Multivariate Testing

Multi–what? Sounds pretty technical. And it is.

The good news is that all of the complex statistical work happens behind the scenes, built into the technology.

With an easy–to–use interface, and a little guidance from experts, multivariate testing can give you the insight you need to give your clients winning results — in a fraction of the time it would take to use traditional A/B testing methods.

A/B Testing

A/B (or split) testing has been used since the early days of the Web to help advertisers understand which page performs best, version A or version B. Visitors are randomly directed to each version, performance is tracked, one version is declared a winner, and performance generally improves.

A/B testing approach

But there are significant limitations to simply testing version A against version B.

First of all, you know which version won, but you don’t know why — which factors influenced its performance. Was it the headline? Value proposition? Imagery? Colors? Price? Customer testimonial? It’s unclear.

While it is possible to use A/B testing to determine which factors are driving better performance, the process is slow and time consuming because tests must be run sequentially. More factors mean more time, and multiple versions of factors mean even more time, which dramatically extends the optimization period.