axbom’s Twitter Archive—№ 15,295

So how do we make sure our A/B-test results aren’t contaminated by contextual variations? Do the same A/B-test many times? #CJAM4