The most popular and comprehensive Open Source ECM platform
Enterprise Search: Hard to Get Right
Enterprise search is hard. While Google has managed to lock up the market for Internet search, they have been far less successful in delivering a solution that can meet the needs of the enterprise. Internet search and enterprise search are really quite different. Google PageRank was the concept that distinguished the Google search engine — pages that are clicked on more frequently and have more links are typically the most interesting and are displayed on top.
PageRank is a fairly simple idea and it has worked exceptionally well for general search, but in the enterprise, PageRank just doesn’t cut it. For example, search over email can be particularly tricky — being able to know who should be able to see and read which emails is very difficult. In fact, the possibility with enterprise search for sensitive emails to accidentally slip into general consumption, many organizations experimenting with enterprise search explicitly omit emails from their set of searchable documents.
Enterprise search is also hard because corporate documents don’t contain embedded tags like HTML documents do. HTML pages are tied together by hyperlinks, but reconstructing similar inter-relationships among corporate documents is much harder.
Of the companies using enterprise search, many are dissatisfied. A recent survey found that 62 percent of scientists and engineers were dissatisfied with enterprise search options.
There aren’t that many vendors that have specifically targeted the challenges of enterprise search: Endeca, Autonomy, and FAST. And the number dwindled to just two after last month’s purchase by Microsoft of FAST for $1.2 billion. The FAST deal brought Microsoft technology and customers. IBM, Google and Oracle are likely carefully considering their next move in this market.