Legal Research

Why You Should Use A/B Testing on Your Website

By Jack Newton

We all run into design-related questions when creating a web page. Questions like “Should this button be red or green?” or “What would the most effective headline for this paragraph be?”. While these decisions may have a dramatic impact on the overall effectiveness of a website, they are often the product of subjective judgement calls by an individual, or worse, a committee.

What if, instead, we could approach such design decisions scientifically? A/B testing makes this possible by treating a web page design instance as an “experiment” where multiple variations of a webpage are randomly presented to page visitors; data measuring the web page’s effectiveness, based on visitor behavior, is gathered, and the best-performing design is selected based on this data.

To make this concrete, imagine your website offers a free newsletter for which a visitor can subscribe. You’d like to maximize the number of visitors signing up for the newsletter. Your existing signup form uses a green button, but you wonder if a red or blue button might perform better. A/B testing allows you to test each of the three variations and to measure which results in the highest number of signups. Each visitor to your website will be randomly presented by one of the three designs. Over a given period of time, say one week, you’ll be able to gather information on each cohort – the “green”, “red” and “blue” visitors – together with their associated signup rates to the newsletter. Here is the outcome of this fictional experiment:

Based on the A/B testing data, we’re able to see that both our red and blue signup buttons generated more signups than the baseline green signup button. Additionally, the red button outperformed the blue button by a significant margin.

A/B testing can deliver tremendous benefits to your website. Simple changes – from changing a button color to tweaking headline text – can increase conversion rates and engagement by an order of magnitude.

A wide variety of tools are available to help with running A/B tests on your website. Google’s Website Optimizer makes it easy to add basic A/B tests to your website, although it requires some level of technical know-how. Performable and Visual Website Optimizer offer easy-to-use, powerful tools for performing A/B tests and require less technical expertise than Google’s Website Optimizer.

If you’re not already using A/B testing on your website, consider giving it a try. You’ll be surprised how high-impact small changes can be.

Full post as published by Slaw.ca on September 13, 2010 (boomark / email).