The Information-Driven Organization
Being called information-driven strikes me as one of the most complimentary things I could be called. I may actually be in love with that phrase. I’ve been infatuated with information since I was young. Looking back, I must have been a real pest, always asking my parents “why.” Perhaps that’s the reason they bought a set of encyclopedias—yes, those heavy print volumes, not the Wikipedia of today. At the time, I thought that all information was contained between the covers of those heavy books, supplemented by more recent information gleaned from newspapers and television news.
It’s odd, I suppose, to envision my 10-year-old self diligently researching a social studies project by poring over encyclopedia entries, but that’s what I did. And even before there was cut and paste, I had to be told not to copy the entry in its entirety into my paper. Plagiarism, you know.
Today’s children grow up in an era of information abundance rather than information scarcity. They know that encyclopedias are not the sole source of information. They’re expected to use word processing, search the web, and create presentations using PowerPoint, even at the age of 10. The internet presents them with an enormous range of information, although some of it is certainly better than others. Being able to discriminate between valid information and unreliable information is a 21st-century skill. Critical thinking and information literacy have never been so important. And that’s true within organizations, not just outside them.
Companies face similar challenges when it comes to information. They, too, confront information saturation. Added to the complexity of finding relevant information is that corporate data is often strewn across numerous locations. Getting a handle on all this information and determining which pieces are valuable and which aren’t is a daunting task. Simply stating that your organization is “information-driven” isn’t sufficient; you have to actually do something with the information. Text analytics plays a big role in being information-driven.
Let me also make a distinction between information-driven and drive-by. I don’t mean drive-by only in criminal terms, as in “drive-by shooting.” You might drive by a house that’s for sale to see if you’re interested, by a school to see if it fits for your kids, or by a restaurant to see if you’d like to eat there. The distinction is that a drive-by is a one-time event, while being information-driven is ongoing.
Understanding Information
Sinequa’s Scott Parker shares my infatuation with being information-driven. He defines the phrase as much more than simply knowing about available information—it encompasses understanding the meaning within that information. As he puts it, “If being information-driven were all about fielding queries and matching on keywords, a simple indexing approach would suffice. The best results are obtained when multiple indexes are combined, each contributing a different perspective or emphasis.” Context is an important element of understanding meaning: Information without context is rarely useful.
Translating that to my 10-year-old self, you need to do more than read an encyclopedia article. You need to look beyond that, consult multiple approaches, and understand what’s going on behind that entry for a comprehensive view of the topic. Obviously, in the corporate setting, it’s more multifaceted than that. It’s not just the written word, since information-driven organizations connect people with expertise and make that expertise transparent. Thus, you need a solution that’s “enterprise-grade.”
Sinequa identifies the components of an enterprise-grade solution as having strong security controls, contextual enrichment, and relevance feedback. Security levels should reflect the organizational structure. The context of language used in your industry and your company should be built-in—acronyms can be particularly troubling. For relevance feedback, the system learns from searches being performed to present more relevant results going forward. Different jobs and different areas of an organization require information tailored to their needs.
Getting to Meaning
When it comes to meaning, Sinequa has several suggestions, based on Natural Language Processing (NLP) technologies. Some NLP functions include automated language detection (particularly important for multinational organizations), lexical analysis (detecting compound words and tagging speech), syntactical analysis (disambiguation and lemmatization of parts of speech), automatic extraction of entity types, and text mining agents. The latter is integrated into the indexing engine to detect regular expressions and to look for complex shapes that give clues about meaning and then normalize them across the enterprise.
Machine Learning and Deep Learning are essential to understanding meaning. If most people in your organization search a particular word or phrase and almost always select a particular use of that word or phrase in your knowledge store, then it’s likely that navigating to that use will return relevant results. This isn’t about rules, it’s about the machine actually figuring out, from analyzing user behavior, what people really want. It extracts and enriches the concepts, drawing out entities and relationships, so they make sense to your users. Plus, they facilitate search result modification and allow for recommendations, both in terms of content and user interactions (collaborative filtering).
Finally, information-driven organizations require good design that is both aesthetically pleasing and understandable. To users, it looks intuitive because it parallels their behavior so unobtrusively.
Information-driven organizations take an encyclopedic view of internal information, incorporating context, determining meaning, maximizing machine learning, and employing intuitive interfaces. They can then discover insights, make informed decisions, and elevate productivity.