Successful online companies are built to continually experiment, and A/B testing is one of the most popular types of tinkering. But this post makes the case against A/B testing, offering three reasons: 1) individual visitors are not all the same: though version A might work best for most of a company’s customers, there will always be that other set that would respond better to B; 2) companies should test the “why,” not the “what”: focus should be on “influence strategies” that are used on a website; 3) don’t exploit when it’s time to explore: optimize the “explore-exploit trade-off” to gauge whether a less popular version today becomes more popular later on. (Econsultancy)
Network
