Experimental Design Primer
Many scientific techniques exist for testing two or more variables; these methods are increasingly being applied by marketers to examine the effectiveness of direct marketing, advertising, Web site design, and other programs. Here’s a brief rundown on the more common types:
A/B test:
Also known as a split-run or champion-challenger test, this basic screening method tests a new version of a single variable against the existing version (the control) to determine which achieves the better response.
Full factorial design:
An experiment that takes into account two or more variables (factors), each with discrete possible values or “levels,” and examines all possible combinations of these levels across all such factors. This enables testers to study the effect of each factor on the response variable and the effects of interactions between factors on the response variable.
Fractional factorial design:
A version of full factorial design in which a subset of variables is chosen; this method is used in cases where the number of experiments using a full design would be too high to execute.
Plackett-Burman design:
A screening technique used to examine the effects of several variables in one experiment and avoid multiple runs of the same basic test.
Taguchi method:
Initially developed to improve automobile and product manufacturing, this technique reduces a problem that might involve thousands of variables down to a relatively small number of experiments, which can be conducted simultaneously, and predicts the optimum outcome by estimating the positive or negative impact of each element.
Sources: Wikipedia, Marketing Experiments Journal, Offermatica, Optimost.
