BMW Salon 2010
Context
Permesso.be is a consumer website, where members leave information on themselves and their family, their favorite past-times, choices and lifestyle. In return, they receive offers, advantages, contests, e-coupons etc… Permesso.be is now the biggest segmented lifestyle opt-in community of its sort in Belgium.
Lifestyle data on the more than 700 000 active members is gathered mainly through surveys. In order to get the members to fill out the presented questionnaires, Permesso.be uses different types of incentives. In recent periods, they gave away cars, full shopping carts, free fuel, holidays etc…
But of course, they are continuously searching for the right incentives, i.e. those that generate the best click through rates to the survey pages. (As once ‘in’ the campaign, drop off rates are steady over most campaigns.)
Therefore Permesso.be used multivariate testing to determine if the level of the incentive influences the click rates from the e-mail to the survey.
Campaign
In 2009, Permesso.be tested the impact of the value of the different BMW visual elements. They wanted to determine if users were more likely to click on a known model or the campaign prototype on top of the campaign or at the bottom.
Only one e-mail campaign had to be developed. With classic tests, this would have meant developing 4 different campaigns. The images with the different values were uploaded into the multivariate tester.
There also was no need to determine test and control groups , as the multivariate tester will test the different versions against each other, until a significant difference can be found. If so, the multivariate tester takes over automatically, using the version that generated the most click throughs.
Results
For this particular campaign, no clear winner was found, i.e. all versions generated roughly the same % of clicks on opened. Permesso.be then decided to route the rest of the campaign through the campaign prototype version.
