One on One with Jeff Catlin of Lexalytics

Jeff Catlin, CEO and founder of Lexalytics [1], has over 20 years of experience in the fields of search, classification and text analytics products and services. He has worked for such companies as Thomson Financial and Sovereign Hill Software. We asked Jeff about text analytics and how companies can use this technology effectively.

FCM: Briefly explain text analytics for our readers who might not be familiar with it.

JC: Text Analytics is a collection of technologies used to extract valuable metadata from unstructured content like news stories, blogs or tweets. Examples of the sort of metadata typically extracted include people and company names (commonly called entities), relationships between entities, key concepts or themes that might describe the key point of a story, and, of course, the sentiment/tone of a story or a person or company described in a story.

FCM: How can text analytics help companies understand their content?

JC: Once extracted, this metadata is often used to help power advanced search applications, where you’d be able to ask questions like: “What is all the negative sentiment about my Company’s CEO?” It is also used in various business-specific applications in PR/Marketing where extracted metadata is analyzed to understand how a new marketing or branding effort is being received. In addition, customer satisfaction or customer feedback departments apply text analytics to help identify commonly reported problems or issues.

FCM: Give me a real-world example of how business can benefit from using text analytics.

JC: Lately, one of the best examples of text analytics in business resides in the world of equities trading where text analytics in general, and sentiment analysis specifically, is used to analyze content used in automated trading systems. Systems like Reuters RNSE (Real-Time News Sentiment Engine) are examples of text analytics being used to score news content to help traders make the right trades more quickly and, therefore, improve their overall returns.

FCM: How could you use text analytics to get a jump on an eDiscovery process?

JC: Text analytics built into a desktop auditing tool is an ideal way to inventory content residing on individual desktops to learn what topics are being discussed as well as to find who is interested in those topics. It represents a powerful first-stage filter to determine which desktops or content repositories will require a more detailed examination.

FCM: Where do text analytics and the semantic web intersect?

JC: The semantic web is essentially a fancy way of saying, “an index of the web that understands the meaning of content on the web rather than just the words.” Text analytics is an essential part of the semantic web because the core technology behind text analytics engines is what allows us to understand who is talking to whom and to understand the context of the discussion. All of the major enterprise search engines include text analytics capabilities in their engines to build semantic capabilities into their products, so while the semantic web itself is in its infancy, the widespread use of text analytics is likely to be a backbone of the semantic web.