Earlier this year CNET published an interview that was originally published by Bloomberg Television: Yahoo CEO Marissa Mayer talks mobile strategy and more. There are two specific points in this interview with which I agree with Ms. Mayer.
2) The Interest Graph - “…your personalization, your context, …to make
sense of the content. It is the Internet ordered for you. ... Now it's
so vast that you can't just categorize it anymore.“
Currently most interest graphs for individuals are defined by a user's previously captured behavior. This behavior could include purchases or web browsing habits. Nevertheless, these variations of interest graphs are limiting because a history of behavior is needed and they assume that a user is already efficiently discovering content that should be interesting to him or herself.
The News Patterns approach for personalization and context is to refine an individual's interest graph with his/her "competition graph." Nearly every business, financial and political decision maker is involved in some variation of competition. In competitive environments, there are archetypical forces that drive what is relevant for players in the environment. Some of these archetypical forces include competitors, buyers, suppliers, substitutes, potential entrants, government mandates and societal trends. Political equivalents include voters, politicians, issues, donors, etc. Many of these forces have thoroughly been documented by Michael Porter of the Harvard Business School.
News Patterns algorithms create a competition graph (interest graph) about a user simply by knowing the employment, investments or campaign of a user. With this interest graph, the universe of potentially relevant news and social media articles are collected, from which patterns are discovered that highlight articles that have a high potential of being interesting for an individual. The result is an efficient discovery or recommendation engine that anticipates interesting content for a user - content that could never realistically be searched for by a busy and often distracted user.
As Ms. Mayer said, users are increasingly looking for the "Internet ordered" for themselves, without the traditional step of them stopping to search for it.
1) Personalization: “…image recognition, voice recognition, translation--these are back on technologies to being able to understand context.”
The News Patterns approach for personalization and context is to refine an individual's interest graph with his/her "competition graph." Nearly every business, financial and political decision maker is involved in some variation of competition. In competitive environments, there are archetypical forces that drive what is relevant for players in the environment. Some of these archetypical forces include competitors, buyers, suppliers, substitutes, potential entrants, government mandates and societal trends. Political equivalents include voters, politicians, issues, donors, etc. Many of these forces have thoroughly been documented by Michael Porter of the Harvard Business School.
News Patterns algorithms create a competition graph (interest graph) about a user simply by knowing the employment, investments or campaign of a user. With this interest graph, the universe of potentially relevant news and social media articles are collected, from which patterns are discovered that highlight articles that have a high potential of being interesting for an individual. The result is an efficient discovery or recommendation engine that anticipates interesting content for a user - content that could never realistically be searched for by a busy and often distracted user.
As Ms. Mayer said, users are increasingly looking for the "Internet ordered" for themselves, without the traditional step of them stopping to search for it.
