For quite a while now, a mate and I have been homebrewing. We've had varying degrees of success ranging from exploding bottles to pouring entire glasses of head to great (drinkable) beer. It's great fun and a cheap way to drink. But one of the frustrating things about it is how slow the process is. From step one to drinking a cold beer you're looking at two to three months, more than enough time to forget the small things you've done along the way to improve/ruin the beer. So on a recent batch, we decided to do something different. Instead of brewing twenty bottles of the exact same beer we brewed ten variations in one batch. Normally it would take two months to know if putting in an extra spoon of sugar has a positive or negative impact. This time, we put one spoon in a bottle, one and half in another and two in a third. Then we tried the same with raw sugar. And again with brown sugar. Then we tried one bottle with honey. In two months, we'll be able to easily analyse ten variations, which under the previous approach would have taken ten times longer. And you should take a similar approach with your highly trafficked website. I've been doing a lot recently around split testing and optimising and here are a few things I've picked up: Every visit is an opportunity to learn something (big or small) Never make assumptions (there's no need with so much data) More often than not, change is incremental (there's no such thing as a silver bullet) Without even realising it, I massively improved my "test and learn" process and inadvertently started split testing my beer at the same time. Funnily enough, those bullet points above apply to homebrewing too....