market research

If you are in the news or publishing industry, you will inevitably ask yourself two kinds of questions about your market: (1) How big is my market? For any given article X, how many readers liked X? (2) How is my market segmented? For any given articles X and Y, how many readers who liked X also liked Y? The answers to these questions are a vital piece of information for your editorial processes, and your business strategy.

But since you are essentially dealing in information, you have a highly differentiated product. No two articles you publish are ever quite the same. This makes market research in your area difficult, and technically nontrivial.

We believe that what you need in order to get better market insights is an analytics package that employs natural language information processing (NLP) technology. Such technology could dig into the meaning of articles, compare them by their degree of similarity, and even find out which bits of content make an article interesting to a given market segment.

robo readers?

Believe it or not, but many of your most important readers are computers. If you publish an online magazine, or scientific abstracts on the web, your most important reader is a search engine. If you publish financial and business news, many of your readers in the future will be automatic trading systems.

Dealing with computerized readers is a great challenge for the news and publishing industries as computers can't, ...well, read. Not yet. But, there are many ways in which NLP technology can already help you make your content more accessible and more attractive to computerized readers. For example, both Reuters and Dow Jones have recently taken over companies specializing in NLP technology, in order to mark up their content for use by automatic trading systems.