IBM Support

Website Personalization - A/B Testing Recommendations Strategy in Digital Analytics and WebSphere Commerce

Technical Blog Post


Abstract

Website Personalization - A/B Testing Recommendations Strategy in Digital Analytics and WebSphere Commerce

Body

Website and marketing personalization are valuable parts of the marketing strategies used by many of IBMs customers, and certainly most retailers. Up to 20% of website sales can funnel through recommendations, so it's important to get personalization right - or as finely tuned as possible. Both IBM Digital Recommendations and Websphere Commerce offer A/B testing tools to help you make the most of the opportunities you have to communicate with your customers. They are easy to use and often require no involvement by IT - which means you will have the ability to make data driven decisions about your marketing programs in a more timely manner.

IBM Digital Recommendations A/B testing tool allows you to easily swap out and test different offer strategies in a recommendations zone on a page. Should you offer on-sale items to encourage conversion, or will that reduce your overall $ earned from recommendations? Should you help clients make a decision about what to add to their basket by offering similar products? Or should you show customers the products that are ultimately purchased together? A/B testing is a great way for you to decide what is effective for your business. Most A/B testing doesn't even involve your IT group - which can lead to faster decision making.

Image 1 - Digital Recommendations User Interface: A/B Test of Product Page Recommendation Strategies

image

*All Digital Recommendations customers have access to A/B testing. Check out The Digital Product Recommendations User Guide to learn more.

In Websphere Commerce 7.2 and later, you take this testing even further-- you can also find the best location on a page for recommendations. How do you do this? In WebSphere this is called Slot Testing and you execute it via a Marketing Experiment, currently only on a cart page or an order confirmation page. For example, test which location will perform better: displaying recommendations on the upper right rail above the submit order button, or further down the page, below cart contents.

Image 2 - Websphere Commerce User Interface: Marketing Experiment Slot Test on Shopping Cart Page

image

This location-based A/B test works together with the IBM Digital Recommendations tool. You create the offers/zones you want to test in IBM DA Recs, then in Websphere create a marketing experiment to serve the recommendations in one location or in another. Read more about Marketing Experiments at: http://www-01.ibm.com/support/knowledgecenter/SSZLC2_7.0.0/com.ibm.commerce.management-center.doc/concepts/csbexpoverview.htm

Thanks,

Beth

-IBM Digital Analytics Team

[{"Business Unit":{"code":"BU055","label":"Cognitive Applications"},"Product":{"code":"SSPG9M","label":"IBM Digital Analytics"},"Component":"","Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"","Edition":"","Line of Business":{"code":"","label":""}}]

UID

ibm11122375