What are predictive analytic technologies? They help recommend actions to buyers or predict customer behavior for businesses. The predictive applications discover and apply patterns in data to predict the behavior of customers, products, services and market dynamics. And the recommendation engines attempt to influence, as well as predict, what customers want, enjoy or need. In the years to come, these technologies are only going to improve and gain broad market acceptance. Here are some of the up-and-coming technologies, along with information about who uses them and their overall strengths and weaknesses.
Collaborative Filtering
What is it? It's a system that matches patterns in the preferences and purchasing behavior of a buyer to those of other buyers. Item-to-item filtering compares an item selected by the buyer with all other items in the database—for example, 90% of people who bought item A also bought item B. User-to-user filtering places buyers in a community of people with similar purchasing behavior and makes recommendations based on what others buy or prefer.Who uses it? Amazon, TiVo Suggestions, Netflix and Yahoo!, for predictive analysis and generating recommendations. Strengths: It's currently the most widely used predictive technology. Weaknesses: It's inappropriate for new or unknown products. It's biased toward high-volume products. Also, it doesn't distinguish between items people buy for their own use versus items they buy for others. Prospects: It's the fastest-growing tool for cultural product recommendations, but it needs further refining in order to personalize recommendations. Search engines such as Google Inc. will be adding collaborative filtering to recommend Web sites.
Attributized Bayesian Analysis
What is it? Don't be scared by the title. It's an attribute-based analysis that looks at why customers behave the way they do by examining the product attributes that customers like or dislike. These attributes can be explicit, such as a movie's rating or a book's title. Or the attributes can be implicit—descriptions such as "thrilling" or "funny." Attributive analysis then predicts or finds other products with similar attributes, using Bayesian inference. Who uses it? TiVo Inc., when looking at new customers or television shows. Strengths: It typically offers more refined recommendations than collaborative filtering does. It's useful when little personalized data is available. The analysis can be quickly developed and may produce surprising choices. Weaknesses: It's difficult to determine and classify attributes. Users often depend on databases of offering attributes that others have prepared. If no database exists, creating one can be daunting. Prospects: It will be adopted in the areas where attribute databases already exist.
Biological Response Analysis
What is it? A family of techniques for assessing biological and neurological responses to content or stimuli, including brain wave monitoring, galvanic skin response and eyeball tracking. Who uses it? Developers of successful television series for children, such as "Sesame Street" and "Blue's Clues," as well as Innerscope Researcher, which assesses responses to advertising and television content. Strengths: It takes the guesswork out of questions involving human response to cultural products. Weaknesses: It's highly invasive, and while it reveals a biological response, it doesn't illuminate the reasons for that response. Prospects: It will probably grow in popularity as we gain an increased understanding of neurological processes—despite ethical concerns.
Social Network-Based Recommendations
What is it? Using social networks to encourage or facilitate recommendation sharing between members. This method assumes that one participant will have similar tastes in cultural products to others already in the network. Who uses it? MySpace Music; MTV SoundTrackStrengths: It can leverage existing social networking sites likes Facebook Inc. and MySpace.com. It's well-suited to new and small businesses lacking a customer base to identify and reach potential customers. Weaknesses: Its recommendations are not scientifically determined. Social network relationships can be superficial and have less impact than other techniques. Large corporations may also have better ways of reaching potential customers. Prospects: It will generally be augmented by other recommendation approaches that rely on data analysis.