Take Out the Guesswork with A/B Testing

Ever wonder what it is about your webpage that could be preventing users from making a sale, or increasing your bounce rate? This is surely a problem that every business owner with a website must face at some point. Now have you ever wondered if there’s a way to fix that problem? With A/B testing, there will never again be a doubt about which style/format/colors/buttons, etc. you should have on your website. A/B testing can help all business owners reach their end goals – to make a profit. This method takes out all of the guesswork, and you can see first hand which version of your webpage will get the best response out of users.

So, what exactly even is A/B testing? Its the best way to determine which versions of your website will produce the best results from your customers. It involves having one version of a webpage (“A”), the “control,” and stacking it up against a second version of that same webpage (“B”), the “variation,” and a specific goal you want to achieve. This way of testing results in hard data and evidence regarding which version of the webpage got you closer to your goal.

Here’s an example of a situation that A/B testing is perfect for. Suppose you determine your main goal to be increasing the number of downloads from your webpage. The “download” button should be then the main focus for your testing. In version A, you would leave the page how it is, and keep it the standard version. In version B, the variation, you might make the button larger, a different color, or different font/size in order to hopefully draw more attention to it. Now, the A/B testing is put in place and website visitors are equally distributed among version A and version B. There will then be measurements of how many people saw each variation, and how many of them actually clicked the download button. When the data collected becomes statistically significant, then you’ll be able to see which webpage is the winner!

The term HiPPO was created by Microsoft and stands for Highest Paid Person’s Opinion. Basically what this means, and what it has to do with A/B testing, is that the opinions of the customer are what’s important – not just the HiPPO’s. For companies to strive, they need to listen to their consumer’s feedback and act/respond accordingly. A/B testing is the exact way to eliminate the HiPPO and see exactly what customers respond to on your webpage instead of following the opinion of whoever is in charge.

Amelia Showalter has had plenty of experience with A/B testing. She and her team worked on the Obama 2012 campaign and did intensive testing on which emails would generate the most donations. To keep it simple, anyone who shared an address with the campaign would the get messages from Barack Obama with subject lines that were tested to trigger donations. Amelia and her team predicted which subject lines would produce the highest amount of donations and what that actual number would be, as well as testing the messages themselves and even the different formatting variations of each email. On some subject lines and emails, there would be up to 18 different variations for testing before the winner would be chosen and the email blasted out! Clearly there were countless tests that had to be run, but in the end it proved to all be worth it. Amelia and her team’s findings were that the uglier the message looked, the more money it would raise in donations! Clearly something they were surprised about. Even the use of mild profanity was a hit for some time.

I believe the biggest take away from the Obama 2012 campaign is that your predictions are probably (more often than not) wrong. It’s so very crucial to listen to the customer and users instead of stubbornly only following your instinct. Amelia herself mentioned in the Businessweek article just how shocked she was that her predictions were always off. She never would have thought that the uglier everything looked, the better response it would get out of the viewers. Without A/B testing, she and her team never would have come to the conclusions that they did.

I found another blog that is A Beginner’s Guide to AB Testing, and explains how you need to make sure you plan correctly for your testing and make sure it matches with the goal you want to achieve through it. There are also A/B Testing Mistakes, such as random testing, that you want to avoid. And here are 5 companies that have successfully applied the A/B testing method, as well as their conversion rate results.

Advertisements

One thought on “Take Out the Guesswork with A/B Testing

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s