Recommendation Engines: Future for Retailers and Content Providers?

I’ll start this article by mentioning a news item that most of you have probably come across in your feed readings: The purge of Digg’s top Diggers. The problem is that top diggers were holding a monopoly on the site’s activity, making it almost impossible for the social news aggregator to organize social networks within the larger social news site’s community.

As Steve O’Hear mentions on this ZDnet article:

The result is that the site’s content becomes even more relevant and social to its users, while at the same time providing even more hooks to advertisers.

If I get this right, the better you can recommend items to visitors, the better you can advertise to those visitors. There are a lot of recommendation engines out there, most of them holding out the promise to find for you the stuff you don’t even know you want.

However, in the recommendation engine area, the business model seems obvious, but the ability to build the technology that will seduce users is less obvious. Here, I would like to show how recommendation engines successfully implement advertising solutions in their product.

On Twitter, someone told me yesterday:

Plus Pandora has the coolest targeted mobile ads going on the iphone. Possibly, greatest app ever… (tweet)

Pandora is one of the most popular apps on the iPhone. Their music-matching technology is perfect to automatically build audio playlists when you are on the go. However, we all know how doomed the music industry is feeling these days, so how could Pandora define a relevant ad revenue model?

Pandora was facing two challenges: How to place ads on the small iPhone screen? Which advertisers to target based on what criteria? The result is quite a success so far, as you can see here.

Another recommendation engine that competes directly with Aggregate Knowledge is France-based Criteo. The startup has developed a rich data infrastructure to serve up recommendations based on behavioral patterns, and swiped a $10 million fund in January of this year. The company has also integrated an advertising system in their recommendation results. Here is an interesting quote I picked from their blog posts announcing their new advertising system:

Another interesting feature is the pay per click business model for advertisers, which is (still) quite uncommon in the uprising behavioural targeting market. While the costs of sponsored links are increasingly high, Criteo offers advertisers an alternative with powerful ultra-targeting graphical advertising. In addition to high volumes of clicks combined with a high quality of traffic, we offer advertisers an exposure of their brand that is not possible on sponsored links. Various tests have shown that one post click order generates an additional 3 post view orders in parallel due to memory of the campaign.

aggregate knowledge

Aggregate Knowledge has recently launched the same solution for their clients. The product is called Pique, and it is now the feature product of Aggregate Knowledge homepage. The product is defined as “discovery advertising”. Pique targets retailers and major media Websites. It offers advertisers the opportunity to leverage Aggregate Knowledge’s network and technology to increase attention and traffic to their own items.

This is the most brilliant form of advertising for users, publishers and advertisers alike. Users get to find the content they are looking for; publishers get another page view per click through the recommendation widget, or a small ad revenue to compensate for the lost visit; and advertisers get a finely tuned ad server that guarantees a well-rounded ROI on ad spendings.

Should advertisers go along with Pique or Criteo for their marketing needs (I hear both in the back)? Having a robust technological platform certainly is a key criteria of success for recommendation engines. The size of the network matters too: Aggregate Knowledge has a network of about 100 Websites; Criteo has about 4,000. However, Criteo works with small publishers (bloggers), and Aggregate Knowledge doesn’t.

Aggregate Knowledge ultimate’s strength is its development efforts in the mobile device arena. Mobile devices are carry-ons with small interfaces. The lack of room on the screen requires that navigation be crystal clear, and that the information be accessible. It’s always easier to retrieve an information from a system that knows what you want. The technology that Aggregate Knowledge is developing is playing a key-role in the assimilation of mobile devices into our daily lives.

2 Replies to “Recommendation Engines: Future for Retailers and Content Providers?”

Leave a Reply