Recommendation Engines are becoming a more common way for companies to target content, increase revenue, and help customers find content and products in easy and intuitive way. There are several approaches to generating recommendations, but perhaps the one most people are familiar with is the approach seen on the Amazon website which recommends products based upon the logic: Customers Who Bought Item X Also Bought Item Y. This is known as Collaborative Filtering, and involves analyzing order (or clickstream) data and computing correlations between products based upon the people who have bought them. It is also a key part of the Netflix movie recommender and is finding its way into more and more websites and online services.

In addition to Collaborative Filtering, there are other approaches, which generally fall into two categories: statistical approaches (of which Collaborative Filtering is an example) and rule-based approaches. Collaborative Filtering and other statistical approaches have the advantage of leveraging a company's order, clickstream, and other data, avoiding the need to manually create complex rules, and letting the data itself tell you about customer's interests and buying patterns rather than trying to guess what will be most effective. Rule-base approaches give the marketing and business experts more control over what products are offered or which content is presented. Hybrid approaches can offer the benefits of both approaches. Recommendation Engines can also incorporate revenue management strategies to find an optimal balance between offering customers the products that they are looking for with business goals such as maximizing profit or revenue, trying new products, moving inventory, etc.

Intelligent Systems has has been working with Collaborative Filtering and Recommendation Engines since 1995, long before many people were first introduced to the idea in Amazon's "Customers Who Bought This Item Also Bought" feature. In addition to Collaborative Filtering, Intelligent Systems has built Recommendation Engines using other statistical models incorporating additional factors including revenue management metrics, as well as rule-based Recommendation Engines.

Some highlights of our previous Recommendation Engine solutions include:

• Collaborative Filtering
• Statistical Models
• Rule-Based Recommendations
• SaaS Recommendation Service