Recommendations and Personalization in 2012

The recent explosion in software-as-a-service recommendation platforms has delivered a powerful personalization tool to marketers in every size company. “People who viewed this also viewed that” isn’t just for Amazon anymore.

I see three evolutionary trends for personalization and recommendations in 2012:

1. Convergence of recommendations, search and ads solutions
2. Vendor consolidation
3. Market expansion

In this post, I’ll delve into convergence.

People discover content that is useful by seeking it or by bumping into it. While you might assume that seeking involves Google or the search box on a web page, other seeking behaviors include going to a web site home page and clicking on a category, or opening an email and deciding to click through. Seeking isn’t hugely entertaining, so most of us would be happy to avoid stumbling around a web site or trawling through pages of search results. Therefore, most of us are pleased to be presented, unprompted, with content that is interesting and valuable. Display ads, email with recommendations, and personalized web pages are this year’s best technique for connecting people with exciting content – before they ask for it. In human terms, ads, recommendations, search, and personalization serve a very similar purpose: discovery.

While the human activities are closely connected, until recently the technologies supporting those activities have been diverse. In the past year, however, one set of technologies has been applied to selecting content for recommendations, display ads, and search results. Recommender systems are information filtering tools. They apply collaborative and/or content-based filtering to predict which items (content) will appeal to a visitor at a moment in time. Systems that select display ads and those that rank search results are also information filtering tools, and some of those systems are beginning to use collaborative or content-based filtering.  Today, some of the leaders in recommender systems offer solutions that also address site search or display ads. Adobe, Avail, Baynote, Locayta, MyBuys and RichRelevance all offer solutions for search and/or display ads in addition to recommendations. I think in 2012 the majority of vendors with solutions on market since 2009 will join Adobe et al in offering two or more solutions.

What does this mean for users, marketers, and vendors? The impact is positive for users and vendors, but somewhat mixed for marketers.

Customers will find that useful content (products, articles, ads) is much more commonly encountered, without having to search or even (in case of ads and email) type a URL. They will find their digital life more entertaining, smoother, and easier.

Vendors of recommender systems have a rich opportunity to offer more services to their current customers. Marketers are very keen to present users with just the right content, whether their goal is to generate a lead by capturing a user’s attention, nudge a user to the next step in the journey with reassuring information, or making a sale by presenting an offer this user can’t pass up. So marketers are the natural clients of solutions that support discovery. Typically, marketing teams are either deciders or influencers for recommendation, site search, and display ad solutions. So the expansion of vendor solutions across the discovery market creates a “land and expand” opportunity for vendors. But I anticipate another significant effect: users become more valuable customers, causing marketing teams to become more valuable clients. Marketing teams that deliver personalized ads and great search results will increase customer loyalty and frequency of purchase. They will reap revenues from more visits and more orders from each customer. Each visit is likely to generate more recommendation-driven revenue. Same number of customer users, more revenue for vendors and their clients. Recommender systems fees are typically a percentage of revenue associated with users clicking on recommendations, so this trend is a huge plus for vendors.

So what shadow could possibly fall on this sunny picture? Marketers strive to deliver a consistent experience to customers, across channels, across customer lifecycle, across campaigns, wherever the customer encounters the brand. Using the converged recommendation/search/ad solutions, the display ads, emails, category pages, landing pages and search results will emphasize similar collections, and each experience is similarly relevant – ads are just as interesting as emails and category pages. This is terrific, just what any brand needs. However, what tools are employed to manage these customer interactions, and who is using the tools? The recommendation-based discovery solutions have their own management console, and it doesn’t manage anything else. The discovery activity creates another structure for campaigns, a silo. Yes, it’s only one silo instead of three, but marketing teams already have too many silos, too many tools, too many gaps between tools, too much overlap among tools. Building a campaign? Well, pull out the Post-Its, because the recommendations and ads portion will be managed in the discovery silo. Who will manage discovery? Will everyone use the discovery console a little, in their campaigns, or would one person be the go-to that everyone else calls on?

What needs to be integrated to create helpful reporting?

Marketers should respond to the changing landscape tactically and strategically. Yes, use these information filtering solutions to improve customer experience and revenue in 2012. Use some of that added revenue to fund development of a technology roadmap for customer experience and for marketing. Longer term, marketing needs to reduce the number of marketing tools it uses and manages.