Search Vendors Desperately Seeking Relevance
If your name isn't Google or Yahoo, being a search vendor is something like being a pro wrestler in the 1960s, with fervent but tiny crowds following you to dilapidated arenas in third-tier cities. Devotees of enterprise search insist that the technology will pay for itself, but the vendors are hard-pressed to prove it.
Enterprise search should be a no-brainer. These companies, often founded and managed by PhDs, use complex mathematical formulas to ferret out data strewn across a variety of data sources. The claim is that such technologies make it possible to efficiently locate information no matter where it is or how it's been stored, whether in email, instant messaging, or what have you.
In practice, though, enterprise search vendors are hard-pressed to show tangible results outside a few specialized areas, often related to the production of data for regulatory compliance. Up to now, that's left them to drum up enthusiasm over the merits of advanced features such as improved search-result "clustering" or "contextual searches" that don't require precise search terms.
Set aside Google, which owns the low end of the market (and for whom enterprise search still represents less than one percent of sales), and Autonomy, which is making hay with compliance-related products in deals worth over $250,000. What's left is a bunch of vendors trying to close deals in the $50,000 to $250,000 range, with little but variations on a theme to differentiate themselves.
Recently, they've taken to calling their technology "search-based applications," a somewhat lame attempt to highlight the potential uses of enterprise search in conjunction with other applications. That, in turn, is supposed to significantly improve worker productivity.
Exalead is among the most active in trying to rebrand its technology as a companion to other applications, as opposed to just being a better information bloodhound. The same is true of Endeca and Attivio, a new company founded by former employees of FAST Search and Transfer, a highly regarded outfit that was acquired by Microsoft last spring. Indeed, acquisition is probably the preferred exit strategy for any of these vendors. (Hint to EMC: I know a vendor that would like a word with you.)
Other search vendors, like Recommind and Brainware, have created new regulatory niches for themselves. Regulations such as Sarbanes-Oxley and recent changes in the federal code of civil procedures can require companies to produce reams of electronic documents almost upon demand, which puts a premium on the ability to locate what's needed as quickly and efficiently as possible.
IDC analyst Sue Feldman published a report in 2004, since quoted ad nauseam by vendors and analysts alike, to the effect that so-called knowledge workers can lose a company $6 million per year in time wasted looking for data. That's hardly an argument you'd want to bring to your CEO at any time, but especially not in this environment.
Nick Patience, an analyst with the451Group, admits that search is a tough sell. Unlike software used to analyze data and generate reports, "the end result of a search is an ongoing process," he told me. So, unlike "business intelligence" software that produces actionable information, such as an assessment of credit risk at a bank, search tools produce answers that merely help kick off the resolution of business problems.
Patience told me that some IT managers are bringing in search technology under the guise of compliance, but harbor more strategic plans. If that's the case, they'd also better find an excuse to bring in a vendor that isn't specialized in legal search, or they'll miss out the more cutting-edge stuff.
It seems unlikely, but then again, so did pro wrestling, roller derby and stock car racing in the '60s. Enterprise search vendors hope they're not going to suffer the same fate as roller derby.