campaign is clicks. Monitor the specific links email recipients click on and modify your next message accordingly.
  • Opens Are Less Important
    The preview pane can really confuse things here. As many email clients have a preview pane that shows the body of an email without requiring the visitor to actually open it, most open rate measurements are inflated. If your recipient uses the preview pane and has images within the email set to load automatically, an “open” will be registered. Again, clicks should be the number you rely upon to measure recipient interest.
  • Forwards
    Many email marketing tools allow marketers the opportunity to play a “forward to a friend” link. Use this and measure how many clicks this link generates. This is a fantastic method of measuring recipient interest as well. Don’t expect huge numbers here, but even a few forwards is a great indication of success.
  • Unsubscribes
    People typically unsubscribe for three reasons: 1) the messages you send them lack relevance, 2) they get too much email from you, and 3) they get too much email from everybody. People are going to opt out. Just monitor this statistic over time, and don’t read too much into it. However, unsubscribes should be measured in a test as a reflection of relevance, or rather lack thereof.

Email testing plans can take many forms based on your specific business model and objectives. Need help putting your testing plan together? Feel free to contact me for assistance.

4 thoughts on “Create an Email Marketing Testing Plan

  1. Great summary and totally agree. A lot of information packed into a few words.

    A couple of weeks ago I released a whitepaper on the same subject. A key point of multivariate that makes it better than A/B is that as a design tries to make elements that work together, to test elements in isolation can be mis-leading. Take a look at the whitepaper if you wish to know more http://bit.ly/ds4Pb5

Leave a Reply

Your email address will not be published. Required fields are marked *