A Simple Guide to A/B Split Testing

conversion, optimization, sign-ups, split testing

June 16, 2017 by

Editor's Note: Elena Tahora is a marketing manager at Web Design Malta. Today, she joins us to discuss how A/B Split Testing can improve the success of your online platforms, and various methods to make it work.

 

Creating a new website, ad, or conversion page can sometimes feel like a shot in the dark, and shooting in the dark is scary for online marketers. However, there is a way to make design choices more smartly: A/B split testing is a marketing technique that compares two separate versions of web content to determine which performs best on your audience.

The basic A/B testing process is easy to understand: after creating an item you want to test, you create small variations, and use trial runs to determine which has a higher conversion rate. The best version wins, and from that point on, further A/B tests can refine your product into something highly effective.

Applications

A/B testing has a diverse range of applications, and all of them can help to make the most of your web traffic. Options include:

  • Headlines and Subheadlines
  • Testimonials
  • Paragraph Text
  • Call to Action (CTA) Text
  • CTA Buttons
  • Links
  • Images
  • Social Proof
  • Content Near the Fold
  • Badges and Awards
  • Media Mentions

Although this list focuses on design features that can be optimized with A/B testing, the method can be used on basic sales copy as well. You can experiment with free trial lengths, pricing structures, sales promotion, delivery fees, and basically anything else that can be changed on your site.

While it's clear that this method can do a lot for your business, A/B split testing is also subject to many regulations which - unless followed carefully - can sink your website quickly. For this reason, it's generally safest to conduct your trials with tools designed specifically for the purpose.

In the following paragraphs, we'll review some of the best commercial A/B Split Testing tools. Let's jump in!

A/B Split Testing Tools

1. Content Experiments by Google Analytics

Google Analytics is such a prominent resource for online marketers that it hardly needs introduction. This suite of analysis tools is a baseline for data analytics across the web, providing users with an opportunity that none can rival: retrieving data directly from the world's largest search engine.

For A/B split testing, Google Analytics provides you with Content Experiments. This feature allows you to split-test page variants for a particular campaign. Advanced reporting options give you a detailed understanding of what worked, and why.

google a/b split testing

The main advantages for Google Analytics are: it's free. And as already mentioned, it taps into the largest source of content information on the web.

The only downside is that Google will not provide you with tools for generating variant pages, and Analytics is a complex tool which can be hard to understand if you're just starting out. For beginners, options with a more visual bent may be helpful.

2. Visual Website Optimizer

Visual Website Optimizer (VWO) is an effective A/B Split Testing tool with a great track record. Prominent case studies range from the car manufacturer Hyundai to computer chip maker AMD.

VWO is also very easy to use, involving the following steps for absolutely any content on your website:

  • Include simple source code from VWO on your webpage to make it active 
  • Developing variations on content using a visual HTML editor.
  • Select goals to optimize for (engagement, revenue, clicks, etc.)
  • Start running your tests and track the test results

VWO has a variety of plans that depend on the size of your business, beginning at $49.00 a month for sites with 10,000 monthly visitors.

3. Optimizely

Optimizely is both well-known and widely used, and specializes in boosting CTA conversion rates. Used by nonprofits such as the Clinton Bush Haiti Fund, it's a great place to go if you want to increase your presence and get more people on board.

Optimizely has a three step process:

  1. You must collect data on your current site performance before deciding your goals.
  2. Predict future anticipated progress before creating variants to test.
  3. With variations set, you can start running the experiment and finish by analyzing the results obtained from the A/B split tests.

Optimizely has the big advantage of offering a thirty day free trial period which allows you to see how it works and what it can do before committing to any payment schemes. Pricing after this period is not published on the site, but reviewers indicate that it is mainly affordable for large or mid-range enterprises.

4. SiteSpect's A/B Testing Tool

SiteSpect is a company that offers extensive testing of business ideas, with a designated A/B testing platform that is very easy to use. The provided visual editing tools make it a cinch to brainstorm small adjustments, and generated analytics data will help your company to understand factors that affect conversion rates, reduce costs, drive revenue and mitigate risks.

Like Optimizely, SiteSpect offers a trial period which can help your company to determine if its worth the cost.

5. Unbounce

Not everyone needs the ability to test every font, subtitle, link or button on a website. Some people pursue A/B Testing for one specific goal, and that goal is common enough that Unbounce has one purpose: optimize landing pages.

Whether you are promoting a webinar, a course, or a single product, landing pages are an easy default destination for potential customers. A/B Testing with Unbounce will help you to optimize for the specific goal of gaining more conversions and sign-ups, and also offers more than 80 pre-designed landing page templates to experiment with in case you don't know where to start.

For some users, this tool has everything they need to get started. Through its simplicity and specialization, Unbounce offers a nice contrast to the range of complex features in Google Analytics.

Going Alone...

As mentioned in the beginning of the article, A/B split testing tools help you to avoid pitfalls that could reduce your site rankings. But in case you decide to forego tools and formulate your own code/strategy, bear in mind that Google is watching you, and it has a few things to say:

  1. Avoid cloaking. This means showing one variant of content to your normal site visitors, and another set of content to web crawlers. This is considered extremely fishy, and it's a sure way to get hit with a penalty.
  2. Use rel=“canonical” tags to distinguish a test page from the main version. This is more a matter of sensible practice than a law: unless Google knows which page is your real page and which is the test page, the real page may go down in rankings because visitors won't be able to find it.
  3. Use 302 redirects. When directing users away from a page to a test page, 302 redirects - as opposed to standard 301 redirects - tell Google that this redirection is only temporary. They'll back off and wait for you to finish.
  4. Get it over with. At some point, you have to get on with business and leave A/B split testing behind. So don't test forever, and don't perpetually use the exact same test. This is suspicious, and you'll get in trouble for it.

You can read Google's full list of detailed guidelines here.

Conclusion

Every online marketer wants more visitors, and after getting more visitors, more conversions and sales. A/B split testing is a great multitasking strategy which can boost response rate to a diverse range of prompts.

Whether you pick a tool or develop a strategy on your own, you can't go wrong by trying new things and seeing what works best.

 

Learn more with these related OMI classes:

Testing, Behavioral Analytics & Metrics Best Practices

Strategic User Experience Design

How to Create Personalized Web Experiences That Increase Engagement

 

Visit the Online Marketing Institute to browse over 400 classes in the digital and social media marketing space.

Your evolution starts here. Try our classes for free.10 day free trial

10 day free trial