A/B Website Testing Is Dead. It’s Been Reborn with Personalization

Evergage Blog

Ideas and Strategies for Real-Time Personalization
A/B Website Testing Is Dead. It’s Been Reborn with Personalization

March 20, 2018 by

Optimizely recently announced that it will no longer support its free A/B testing plan starting at the end of last month, leaving around 70,000 websites without a testing tool. A recent guest article from Yvonne Koleczek at Venture Beat suggests that “Optimizely has finally decided to call time of death on oversimplified CRO attempts. After all, bad results reflect badly on the tool itself, and with Optimizely X the company seems eager to distance itself from its past.” She suggests that by focusing on its enterprise offering, Optimizely X, it is throwing its efforts into personalization instead.

I completely agree with this assessment, and I’d like to expand on it a bit in this blog post.

The Power of Personalized Experiences

A/B testing had been heralded as a cure-all for marketing. With A/B testing, you don’t need to guess what will be effective. You can run a few tests to find out for sure. Trying to decide between page design A or page design B? Test it! Headline A or headline B? Test it! Green button or blue button? Test it!

This data-driven approach to marketing is absolutely essential. You should never just make guesses in your marketing, and testing will always be important.

The problem is that A/B testing only tells you what works for the majority of people; it doesn’t tell you what works for each person. In other words, if you select an experience that won an A/B test, you’re still missing the opportunity be relevant to a lot of individual people.

There are many situations where different experiences should be tailored to different audiences. For example:

Promotional offers should be targeted differently to loyal shoppers versus occasional shoppers.

A particular product category may be more appropriate to highlight to visitors in one industry or geolocation versus another.

Some pieces of content can grab the attention of prospects at the bottom of the funnel while others are better suited for those higher up in the funnel.

But that’s just segment-based targeting. Machine-learning algorithms can be used to deliver personalized experiences to each and every person at the one-to-one level.

Each of the different experiences you select for each segment and each algorithm you create should be tested. You need to measure the performance of each campaign to ensure it has a positive impact on your business metrics. And you should continue to A/B test and tweak each campaign to make it even more effective.

Essentially, testing is just as critical as always. But testing how an experience performs for your whole audience doesn’t cut it anymore. It needs to be combined with personalization to achieve maximum results — and to deliver the best experiences for your customers and prospects.

Testing and Personalization Require Advanced Technology

Of course, this type of testing requires much more powerful technology than needed for basic A/B testing — and it can’t be given away for free. Optimizely knows this, which is why it’s abandoning its free testing product.

But a shift to personalization is not an easy move for a solution that was built solely for A/B testing. An A/B testing tool does not need to track and store in-depth behavioral data and comprehensive attribute data at the individual level. It only needs to recognize that someone is a part of the test and determine the experience to show. So to expand its initial testing capabilities, Optimizely has built a lightweight shadow profile that is stored in the cloud, while most of the visitor profile is stored in a mix of the visitor’s browser and cookie. To enable real-time personalization, Optimizely pushes all the data needed to make the personalization decision to the visitor’s browser. This information must fit inside Optimizely’s JavaScript. Therefore, marketers must make tradeoffs between response time and the extent of their personalization. (To learn more about Optimizely’s limitations as a personalization solution, read this blog post.)

This is not the case with Evergage. The Evergage platform was built with one-to-one personalization in mind from the start. All data collected on an individual is stored in his or her unique customer profile, and that data can be used to make a decision about which experience to show in the moment. And marketers can use Evergage as a customer data platform (CDP) to integrate multiple sources of data together to get a clear picture of each and every customer and prospect — and then use all that data to deliver more relevant experiences to each person.

Final Thoughts

This recent strategic move by Optimizely shows us that A/B testing can’t be done in isolation anymore. It needs to be combined with personalization to be effective. But Optimizely X isn’t a truly effective real-time 1:1 personalization solution. Evergage is. To learn why Evergage is the preferred personalization platform for leading companies across industries, request a demo today.

Share This
testingpage2