Why? There are many reasons-it's intimidating, we haven't been trained in it, it takes more effort and time, it's too hard to get the necessary data from corporate systems, it's not ingrained in the marketer or corporate culture, etc.
Testing should be a regular part of a marketer's routine and virtually everything-including headlines, content, and offers via website, direct mail, advertising, and email-can and should be tested on an ongoing basis.
Multivariate testing can ultimately be faster and cheaper and tell you more but even simple A/B testing over time will make marketing strategies more efficient.
One obstacle for marketers is either figuring out how big the sample size should be or if they can rely on the results of the sample they've used (power analysis).
There are a variety of useful calculators on the Internet to do this but I thought I'd point you to a favorite, on the website of Longbow, a predictive analytics and direct marketing SaaS from Loyalty Builders (my old employer). It's a favorite because it is fast and easy to use and has robust documentation for all levels of users:
SAMPLE SIZE AND POWER ANALYSIS CALCULATOR
Here's a sample output computing sample size. Note that it shows the curve of the relationship between sample size and power allowing you to quickly see the trade offs of choosing different levels:
One final note-this calculator became a great viral tool for us. It shows up on the first page for a "power analysis" search and became the second most visited page on the Longbow website after the home page.
No comments:
Post a Comment