Marketing solutions provider Loyalty Builders released today its “Sample Size and Power Analysis Calculator”. It’s a free Web-based tool (part of Loyalty Builders’s Web-based direct marketing system Longbow) that is intended to help marketers determine whether their results are accurate (power analysis) or whether their sample size is sufficient.

The purpose of this calculator is twofold:
1) help users who are testing determine the ideal sample size; and
2) encourage mathematical marketing practices.

The difference between deciding whether or not you want to use the power analysis component or find a sample size, Klein explains, is: if you’ve already done the test, use the power analysis tool to gauge the value of your statistical results; or if you haven’t done the test and are still in the planning stages, determine the power you want — e.g., 85 percent, 90 percent — and calculate how big your sample size needs to be to achieve it.

Testing, says Mark Klein, chief executive officer of Loyalty Builders, is the dreaded chore for most companies, which unfortunately, often gets swept under the rug. At the high end, he estimates that “no more than one third of the companies that we interact with have some interest in using testing.” Note: one-third only have some interest.

The reason companies don’t test, he says, is because: a) they don’t know how; b) they think it’s too time-consuming; or c) they think it’s too costly.

In his previous life as a physicist, Klein agrees that certainly, if companies are going about thinking they have to test one variable at a time, it’s going to be an excruciatingly long process. But technology is available now that makes basic A/B testing and more complex multivariate testing significantly more feasible, both in terms of time and resources.

There are three core techniques you can use to improve your marketing campaign, Klein says — segmentation, personalization, and testing — and they all have a fundamental theme: they require math.