-
A/B and Multivariate Testing
There are countless decisions to be made when designing a web page. Variations of colors, fonts, layouts, descriptive text, calls to action, product images, and navigation all combine to produce a nearly infinite range of possibilities. Many times, designers, marketers, or site owners will simply go with what feels good. They find something that looks nice and go with it.
What if this process could be turned on its head, so that customers tell you what works best for them? Better yet, what if they could tell you in a completely passive way, that required no effort on their part? No surveys, no learning curves, no hassle, no guesswork: welcome to A/B and multivariate testing.
We begin by identifying key elements of your pages we can change to find the most effective combinations. Then, we randomly swap page designs, or components of page designs, and collect data about how effective they are in getting your customers to reach a goal. Customers receive variations of your site at random, and their usage patterns tell us what is most effective; again, your customers don't have to do a thing. A goal could be anything from a product purchase to a sign-up for an e-mail newsletter; your goals will be based on what you’re trying to accomplish with your site. Once we collect enough data, we implement the design that your customers decided to be the most effective, thereby making your page as effective as possible. Since only 2-3% of visitors generally accomplish a goal, a page design that increases goal completion by as little as .5% has a tremendous impact on your profitability.