Focus on Convera: A Conversation with...Dale Hazel

Trust, Hope and Faith: What to Expect from a Search Tool

It’s rare in my work that I get to address the higher subjects in life, such as Trust, Hope and Faith. When writing about technology and business processes, such greater topics just don’t come up that often. Maybe they should, but they just don’t. Usually.

So it was quite pleasant that my recent conversation with Dale Hazel, Senior VP Marketing for Convera, took just such a lofty turn. First, you need to understand a little bit about Convera. More than 20 years old, Convera (formerly Excalibur Technologies) has about 800 customers worldwide, a quarter of which are in government. And breaking down their government customer base, the largest chunk is in the intelligence community—DoD, law enforcement, etc.—in both the US and abroad. The marketspace? Mission critical search and retrieval!

On the commercial side of their business, a recent uptick in sales of RetrievalWare 8.0 (Convera’s primary product family) among their investment banking/financial services, regulatory compliance and life sciences/pharmaceutical customers, suggests that Convera plays in some pretty sensitive areas.

So it stands to reason that security plays a large role in the day-to-day world of Convera. But trust? “RetrievalWare has always been good at optimizing recall,” explains Hazel. “You want to be sure you don’t miss anything in a repository that could be relevant to your query.”

The Trust Factor

“There’s a lot in common between intelligence gathering and things like the discovery process in, say, pharmaceutical research,” Hazel continues. Likewise, Hazel explains, applications in the financial markets such as compliance, fraud detection and accounting forensics share many of the same technology demands—trustworthy intelligence gathering, no-rock-unturned search and retrieval. “These commercial requirements are very closely related to the demands placed upon technology that serves the intelligence and government areas,” Hazel says.

So this is one way in which “trust” became part of our conversation. There’s a high premium placed on “trust” when the potential result of a query is likely to reveal the discovery of a regulatory violation, the prevention of a product liability or a plot against homeland security.

But there’s another level of “trust” that applies to the implementation and adoption of any technology tool. It’s the trust that end users develop when the tool does what it’s supposed to do. That’s a particularly tough ticket for search and retrieval products, since, in many cases, the typical RetrievalWare user doesn’t know what it is he or she really wants to find. Convera terms it “discovery.” It’s central to most KM theory, and can be summed up as: “I need to know what my company knows, but doesn’t know it has.” Many wags refer to this problem as finding a needle in a haystack. I think it’s much harder: it’s more like finding a needle in a needle stack.

So, facing this level of demand on the ability to retrieve information, one arrives at a familiar crux-point, a place that today’s technology decision-maker often finds him or herself: Do I spend money on new technology? Or do I focus on existing systems and try to derive the value from them? Convera can arguably answer from either point of view. “One of the ways we look at search, categorization and discovery technology is that it really can be the integration fabric across many silos of information,” says Hazel. “You’ve got all these enterprise applications, and archiving systems that put out or hold data in hundreds of formats. Plus they each have their own security mechanisms to enforce, etc.

“We can plug into over 200 formats (for example, groupware platforms like Notes or Exchange) or document management (like Documentum) or portals from BEA, Oracle or IBM.” Hazel continues: “Not only do we connect to them, but we enforce the security measures that are in place. So from a single platform, technologically, we can look for information from wherever it may be stashed away”... leading, no doubt, to satisfying, accurate results, and yes, trust in the system. “When people trust the technology, they use it more. Our view is that to achieve a return on investment in large enterprise deployments, you’ve got to get both rapid and complete user adoption. This is where most ROIs in technology are lost. People don’t use the systems that have been deployed,” says Hazel, adding, “when users don’t trust the systems, they look for alternative ways to find the information they need.” In this case, ROI will be determined by how many people in an organization are actually able to save time.

Hope Springs

Then there’s “hope,” and it can be derived from understanding the difference between “search” and “discovery,” a distinction with which Convera seems very comfortable. Basically, with “search,” you know what you’re looking for. But discovery is much more complex, and can sometimes lead to new revenue opportunities (further enhancing ROI) or the fortunate revealing intelligence clue.

RetrievalWare includes a powerful technology that can, I reckon, create a level of hope for the user. Called “dynamic classification,” it is, in essence, a means to look through multiple repositories of information from several different perspectives to arrive at difficult-to-see relationships, and to then “connect the dots” among them. Hopefully (there it is again) the relationships among these pieces of information can be determined in such a way that might never have been possible using a different kind of search tool.

An example: In a homeland security application, you can layer factors together, such as “terrorist organizations” plus “geography” plus “type of weapons” to arrive at the commonalities of information that might be held in many disparate locations and formats. Dynamic classification provides the key data points that are most relevant, considering the various query contexts you use. And viewers within RetrievalWare can slice and dice the results into whatever table or matrix view is best for the user. “In many search interfaces, you get back a list of results, but it’s not always clear that the most relevant response is really near the top, because the user is operating in multiple contexts simultaneously”, explains Hazel. “But if you compare results with other axes of interest, it’s much easier to arrive at what’s really relevant. You are much more quickly able to identify the things that are relevant to the multiple perspectives you (as a human) have on a particular subject.”

You Gotta Have It

“Faith” in a technology comes when it responds in appropriate and familiar ways, speaking the language and displaying sufficient amounts of shared knowledge. Convera accomplishes this basis of faith by creating taxonomy and classification “cartridges,” pre-designed for certain vertical market applications. For example, the pharma cartridge is pre-filled with tons of drug-related research data, jargon, alternative nomenclatures, and such. So when the tool interacts with the researcher, it can respond in a language familiar to the researcher, and can find interrelationships that might have been hidden from a less “educated” search tool.

So, without stretching the metaphor too far, what a search and discovery tool provides, at the end of the day, is far more than just “search” and “retrieval.” It should strive to meet a higher standard.

Andy Moore has held senior editorial and publishing positions for more than 25 years. As a technology writer and editor, Moore speaks with dozens of senior executives and industry experts each month. In his role as Editorial Director for the Specialty Publishing Group, Moore oversees the contributions to the series as well as conducting market research for future topics of interest for the series.

Moore was the editor-in-chief of KMWorld Magazine and is now its publisher.

KMWorld Covers
for qualified subscribers
Subscribe Now Current Issue Past Issues