Download our FREE ebook of 43 A/B testing case studies from the world's leading companies for test ideas and inspiration. Download Now

Optimizely Blog

Tips & Tricks for Building Your Experimentation Program

X

Download our FREE Testing Toolkit for A/B testing ideas, planning worksheets, presentation templates, and more!

Get It Now




Today, we are excited to announce the launch of a new Personalization capability we are calling Adaptive Audiences. Optimizely’s Adaptive Audiences gives marketers an entirely new way to personalize experiences. It gives you the control of rules-based targeting, with the reach and scale of automated personalization. The idea is to use machine learning to match people with dynamic audience segments — not rigid rules — based on real-time user behavior.

Optimizely Customer Crate & Barrel has already put this new technology to use. With the launch of Crate and Kids last year — a new children’s collection — they saw an opportunity to tailor the experience of their Crate and Kids site. “As a destination for children’s furniture, home decor and toys, we have a lot of products and want to target the parents shopping for their children. The challenge is that we have baby-specific products as well as products geared towards children that are a bit older. We were wondering if building out content for these two groups would be worth the effort. Now, Adaptive Audiences lets us differentiate between these two groups quickly, with great accuracy and provides significant lift to performance. Showing customized content to these two groups is now our always-on personalization strategy for these pages” says Christine Garvey, Senior Manager of Personalization & Optimization at Crate and Barrel.

Traditional Approaches to Personalization vs. Adaptive Audiences

When it comes to personalization there are two technologies in the market today (1) rules-based targeting and (2) automated 1:1 personalization. Generally, rules-based targeting defines personas using rules like: “furniture shoppers are visitors who have viewed 4 furniture items in the last 7 days.” This approach has proven to be fairly powerful — especially for mature organizations that have a team of analysts to define these rules. As marketers, it gives us fine-grained control. But this approach can also be hard to scale when you have to constantly adjust rules as customer behavior changes.

The second approach is automated 1:1 personalization.  With automated personalization, you’d create dozens of different user experiences and let an algorithm decide when to feature each one. But we know this is not always a winning experience for the end user. I may browse a couch for a potential gift idea, and now all I’m seeing for the next couple of months are the same sofas following me around the web.

With Adaptive Audiences, Optimizely saves you time and resources by 1) enabling you to define an audience simply by specifying a few keywords that capture the essence of this audience, and 2) continuously deciding if an individual visitor is part of this audience or not.  Optimizely uses word embeddings from the cutting edge of Natural Language Processing (NLP) to understand the interests of site visitors. Only taking a handful of plain words as a guide, Adaptive Audiences dynamically identifies the most relevant people to target.

Personalize Quickly and Broadly

Traditional machine learning-based personalization products often depend on a significant amount of data upfront to deliver truly personalized experience.  This requires both instrumentation and collection time, quickly adding up to multiple months if not more.

With Optimizely, customers can be up and running personalization campaigns targeted at Adaptive Audiences immediately because we use a language model pre-trained on a large, diverse corpus of text.  Visitors only need to visit one page for Optimizely to start predicting their interests on the site.

Adapting to Constant Changes in Consumer Behavior

Your customers’ behavior changes.  They change based on the time of day or the day of the week.  They may also change simply because the visitor has already seen a product or article on your site.  Rules-based targeting is difficult to scale because you constantly need to update your audience definitions.  Black-box automated personalization can take a while to adapt to these changes, lagging behind your visitors’ preferences.

With Adaptive Audiences, you don’t need to keep changing your audiences around as browsing patterns, seasons, content or products change. Optimizely’s machine learning algorithm harnesses real-time user behavior to match people with audience segments.

The Future of Personalization

We are really excited about the future of Personalization at Optimizely. In addition to launching Adaptive Audiences today, we are also upgrading our Recommendations offering to bring a whole new set of algorithms to market and give you the power to build on top of these. We’ve also recently launched Flexible Segmentation, which lets you deeply analyze and segment your results to identify differences in how your customers are responding to an experiment or personalization campaign. The future is bright for brands who want to personalize experiences for their customers. We are excited to be helping leading brands like Crate&Barrel, Brooks Running, and Atlassian transform their Personalization programs.

To learn more about how you can start using Optimizely Personalization today, visit our product page. We will begin rolling out Adaptive Audiences to all Optimizely Enterprise Web customers starting July 1st. To learn more about getting started and see a personal demo, please contact your Customer Success Manager or get in touch by requesting a demo up here.

Please join us on July 17th to hear from Christine Garvey, the Senior Manager of Personalization at Crate&Barrel as she shares the Crate&Barrel Personalization story.

Optimizely DXE Report