The Evolution of A/B Testing in a 1-to-1 World




  • — June 14, 2018

    The Evolution of A/B Testing in a 1-to-1 World

    Optimizing your customer experience requires a dedicated, ongoing commitment to testing. This is true across channels – from your website to your email and ad campaigns and your mobile app experiences. To drive results, you need to test, test, test, and then test some more. For those reading this blog, we don’t expect this to come as a surprise.

    What may come as a surprise, however, is that testing has evolved considerably since it was first adopted by digital marketers in the early 2000s. Traditionally, testing has focused on broad audiences. One might call this a “testing for all” approach. Here, marketers are looking to prove or disprove theories using champion-challenger tactics to determine what works for the majority of their visitors.

    For example, a company focused on demand generation may want to test different hero banners on its homepage. To execute this, a marketer would split traffic between experience A and experience B (which could also be the default experience as the “control”). The marketer would then analyze and compare the results of each experience against key goals such as click-throughs and conversions. This simple use case has been highly effective for companies over the years, and there are scores of technology tools that have been successful in helping businesses facilitate these types of tests.

    But in an environment where companies are now able to deliver true 1-to-1 personalized experiences across channels, the role of testing is changing. So we must ask ourselves, is figuring out the experience that works for the majority of visitors still applicable? We think it’s a good start, but the best experience for the majority of visitors isn’t the best experience for every visitor.

    Here are three ways that testing is changing to optimize visitor experiences and how Evergage is at the forefront of this evolution.

    1. Targeted Testing: With the right solution, business users can apply rules that enable A/B and multivariate testing of experiences for user-defined segments of visitors. Referring back to the demand generation example above, instead of testing the hero banners to find which version produces the best results for the majority of visitors, companies can test to determine, among other things, which hero experience works best for visitors from specific industries (e.g., retail, financial services, healthcare), referring sources (e.g., Twitter, LinkedIn, Facebook), locations (state, region, country), and more. For example, they can test different versions of a retail-specific hero experience by only showing those experiences to visitors from the retail industry or those showing a clear interest in retail-related topics.
    2. Testing Algorithms: Machine-learning algorithms can deliver true 1-to-1 experiences in the form of individualized product and content recommendations, tailored onsite search results, dynamic promotional offers, and more. The only way to know if an algorithm is fully optimized, though, is to test it. An advanced solution enables businesses to A/B test customized algorithms in order to figure out which is most effective. A solution like Evergage can even provide simulated renderings of recommendations for specific individuals or groups (in other words, they allow you to test recommendations by showing you the experiences they would deliver before you deploy them). Taking things a step further, companies should also be able to test machine-learning algorithms within a specific segment of visitors (e.g., high-value customers or first-time visitors).
    3. Cross-Channel Testing: Testing should not be limited to a single channel. To fully optimize experiences, you need to test every customer touch point. Advanced solutions not only allow a company to run tests on its website, but also in a mobile app, logged-in environment, email and more. And the first two points are equally relevant here in that within each channel, an advanced solution would enable users to run targeted tests of machine-learning algorithms and experiences targeted to different segments within specific channels. Among other things, advanced solutions enable companies to test the timing, content and subject line of triggered emails, the content of push notifications and web notifications, the relevance of machine learning-driven, open-time banners and promotions, the effectiveness of rule-based web sequences, the effect of targeted surveys, and the effectiveness of mobile app recommendations.

    Final Thoughts

    The notion of running a test without thinking about the cross-channel journey, targeting, and 1-to-1 experiences is antiquated. Evergage regularly outperforms or even displaces pure-play testing tools, once business professionals begin to realize the limitations of these solutions for optimizing customer experiences at the segment, micro-segment and individual level. If you’re considering testing today or in the future, you need to do so in conjunction with personalization. Once you start looking for a solution that can help you deliver the most relevant experience for each individual, you’ll find that Evergage has one of the most advanced solutions for testing in today’s world.

     

    Digital & Social Articles on Business 2 Community

    Author: T.J. Prebil

    View full profile ›

    (5)