Discovering Hidden Information Treasures in ECM

The specialists recognized that the ways of organizing information were constantly changing, yet no system—for example, a taxonomy-based one—could handle the flux. The seemingly impossible problem to solve was this: there was no way to predict in advance precisely how any given information might get reused, since each new project was unique. A newly discovered ocean trench in the Gulf of Mexico might share a geological similarity to a peak in the French Alps, making it possible to inform a new drilling project with existing data. Yet until the new trench was discovered, no one could have known to classify their topographies in a way that they might eventually be juxtaposed. This problem is the complement of an information consumer's dilemma: how can he retrieve information when his precise task is unlike any other that has come before?

Information-access platforms offer a solution. Though the search engine has become the most familiar approach, information access has advanced far beyond search. Search engines work like wildcatters sinking an oil well: they have a hunch on a location, then stab in the dark, hoping for a strike. For example, to find information about how the French Alps are similar to our newly discovered trench, we might search on "French Alps" to discover which of our tens of thousands of reports might mention them. Typically we either get back no results (perhaps we've over-specified our query, or a classifier used different terminology), or more commonly, we get back too many results to make sense of.

Information access works more like ground-penetrating radar: it summarizes all of that hidden information in a way that can guide our drilling before we begin. Through a combination of techniques like search, guided navigation, charts and reports, it shows us where we might have the most success exploring. As with oil exploration, the map is not the world, and as soon as we forget the map is not the world, we are lost. An iterative process is required, where we step closer, constantly remapping, updating our host of summarizations that will help us determine our next steps.

The oil seeps sought by drillers near Pittsburgh and the migratory paths hunted by the whalers are the stuff of wayfinding, the set of techniques people and animals alike use to find something, like food, whose location they don't yet know. The process has been well studied, and follows predictable steps: people follow a scent, looking for fresh cues that the trail is growing hotter or colder. They constantly re-evaluate and concentrate their efforts in the most fertile directions.

People foraging for information face the same problem, and need a powerful set of wayfinding tools to support them. In this process, they must:

• Predict: Which available navigation or search might get me closer?
• Evaluate: Does the rich feedback of the UI tell me I've gotten closer to my goal?
• Adapt: Now that I understand the terrain and types of options available, can I better formulate my goal?
• Iterate: I follow the information scent, stepping towards my goal; and
• Revise: Unexpected discoveries along the trail help me formulate new goals.

Search engines alone don't support this full range of behaviors. Information access succeeds by helping each individual follow the idiosyncratic path he or she must take to discover unpredictable information.

How could BHP Billiton's knowledge management experts help searchers complete tasks that were impossible to predict in advance? The answer was also in wayfinding. Instead of predicting the precise path people would take to reach information, they instead needed merely to expose evidence of an information scent. This came in the form of tags and metadata; structure—explicit or implied—that could help searchers evaluate the suitability of any particular set of data. From there, the information access platform was able to summarize that evidence into constantly updated "maps" that could inform our searchers as they completed a goal.

Big Treasure in Small Decisions

We've been talking about big treasure: oil. There are also comparable big treasures in ECM systems to be found, like forgotten intellectual property, or big costs to be avoided, like regulatory compliance mandates. For example, one banking customer of Endeca's estimates it is going to save tens of millions of dollars by making just a single set of regulatory documents easier to discover.

The biggest treasure in ECM is in the aggregate of all the little treasures to be found. That's because information access makes possible, for the first time, better informed decision-making on ever-smaller decisions. A typical investment in ECM might have its roots many years in the past, and in that time it has successfully captured far more valuable information than has been retrieved. Information access is finally unlocking that value, and as you might guess, the potential is tremendous.

Just one example: IBM reported in the Wall Street Journal that it saved $500 million internally in a single year on its human capital management initiative. Built with the Endeca Information Access Platform, it supports decisions on how to staff consultants on projects. Each staffing decision is complex, and never precisely like any that has ever come before. Yet information access removes guesswork from those decisions, providing rich evidence of the choices available at each step of the process. $500 million represents the accrual of modest savings at each tiny decision in the step, aggregated up many, many times over. The biggest treasure in ECM is found when it becomes affordable to drill for its smallest treasures.

---

Endeca (www.endeca.com), headquartered in Cambridge, MA is a next-generation information access company uniting the ease of search with the analytical power of business intelligence. The Endeca Information Access Platform combines patented intellectual property, breakthrough science and a deep focus on user experience to help people find, analyze and understand information in ways never before possible. Leading global organizations like ABN AMRO, Boeing and Cox Newspapers rely on Endeca to increase revenue, reduce costs and streamline operations through better information access.

Jonathan Bracken is Endeca's Enterprise Solutions Manager for Knowledge Management. Prior to joining Endeca, Bracken worked in consulting for 6 years. First at Deloitte Consulting where he managed business and technical teams in IT implementation projects for hi-tech, manufacturing, retail, life sciences and pharmaceutical industries, and then as an independent management consultant working with a Boston-based footwear client. In this role he managed a major implementation of Business Intelligence and the launch of a customized footwear ecommerce business. Prior to that, Bracken was a project manager for new vehicle development at the Ford Motor Company in Dearborn, Michigan. He has a B.A. in History from Brown University and an MBA from the University of Michigan.