How to Do A/B Testing

Photo of author

By Boris Dzhingarov

A single version of a website offers one design, offer, or angle to interest the visitor. Similarly, emails to a larger email list can suffer from the same problem too. When the design, the offer made, or something else doesn’t sit right with the person receiving it, then it will be less successful than it could have been.

Given that getting a new website visitor, email subscriber or another type of approach from a potential customer is so valuable and sought after, tweaking what they see by perfecting it is crucial.

In this article, we describe different approaches around A/B Testing to better understand how it can be performed.

A/B Testing Websites Manually

When you’re looking to perform basic, ad hoc A/B testing on your website manually, then you’ll avoid using tools specifically designed for this purpose. That’s a shame because they’re quite powerful and provide useful statistics to provide greater clarity about what’s working and why.

A/B Testing Using WordPress

Testing different site designs is done in a variety of ways. When done manually, it’s possible under WordPress to use a staging area (when the web host supports and provides one) where the second version of a website is setup. A backup can be created of the original site and then the staged site can be moved to become the live one.

Looking at the traffic stats and using a free heat map plugin, it’s possible to see the activity on a per-page basis. It proves invaluable to see whether visitors spend more time on a redesigned home page than before and what parts of the page drew their attention more than previously.

Related Articles:  How Captection Services Help Your Business

Other Platforms

Sites that are hand-coded can be swapped out for a time.

Also, when a site runs on PHP or another dynamic language environment, it’s possible to programmatically swap between two site designs to show some visitors either version in a calculated manner.

A/B Testing Websites with Google Optimize

Given the growing interest in A/B testing, Google has come out with their online software to help website owners do it themselves. Similar to their Google Analytics for website statistics, Google Optimize aims to do the same thing for A/B testing of websites.

The Optimize interface allows administrators to perform a variety of tests like A/B, plus redirection tests, and multivariate ones too. So, you’re not restricted to A/B testing alone.

Primarily, the system works best for businesses that are advertising on the Google Ads platform but wish to test out different landing pages. The system ties neatly into Google Analytics too, so it logs ad clicks, directs the visitor, and tracks the date – all in one.

There is a free plan. It’s currently restricted to only three “experiments” at one time, otherwise, there are paid plans if you want to perform several different tests on the site.

A/B Testing Websites with Optimizely

Optimizely was designed with A/B testing in mind. It was a leader and still is a leader in the field.

The founders even wrote the book – literally – on A/B testing.

Their focus is keenly on providing better results for people using their service. Therefore, while it’s possible to test page designs and other aspects using Optimizely that don’t have an immediate bottom-line effect, they’re also excellent at improving landing page performing using split testing.

Related Articles:  10 Changes You Must Make To Ensure Your Website Is Compliant for Disabled Users

Testing different promotions on the homepage to see which wording or imagery or discount is preferred by visitors is certainly possible here. They’re also useful when working with calls to action, opt-ins for newsletters, and other customer activities on the site.

With Optimizely, it’s their sole focus to provide the best in testing and it shows. While the Google Optimize tool is certainly worth a look, a skilled team of people who know A/B testing inside and out is going to get better results for you when using Optimizely. It’s also been out much longer and battle-tested, whereas the Google tool is far newer.

A/B Testing of Email Opt-in Forms

Opt-in forms are an important area. In an age where Google has seen fit to penalize companies for more aggressively promoting their newsletter with pop-ups or pop-under windows, it’s become more difficult to gain new subscribers.

New Rules Making Getting New Subscribers Difficult

Link magnets are also currently a sore point with some companies. They offer them – usually a short eBook or another promotional item – as a free bonus when signing up. However, due to the new European rules on GDPR for emails, it’s now necessary for new subscribers to tick a checkbox confirming they don’t only wish to receive the promo, but the ongoing email newsletter too. This makes it doubly hard to get new subscribers.

A/B Testing on Opt-in Pages Worthwhile

Using A/B testing, it’s possible to pick more than one design for the opt-in section on a home page (or elsewhere on the site). Then the designs can be altered depending on changes to the opt-in rate.

Related Articles:  Are Celebrity Brand Endorsements Worth Paying For?

It’s vital to using testing on opt-in pages because every subscriber counts. Once on the list, there are repeated opportunities to market to them, build the brand, and create a return for the company.

A/B Testing of Contact Us Forms

It often goes unnoticed that even contact forms can be tested for usability and general appeal.

For instance, while an AJAX form will respond to entries and actions on the contact form in real-time, they may not be compatible with all users. Seeing in the analytics that people are bouncing away from the contact page without sending a message should encourage designers to take a second look at the page’s design.

Perhaps sometimes it’s confusing people. And if you cannot figure out what, then trying different design ideas can bring the surface what the problem has been.

Lastly, don’t only think about A/B testing from a website and advertising standpoint. Testing email subject lines for a new email blast to 10,000 subscribers is well worth it to improve the average open rates. In conclusion, there’s all manner of different ways to get a net benefit from A/B testing and beyond.