6 Steps to Scientific Email Testing

Email is unique in that it’s a low cost, high performance channel that can be used to really learn about your database.  By developing an ongoing testing plan, you can collect some valuable behavioral data on your consumers that will inform all of your programs.  (Check out this ClickZ article on using the applying the scientific method to behavioral marketing across all digital channels: http://www.clickz.com/3633553)

Testing means more than a haphazard A/B split every month or two.  It means developing a long term plan and sticking to it.  The good news is that the method has already been developed – it’s the Scientific Method, used by scientists of all kinds the world over.

Here are the 6 steps to the method, and how to apply them to your email marketing plan:

  1. Make An Observation – Identify the metrics that you want to improve.  Pick one at a time, and start with the one that is the easiest, and will make the most impact.
  2. Ask A Question – Determine what elements of your email affect the selected metrics. (i.e. if you’re looking to boost open rates, you might want to take a look at your subject line)
  3. Form A Hypothesis – This is where you decide what you’re going to test against your standard.  Don’t enter into this step blindly – do your homework and make an informed guess on what the B version is going to be, and why it should outperform the A version.  Write down your hypothesis.
  4. Conduct an experiment – Run your test, and do it right.  If you’re testing subject lines, then make that the ONLY difference.  Don’t tweak the copy or change the offers or move the CTA.  One thing at a time – that way you’ll have fully attributable results.
  5. Evaluate the results – When the numbers start pouring in, pull out that hypothesis you wrote down, and use it to evaluate the data.  Did your B version outperform the standard A version on the metric you were testing?
  6. Do it all again – So you found your winner, the better of the two, and have improved your program.  But, there’s something out there that’s even better, that’s going to boost performance even more.  How do you find it?  By continuously repeating the process.