In the April 7th edition of BtoB's E-Mail Marketer Insight Morgan Stewart, director strategic services for ExactTarget, suggests that instead of relying on various studies, we should consider conducting your own research.
And he explains how to do this:
- Create a single e-mail message that is not date-bound or time-sensitive. If the subject line is "Friday only" and you send on Tuesday, your results will suffer. The message should be exactly the same so that other variables, such as subject line or content, don't skew the results.
- Create randomly selected subscriber segments, each with the same (or nearly so) number of subscribers. Make sure to record the final number of subscribers within each segment.
- Select a time of day that messages will be sent each day and keep it the same for each day you test.
- Capture results. The simplest way to compare results side by side is to wait a week or more after the last test campaign is sent. More than 99% of people will have responded within that period. For best results, record the open rate and click-through rate for the campaigns at equal intervals past each send date. An interval of 72 hours is good.
- Compare results. Generally, just looking at the results side by side will tell the story. However, for a more scientific analysis, use a significance calculator (such as the one found online at www.surveystar.com)to compare campaign results two at a time to determine if the results are significant or just the result of random chance.
I welcome your feedback and questions on this subject!