October 18, 2010|
A big message we try to deliver to our clients is that the launch of your new website is the beginning of the project, not the end. Your web program is an ongoing campaign, not a finite process like the production of a print brochure. For your web program to reach its potential, you should enter into a cycle of experimentation and optimization based on results.
An important tool in improving your website after it is launched is A/B testing, which involves changing variables on a web page in an effort to measure the impact on response rates. Let me give you a real world example. The Barack Obama website has a landing page (screenshot below) that encourages folks to sign up for their email list. You see this page the first time you visit www.barackobama.com and if you click on a link to the site from online ads. From visiting the site often over the last few years, I can tell you that they are constantly optimizing the page to improve response rates.
To A/B test this page, you first need to identify the variables that are in play. The variables are generally a combination of message (photos, wording) and geekier stuff like the physical placement of elements on the page. Here are the important variables I can identify on the page above.
- The picture that is used.
- The placement of the sign up box.
- The language used to encourage you to sign up – “Join the President”.
- The language used on the sign up bottom – “Get Started.”
To maximize the number of sign ups this page attracts, you would create a variety of versions with different variable combinations. Use different photos. Move the location of the sign up. Change the language on the button from “Get Started” to “Join Us.” Experiment with the “Join the President” text. You would be shocked at how much all these little variables can change the percentage of people that fill out the form to give you their email.
After experimenting for a a few weeks with a few different options, commit to the version of the page that has the highest conversion rates (in this case it would be the percentage of users who give you their email). Revisit in a few months to see if things have changed. All of this is made pretty simple using Google Optimizer.
The great thing about the web is that concrete metrics are readily available. You should use that data to constantly improve your website. By using tactics like A/B testing to continually optimize, you can go a long ways toward insuring that you get the most out of your investment in your web program.