January 2010 – Online Insights: Merchandising

Next Generation Recommendations

Now that we’ve entered the New Year, online retailers are taking a close look at their holiday sales and overall business results from 2009. And with new 2010 budgets in place and a more promising economic environment ahead, retailers are likely evaluating what new e-commerce technologies they need to meet the growing expectations of today’s savvy online shoppers and to outperform their competitors in 2010.

What’s on your “must have” list for 2010? Is it social network integration? Marketing or site optimization? Better ratings and reviews?

What about personalization?

It’s certainly not a new topic. Personalization has been around for many years and has many types and nuances, including both “explicit” (user-controlled) and “implicit” (automated). It is the automated variety of personalization–the technique of personalizing the shopping experience for each shopper automatically based on his or her unique interests–that has caught fire of late. Also called “predictive analytics” or “micro-targeting,” the goal is to harness available knowledge about each customer to automatically show the right products to the right shoppers at the right time. The benefits? More engaged shoppers, higher conversion rates, increased order values and rich behavioral analytics to help inform merchandising strategies.

The good news for online retailers is that automated personalization technologies have evolved and matured. The bad news is that there are dozens of approaches and solutions to consider, each with strengths and drawbacks. The goal of this article is to explore automated personalization–how it has evolved and where it is today–to help you decide whether it should make your wish list for 2010.

Initial attempts at personalization began early in the decade with online cross-sells. Online merchants–using their in-depth knowledge of the product catalog and consumer purchase behavior–added cross-sell suggestions to product detail pages and to the shopping cart.

Yet merchants quickly realized that cross-selling in the online world–while valuable–was difficult and time-intensive. With large, rapidly changing catalogs, constant inventory changes and lean merchandising teams, most online retailers could only realistically apply cross-sells to their most popular products. Even then, cross-sells were not personalized to each shopper’s tastes or current interests.

Over the past few years, automated recommendations engines burst onto the scene. The first true implicit personalization technology to use real-time data analysis, recommendations engines were marketed as smarter, more efficient alternatives to “one-size-fits-all” cross-selling. Using click-stream analysis, collaborative filtering and statistical modeling, recommendations engines automatically display personalized cross-sells and product suggestions for each shopper, making it possible to cross-sell the entire catalog or the “long tail.”

As recommendations engines matured, online retailer adoption increased. Analyst estimates vary, but most suggest that 20 to 30 percent of electronic retailers today employ some type of automation for product cross-sells (either their own technology or a commercial solution). Sites like Amazon and Netflix have brought recommendations to the forefront and a growing number of commercial solutions have made the technology available to virtually all online retailers, regardless of size.

So where does that leave us today? Are automated cross-sells the end game, or just the beginning? Why aren’t more retailers utilizing recommendations? And can the predictive targeting technology underneath recommendations engines be applied to other aspects of online merchandising across the site and beyond it?

THE FEAR OF AUTOMATION
One reason for slow adoption of recommendations may be the “fear factor.” Many solutions rely solely on click-stream analysis or past-purchase behavior to recommend items. As a result, their relevancy is often off the mark. Suggesting items shoppers aren’t interested in buying is a waste of virtual shelf space. Other solutions can’t easily be controlled and often unknowingly violate merchant strategies, such as showing a certain brand of product with a competing brand, or showing an out-of-stock item.

Here’s a simple analogy that helps explain this fear of automation. Ask yourself if you’d trust an in-store associate to recommend products to shoppers solely based on what similar shoppers in the store looked at and ultimately bought. Or, would you ensure that the associate fully understood your products and merchandising goals, so that the products he or she showed and recommended to each customer matched the customer’s interests as well as your business’s interests? Most online merchants want the latter if they are to automate recommendations–and rightly so.

Fortunately, automated recommendations technologies have matured. Next-generation solutions have more predictive relevancy, deeper catalog understanding and more fine-tuned control. As a result, the products they show to each shopper are both more relevant and more aligned with merchant strategies.

AUTOMATION PLUS MERCHANT CONTROL
Innovative online retailers are using these next-generation technologies to push the personalization envelope and exploit their recommendations engines in new ways. They are building sophisticated, automated merchandising “campaigns” that blend the power of automation with merchant control and strategy.

One top-100 retailer, for example, has more than a dozen, finely tuned, automated cross-sell campaigns on its site. On certain product detail pages, recommendations are limited to products that cost less than 50 percent of the price of the product in view. When a shopper’s gender is identified, recommendations are limited to items only for that same gender. When a shopper is browsing in certain categories, recommendations are focused on that category, or from like categories; when a shopper is looking at a sale item, recommendations are limited to other sale items. In all of these campaigns, out-of-stock products are excluded automatically, and some high-inventory products are pushed more than others. Clicks, conversions and cart values resulting from each campaign can be analyzed to optimize results.

Another retailer is using its recommendations engine to automate its online best-seller and gift-guide sections. Sophisticated, automated merchandising campaigns instruct the engine to select and display the best-selling products or gift suggestions most relevant to each shopper. If I were shopping on this site, I’d see best-selling products and gift suggestions more tailored to my tastes and needs, while you would see products tailored to yours–making both of us more likely to find and buy the right product or gift. In addition, the site’s merchandising team has more time to focus on data analysis to make sure its assortment and gifts are relevant to customers.

If implicit personalization, micro-targeting or automated recommendations are on your list for 2010, you understand the value and importance of differentiating your online store by personally engaging each online shopper. As you begin your evaluation, keep an eye out for solutions that don’t just automate, but which also give you the power to control and guide that automation to align with your strategies and meet your business goals.

Ryan Hoppe is the director of marketing for ATG’s e-Commerce Optimization Services. He can be reached at rhoppe@atg.com.