While enterprise search engines are not a new concept, what’s new is the volume of information enterprises are currently dealing with – not to mention the expectations of users.
The technologies available from search companies are much more sophisticated than anything one would find with just a Web search engine. And it pays to have a better search strategy because what you’re trying to do is make workers as productive as possible, said Susan Feldman, vice-president of content management and retrieval software with IDC Corp.
Employees spend roughly 25 per cent of their time looking for information – and that’s a conservative estimate. “If we ask people what they do all week, it adds up to an impossible number of hours,” she said. “There’s no way you can fit it into a 40-hour work week.”
This creates problems, tensions and frustrations for people who just want to get their job done, but just can’t cram everything into a standard work day.
But finding information comprises a set of activities or tasks, rather than just searching – which means you need more tools in your toolbox. If you’re searching for information in the standard sense, it means you have to match the words that you choose to put in the query box with what some expert has decided to use as his or her terminology. “No one has ever said that’s a great idea,” said Feldman.
If you have a lot of documents in your index, you’re going to be able to pull something up with a key word search. “If what you’re looking for is a specific document within the enterprise, you’re not looking for a piece of the haystack, you’re looking for the needle in the haystack,” she said. “And that’s a lot harder to find.”
A few good matches are usually all that’s required in consumer search, but that’s not the case in enterprise search. “The holy grail would be to figure out the idea that I have behind the query and match the idea,” said Feldman. “That means you need to understand the concept in an abstract sense and all the patterns of words that are tied to that
concept.”
Search tools are starting to come on the market that can extract concepts or ideas, which means if a user searches under “hypertension,” he or she will still bring up “high blood pressure.” There’s also questioning and answering, where the search tool pulls up a sentence within a document that answers a specific question, which is starting to appear on health-related sites.
But when it comes to the search engines we’ve grown to know and love – the Googles, the Yahoos, the AltaVistas – there are a few caveats, said George Goodall, senior research analyst with Info-Tech Research. “They have been built and optimized around the structure of the Web,” he said. “There’s a whole linking structure behind the Web that determines certain relevance.”
With page rank, for example, which is the Google algorithm, if you type in five words, it will return all the documents that contain those five words. But when we look at internal document stores filled with Word documents and spreadsheets, they don’t link. “Who writes a Word document linking to other documents on those network shares?” said Goodall. “So that internal structure is very different, but the thing is users still demand the easy functionality of a Google.”
Another issue is that enterprises may just have too many search tools running at the same time, such as Blinkx on the desktop, Google Mini to manage network repositories and SharePoint for basic document collaboration. “Now it’s not a matter of finding the right document but having to go through all these different search tools,” said Goodall.
The use of personal or desktop search in the enterprise is causing a lot of concern for enterprises because of the bandwidth that’s required, said Gregory Dowling, senior analyst with Jupiter Research.
If you have 10,000 employees and all of them are using desktop search and indexing network resources, bandwidth is going to suffer.
“The long and short of it is leveraging multiple tools is an IT drain, and it’s also confusing for the end users because they’ve got to reference two different applications and two different logons,” said Dowling.
Centralization allows users to index local, remote and enterprise-level resources – and that’s the ultimate goal to truly take advantage of enterprise search, he said.