By Angela McCalister, Director of Client Services, ATG Recommendations

As the Director of Client Services for ATG’s automated recommendations service (an increasingly hot technology in the industry), I am bumping into some unexpected questions from retailers about how the tool actually goes about making recommendations to consumers. They’re seeing that some of the recommendations it generates don’t seem to make sense at first glance. As I think about it, the questions are very understandable and probably more common than I had realized.

Predictive Recommendations– A quick primer
First, here’s just a little background if you’re not familiar with what an automated recommendations service does. It’s basically a predictive merchandising solution. This tool, which operates as an on demand service running on top of an existing commerce site, recommends products for individual consumers, based on each person’s specific shopping behavior. Sometimes called “predictive merchandising tools,” these kinds of services are helping merchants generate higher average order value and increase conversion rates.

But don’t let this simplistic explanation mislead you into thinking this is simplistic technology. There is a great deal of extremely sophisticated science that goes into creating the algorithms that drive the service. Behind the scenes of a shopper’s online experience, the services is observing behavior and employing very complex predictive models to collect all sorts of behavorial and demographic data points, and then recommends products to those shoppers based on the collected behavior of their shopping session and shopping sessions of other similar consumers.

These tools actually “learn” shopping behavior and grow smarter with every session they observe. As a result, the tools know when I am interested in purchasing a new pair of jeans vs. when I’m looking for some really high-end perfume.

What does a duffle bag have to do with a black blazer?
But back to that question I referred to earlier. One major clothing retailer recently asked me, “If this tool is so great about recommending the right product, why doesn’t it always recommend the matching pants with the jacket of a clearly identify pant/jacket set?” That is a great question! It turns out that while retailers market pants and jackets as sets, many women actually buy clothing as separate pieces – either the jacket or the pants, but not both together. In addition, the automated recommendations tool has learned over time that the shopper is far more likely to buy the black blazer in the same shopping cart as a duffle bag. No human being would ever have predicted that combination, but the recommendation service has observed the correlation over time. So why not serve the shopper product ideas which include both the merchant-defined clothing sets and the products discovered by the automated tool? By the way, yes, the tool can be tuned to display catalog style sets (but wouldn’t that spoil the fun?).

Best practice: Have predictive technology coexist with merchant-defined rules
Back to that retailer. As he and I continued talking, it became clear that the best practice is to combine manual and predictive merchandising techniques as a comprehensive strategy. The retailer will continue to market products in sets, as doing so helps create a fashion look and suggests buying options. But they will also use a predictive recommendations engine to get inside the mind of the shopper and suggest products which meet the shopper’s immediate desires.

Data from the automated recommendations sessions will suggest new trends and will help the retailer lead the fashion scene. So this retailer will have the best of both worlds – an ability to promote a certain look every season, and to lead the fashion industry in trend identification.

Last week consumers opened up their wallets to shopping this past Cyber Monday. Market research firm Comscore revealed that the sales volume was 15% higher than last year’s Cyber Monday sales. Even more evidence that e-commerce is proving to be a bright light is that the four-day period between Black Friday and Cyber Monday saw a sales increase of 13% – from $1.9 billion to $2.2 billion: more here.

By Bob Burke, CEO of ATG

The Wall Street Journal has recently run a couple of interesting stories related to customer loyalty, online marketing, and social computing. One of these articles (Marketing in the World of the Web, by Tom Hayes and Michael Malone) centers on the power of social networks to “reconfigure individual behavior.” It argues that retailers need to leverage this social movement to help their efforts to create a more loyal consumer base, thereby increasing sales and profits. Offering their thoughts on a “Marketing 3.0” world, the authors focus their comments on 5 key themes:

- Gaining customer attention before winning loyalty
- Appealing to specific communities of interest vs. the general customer base
- Targeting the “meganiches” of highly specialized, yet relevant, online communities
- Exploiting the viral power of social communities
- Making offline and online channels work in tandem

Which got us to thinking…
This entire topic is near and dear to us at ATG, and to our customers who are facing the challenge of building revenue in an environment where, for many online businesses, the majority of revenues are starting to come from an existing, steady, loyal customer base more than from newly acquired customers. Rather than focus on customer acquisition tactics, the companies that will thrive during a downturn will be more focused on retention and finding a new twist on their marketing strategies. That twist just may come from the growing “cloud” of online social communities that the article discusses.

Successful online businesses are coming to understand that user generated content and social communities can be critical elements of a state-of-the-art online customer experience. They’re learning how to translate that understanding to increased customer loyalty and sales conversions. However, many other retailers are still trying to figure out how to use this powerful voice of the consumer, and the wisdom of the crowds, to their advantage.

As social networking continues to grow, there are new participants, there are new outlets, and yet there are limited connections between a “Web citizen’s” goals and a merchant’s goals. The opportunity here is for retailers to engage and leverage their “zealots” – those consumers deeply attached to a company or its products and services – to become stronger advocates. In addition to coming back themselves, they have enormous potential to spread enthusiasm to their friends and other contacts. From the consumer perspective (that of both zealots and more passive consumers), retailers need to provide as many easy ways as possible to share opinions, keep the zealots and their networks informed, and reward their efforts to share their enthusiasm.

Find and nurture the zealots
To be successful, merchants need better capabilities to identify and recognize their zealots and they need easy ways to provide targeted, dynamic content such as product updates, promotions, and special offers to encourage them to share their enthusiasm far and wide throughout all their social connections. They also need the tools to track results, learn from, and refine their marketing efforts to their zealots.

In order to help merchants turn these loyal zealots into profit engines, effective commerce platforms not only need to provide integrations with top social sites, they also need to give merchants the power to:

- Automate information updates to web pages both in the store, and elsewhere on the site
- Allow shoppers to post content from a merchant’s site to their desired social pages
- Automate information updates to their subscribers

In order to properly empower the social shopper to be a stronger ambassador, merchants need to begin targeting and segmenting consumers in order to provide contact and products that are relevant and meaningful. They need to know what their power shoppers and power content contributors are doing on their site, so merchants can serve them special, appropriate content that speaks to their status and ability to influence others.

Use a cross-channel strategy to effectively reach social communities
There is also an urgent need to ensure cross-channel consistency when customers engage with a merchant. Customers don’t necessarily think about what “channel” they are accessing when approaching a merchant. They don’t understand why the experience they have on a company’s web site should be different from the customer care or in-store experience. They merely think about their interactions with a company as all part of a whole. To that end, retailers can provide access and sharing via mobile channels. Mobile commerce is reaching critical mass overseas, and in the US we are on the cusp of a breakthrough. Engaged social shoppers will continue to evolve and look to utilize their smart phones to research, shop, and share product data among peers and colleagues.

These are initial steps that we’re seeing top merchants both considering and successfully implementing in an effort to better engage their zealots, incent them to share, and drive long-term loyalty and lifetime value. In today’s social commerce environment, these tactics are decreasingly seen as “nice to have” customer experience offerings. They may soon be key to a retailer’s ability to survive – and thrive.

The Boston Symphony Orchestra announced yesterday that 179 music downloads are now available directly from its Web site, www.bso.org, with more coming in 2009.

The Boston Globe story quotes Rich Bradway, the BSO’s associate director of e-commerce and new media, who said, “if we want to try and start putting out recordings of live performances, we have all the technology in-house now…. Now that we have the e-commerce side of it, it completes the circle.”

To learn more about the BSO’s e-commerce initiatives with ATG, read their case study.

Congratulations to the CVS.com team, for their well-deserved recent award from eHealthcare Strategy & Trends magazine.

CVS.com was named ‘Best Overall Internet Site’ for 2008 in the consumer healthcare product sales category for the eHealthcare Leadership Awards. The site was chosen by an independent panel of judges among more than 1,100 entries competing in one of 17 areas across 12 different award categories.

We know that CVS/pharmacy will continue to innovate both with their health and prescription services and their retail business, and we can’t wait to see what’s in store for 2009.