We’ve gotten used to being able to find information online on anything, anywhere in the world, and in under 30 seconds -- from details about medical symptoms to trivia such as the population of Irkutsk or the actor who played Dobie Gillis. But within your organization’s systems, you can spend the whole afternoon trying in vain to track down a single document. Why is it that while Internet search is better than ever, enterprise search often seems worse?
When it comes to creating and implementing an effective enterprise search strategy, the anticipation that end users derive from their everyday experiences on the Web can cast search experiences inside corporate firewalls in a pretty bad light, according to analysts and consultants.
“Your expectations as a consumer have raised the bar for everything in the enterprise -- and that bar is set really high,” said Scott Liewehr, principal analyst at Digital Clarity Group, a content management consultancy in New Hyde Park, N.Y.
Google’s online relevancy model derives from millions of people conducting millions of searches. What terms did others use when searching for information about the old TV show The Many Loves of Dobie Gillis? What links did they ultimately click? Metadata builds up around those searches, pointing to actor Dwayne Hickman as the answer to the question of who played Dobie in the show, which aired from 1959 to 1963.
Find out more about enterprise search strategy and software
Read about who should own enterprise search software projects
Learn about how NASA and Google tackled a major enterprise search project
Discover how the HP-Autonomy deal might be a game changer in search and information management
One of the main issues with setting up an effective enterprise search strategy is that there aren’t millions of people in your enterprise. Check your internal search logs and you might discover that the majority of queries are only entered once or twice.
“There’s a feeling,” Liewehr said, “that with a field this small it ought to be easy, but it’s the fact that the field’s so small that makes it harder. You lose the benefit of the crowd.”
Furthermore, organizations typically don’t have teams dedicated to making everything easy to find; in most cases, many people are creating or capturing and storing content with little attention paid to information governance. At an online retailer like Amazon.com, by comparison, a relatively small group of people is in control of making sure that there’s robust metadata surrounding the content on the company’s website, Liewehr said. For search purposes, he added, that information is as important as the content itself, if not more so. But in the enterprise, making such information available to co-workers is often an afterthought at best, he said.
Fellowship of the search
Web pages, like the One Ring, want to be found. The same can’t always be said of enterprise content.
“Really focused results that hit the mark are the result of organizations … putting great care and thought into the building of a content repository with clear metadata,” said Lynda Moulton, principal analyst at LWM Technology Services in Harvard, Mass. These are typically companies that go to great lengths to make sure their content gets the attention they want it to have, she added. Often, that includes employing content managers with the specific skills to build metadata for an enterprise search strategy that leads to consistency and clarity for the target audience of users.
Very few content management professionals have explained to their superiors that it takes more than simply software to enable good enterprise search, Moulton said. Clients tell her that business executives will often mention their good luck with Google and ask for the Google Search Appliance to be installed internally without realizing the backroom effort that makes Internet content so findable and, therefore, so valuable. “Humans add that value,” she said.
Throwing technology at it not enough
What about auto-categorization tools or text mining software, which respectively use statistical and semantic analysis to elicit the important attributes of documents?
“Throwing technology at the problem is only going to get you 60% to 80% of the way,” Moulton said. “You have to make sure the vocabulary matches the vocabulary of the people doing the searching.” That means building in-house taxonomies and treating them as infrastructure that requires ongoing attention and updates.
Most organizations don’t have a good “knowledge map” pointing toward in-house content experts, or for that matter, a sense of how that content is relevant to the organization’s business efforts. Moulton called that a “human management problem.”
Whether a person is searching for information or using text mining software to find it, companies must make sure the target content is up to date and isn’t indexed redundantly. And before starting, it’s important to know where different content is stored within corporate systems. That’s easier said than done.
Divided systems add enterprise search challenges
Enterprises usually have many different databases and content management systems scattered across different networks. The source system containing the content someone’s looking for might not even be available to an enterprise search engine. If it is, there likely are different levels of security to navigate -- sometimes far too much security. Sorting that all out takes people, time and money.
“I’ve talked to clients who’ve tried every best practice in the book and they’re still struggling,” said Leslie Owens, a senior analyst at Forrester Research Inc. in Cambridge, Mass.
The resources required to even approach familiar consumer search experiences can be considerable, said Whit Andrews, an enterprise search analyst at Gartner Inc. in Stamford, Conn. “Google,” he said, “has enormous -- if not infinite -- resources available to them to deliver these amazing [Web search] experiences.”
The problems of enterprise search are not primarily technical problems, and technology can’t solve them by itself, Andrews and other analysts said. Internet search experiences rest on countless hours of work to make Web pages easily findable. Providing the same level of capability in an enterprise setting is a steep hill to climb, and investments in people and content management processes are crucial to an effective enterprise search strategy.
Mark Clarkson is the author of seven books and hundreds of articles on all things technical, from CAD/CAM to Photoshop. He lives and works in Wichita, Kan., where he rarely leaves his basement and maintains his website at markclarkson.com.