A Quick Review of Our Favorite A/B Testing Tools

CL Blog Banner

In this post

If you’ve been involved with digital marketing over the past decade, you’ve surely heard people sing the praises of landing page optimization using A/B testing.

“We doubled our conversion rate by changing the page headline.” Or, “We didn’t think an asset was going to work, then we changed the form fields, and downloads went through the roof!”

What is less common though, is an honest appraisal of the tools available to execute these tests, and what you should look for when choosing a solution.

Table of Contents

Identify Your Needs & Challenges

The first step in making this assessment is to evaluate your unique situation – mainly your staff & their workload, your website’s infrastructure, integrations, and agency relationships.  Developing a matrix that outlines this information is vital to choosing the right tool.  

A few considerations include:

  • What internal teams (marketing vs. IT) or external partners, will be designing, setting up, and running the tests?
  • Does my IT staff have the bandwidth to help generate landing pages, set up A/B testing, and execute on-the-fly page updates, based on the results you’re seeing
  • Are you interested in running tests within the company’s mobile app?
  • Do you want those involved in the testing to have access to the raw code of your website?

Once you have a clearer understanding of your specific needs and challenges, it’s time to start evaluating the plethora of options out there. At Closed Loop, two of our favorite tools in this arena – albeit very different – are Unbounce and Optimizely.  It is important to note that while these tools may be competitors in a sense, each have unique features that set them apart from others in the space.

Tool Comparison

Unbounce

Unbounce is really the best tool for rapidly designing, deploying and A/B testing multiple landing pages. After a subdomain is set up on your website, marketers are able to rapidly create customized landing pages for each campaign. This makes Unbounce a step ahead of most other traditional A/B testing options out there.

Strengths:

  • Literally hundreds of mobile-friendly, fully responsive landing pages can be launched quickly without the need for the IT department
  • Ability to create dynamic content, including headlines, based on the referring site’s URL parameters.  This can be hugely valuable when developing landing pages for PPC campaigns, where the traffic stems from specific keyword themes.
  • Streamlined interface that allows for fast page updates, and A/B test setups
  • Integrations with MarketoHubspot, WordPress, Salesforce, and Constant Contact (full list) which gives this tool a leg up for B2B marketers
  • Ability to create Facebook page tabs that can be used for contests, promotions, events, and more
  • Landing pages run on a dedicated subdomain, in their own separate test environment, which puts them under the control of marketers, not IT.  Just make sure that cross-domain tracking is enabled for your web analytics to work properly.

Drawbacks:

  • Although you can track multiple goals per page, conversions are reported in aggregate so you won’t see performance for individual CTAs. It sounds like this is on the roadmap for future releases, however.
  • You can only run A/B tests on pages set up within Unbounce. In other words, you cannot test an Unbounce B version against an existing A version already on your site.
  • If you ever decide to cancel your subscription, you lose your Unbounce pages. You’ll need to replicate them and relaunch in your own environment.

Optimizely

Optimizely is an A/B testing and personalization engine that allows for streamlined testing across all digital experiences within an organization.  Optimizely’s bread and butter is highly statistically significant A/B testing of various aspects of one’s website or mobile app.  It has a WYSIWYG editor for changing out things like headline, text, formatting, etc. (once set up), but does not provide any assistance with the actual page design.

On the personalization front, Optimizely is quickly expanding upon its personalization features, which allow users to dynamically change the content that is displayed to a user, based on their past browsing activity.  For example, if someone has purchased a green hat in the past, Optimizely can leverage this information and show green hats in similar styles, when that user is browsing the eCommerce site two months down the road.

Strengths:

  • Full-site & app (iOS and Android) A/B testing
  • Multi-goal testing allows users to decide how changes impact multiple conversion points (i.e. form fills, social clicks, demo requests)
  • Real-time behavioral targeting that tailors content and its order, based on predefined rules (changing display order, for example)
  • API access, allowing for deep experience customization
  • Full import/export ability allows you to mine data, such as test results
  • Integrations with Google AnalyticsMixpanelKISSmetrics, and WordPress (full list)

Drawbacks:

  • No template support
  • Some IT knowledge is required for more complex feature implementation
  • Some features and integrations only available through upper tier subscriptions

Conclusion

Unbounce and Optimizely are both great tools to add to your marketing tool belt, and each one is great for specific situations. Unbounce is Closed Loop’s preferred choice for B2B demand generation teams looking for landing page optimization.  Optimizely is a good choice for web development teams looking for website optimization. A thorough analysis of your needs should be done before deciding which solution is right for you.  Regardless of the solution you choose, any software that helps you streamline landing page generation and testing is hugely valuable and should be included as part of your overall digital strategy.

GET THE CLOSED LOOP NEWSLETTER​

All the latest digital advertising tips and industry insights delivered to your inbox every quarter

  • This field is for validation purposes and should be left unchanged.

You May Also Like...