Everyone knows how to test in email marketing. Every now and again you pick something to that might respond to a slight change, you split your email marketing list and if the returns show one gives better results than the other, you use that forever more.
Whilst this might seem a bit patronising, it is true for some, and let’s be thankful so many do it that way. It makes our emails more competitive. What is shows, however, is that the basics of testing are not universally known. Here are some to be going on with.
1/ Reduce the differences between the test email and the control as far as possible. This means you will send them at the same time of day, they will be the same size, and everything other than the test subject will be left as is. The landing pages must be the same. Variation, even small items, will affect the results.
2/ Initially test just the one factor at a time. There are ways of testing more than one factor at a time an we’ll describe the methods in a later article. By testing just the one change means the results are easy to interpret.
3/ Use a statistically significant testing sample to produce dependable results. Keeping the test sample low means that if the change lowers completion rates, you have not lost too much money. If you have smallish email marketing lists then a 50:50 mix might be your only option. Whatever your circumstances, the larger the sample the more dependable the results.
4/ The results are dependable so believe them, even if you don’t agree. They will be more accurate than your assumptions project planning tools. The result you want is one that goes against common sense because that might be one that has been discarded by your competitors.
5/ You should have tests going on all the time. You will never have the only right answer as circumstances change frequently. There’s fashion, politics, the economy and Brexit, all of which have an effect.
That’s the basics. In future blogs we will cover some of the more involved methods of testing.