• MichaelaKennedy

    I suppose no degree or style of testing is perfect, hence the term testing! Great post, Kohki. It never occurred to me there would be a time not to run split testing, unless there just weren’t enough numbers to glean an analysis, like a small email list, for example. I still have seen intriguing results, though, when a test was done between a promo and a non promo. surprisingly, the non promo did better. It was for a travel service, where people were just a lot interested in the service itself rather than the contest that was being run. If a promo involves too much work from the visitor, it can backfire.

  • http://tinyurl.com/ltjpt98 http://tinyurl.com/ltjpt98

    Interesting post!