Searching for ROI

Several times this year I've read proclamations from journalists and consultants that 2004 will be "The Year of Search." My first reaction was "didn't search already arrive?" Is there anybody with Web access who hasn't heard of Google? Doesn't every personal computer have a file and an attribute search system? And doesn't every intranet and email system have some facility for searching? On reflection, I see three reasons why search has indeed emerged on corporate radar screens. First, the relentless growth of storage and information types is overwhelming us all. Personal computer hard-drive storage has increased from megabytes to gigabytes. Aggregate corporate repository sizes are moving from terabytes towards exabytes (millions of gigabytes). Sooner or later, like the proverbial frogs in a pot of water reaching the boiling point, some grow accustomed to the increasing heat. Smart, survivor frogs jump out of the pot like crazy.

Secondly, these pundits don't mean the individual OEM search systems typically bundled with applications. Using a search system for each of a dozen corporate information repositories is increasingly being viewed as a problem, not a solution. Better to invest in a single comprehensive search system, which reduces licensing, support, and training costs, and can perform "joined" searches across repositories. Lastly, today's search systems are increasingly more sophisticated than mere content-and-attribute searching, routinely including the ability to work with taxonomies and to categorize legacy content based on taxonomies.

So if enterprise search is such a winning idea, why is it still such a hard sell? After several years of flat-to-declining IT spending, many projects are competing for scarce dollars. Some IT organizations have moved under the dominion of finance departments, where the routine question is, "What return on investment will search yield for the enterprise?" ROI is always a healthy question and it is always difficult to answer. To get some guidance, I posed the same set of questions to leading enterprise search vendors Verity, Autonomy, Convera, and TripleHop Technologies.

First question: "What are the two or three best approaches to justifying the purchase of an enterprise search system?" Each vendor approaches it somewhat differently. Verity's media and analyst relations manager, Winifred Schum, suggested finding these benefits by improving knowledge worker productivity and realizing more value from other enterprise applications like CRM and ERP systems. Kris Marubio, Autonomy's marketing VP, suggests "figuring out exactly what your company wants to achieve." TripleHop's president and COO, Mathias Turck, pointed out numerous studies showing that employees waste 25% or more of their time looking for information or recreating what already exists somewhere else. In addition to ROI, Turck listed "Return on Assets"—essentially getting more value from existing intellectual capital—as equally important. He also referred to "fear sales" in the area of compliance. Convera's director of corporate communications, Anna Van Lier, agrees saying that with Convera's Compliance Solution "in a matter of minutes users can comb through several terabytes of email" to determine compliance with an SEC regulation.

Nobody asks for an ROI to prove the value of email systems. They're just part of the corporate infrastructure even though we are all finding it increasingly difficult to find filed mail. Isn't the value of search equally obvious? Do you really need to identify hard benefits like time saved, or is a softer approach like leveraging intellectual capital sufficient? This led to my second vendor question: "Is the trend to justify search systems based on hard benefits or soft ones?" All four vendors agree that both hard and soft benefits are required to build a case for an enterprise search system, but hard benefits still have the upper hand. Turck says that TripleHop can work hand-in-hand with customers to build convincing ROI demonstrations. Autonomy cites an impressive example of a customer hard benefit at an aerospace firm. Using Autonomy's CEN, engineers in different locations and business units discovered they were working to solve the same wing construction problem. They avoided duplicating work and saved millions. Verity's Schum points out that time saved searching can "mark the difference between market success and failure" of products like new drugs, where benefits are measured in millions of dollars per day.

I still find it somewhat puzzling to appreciate the emphasis on hard ROI. I've worked in large and small corporations in many different industries. Never have I found one organization that, at a project's completion, was willing to review projected returns to see if they materialized. Verity's Schum agrees and says, "It's odd that an enterprise search solution still needs justification given the exponential growth of data and content." Still, even though benefits of enterprise search are increasingly obvious, you'll need to search and find proposed benefits to get corporate buy-in.