In my hard copy version of the Feb. 18, 2012, Wall Street Journal, I noticed an essay about human behavior, "How Habits Hold Us." Jonah Lehrer explained how habits hold in the brain. The example he used concerned an air freshener, and how advertisers made a product successful by hooking into a sequence of behaviors.
I wondered if there is a way to use such psychological "hooks" and "loops" to create an enterprise search system that actually meshed with an employee's information seeking behaviors. The quote Lehrer used to kick off his article was one by William James that I had encountered in 1962 in a required psychology lecture: "Ninety-nine hundredths of our activity is purely automatic. All of our life is nothing but a mass of habits."
With the consumerization of information technology, traditional enterprise systems were facing more and more resistance from users, and new competitors were making headway in the lucrative enterprise search market. With familiar names disappearing inside mega corporations like Hewlett-Packard (HP), Microsoft and Oracle, enterprise search was in flux, and to some of the pundits who frequent the LinkedIn enterprise search forum, disappearing or becoming a commodity of little business value.
Enterprise search is a touchy-feely service. If you have interviewed potential users of an enterprise information system, you probably have heard, "I prefer a system that works just like Google" or, "I want the system to provide just the information I need." Those types of statements make clear that search is a subjective concept.
When you run a survey on SurveyMonkey or another do-it-yourself system, the results make it easy to determine how many of the organization's search users want a Google-style system or want what marketers call "answers."
The problem with this type of pulse taking is that Google delivers anywhere from 65 to 90 percent of the Web search results. The variance is due not to Google's marketing; the swing is a result of different third-party measuring tools. In an enterprise survey, most of those completing it will be Google users. Therefore, enterprise search systems will be measured against the world's most popular search engine. The problem, of course, is that Web Google depends on popularity and links to documents. In an organization, most documents are not popular. Word and PowerPoint documents usually contain fewer links than Web content.
Multiple Search Systems
The notion of answers instead of lists is a reaction against the brutal routine of search-scan-open-browse-close-repeat. A user, accustomed to the effectiveness of iPad applications that deliver specific information about "the weather" with a click, eliminate the search-scan-open of the gerbil exercise wheel. Gerbils seem to run without tiring. Unfortunately enterprise search users do tire and often abandon the enterprise search system.
In my experience, asking enterprise search system users what they want delivers quite predictable answers. The problem is that pointing out what users want and gauging the dissatisfaction with the incumbent search system produces an additional job: licensing another search system. Most organizations have multiple search and retrieval systems, which forces the person looking for information to check different ones. The result? More annoyance.
When I expressed my concern with traditional surveys, a colleague suggested that I check in with Dr. Linda McIsaac, whose work involves a next-generation method of determining employee preferences. I asked McIsaac if she would update me on her methods for obtaining statistically valid data about an individual's or a group's preferences. Her company is Xyte, which uses her method described as "human behavior technology." Her work makes it possible to predict employee behavior and translate it into tangible business results. The Xyte approach, according to the company's Web site, is grounded in neuroscience and psychology.
McIsaac explained, "A group of my colleagues and I were discussing how so much emphasis is put on technology for increasing productivity in business when really productivity is based on people. I know you have pointed out that search and retrieval systems in organizations deeply dissatisfy those who use them. But search is one glaring example of a deeper problem."
Search, as Martin White and I explained in Successful Enterprise Search Management (Galatea, 2009), is a dissatisfying experience for more than half of an organization's employees who use the enterprise findability systems. Regardless of the enterprise search vendor, a high dissatisfaction score sets up a cycle of license-deploy-replace.
McIsaac pointed out that the firm's Xyting Insight method provides a "system science model and a structure that identifies the way people function intellectually by classifying their cognitive information processing capabilities presented as a logical system of relationships among various human capabilities."
She said, "In simple terms, Xyte's system provides a structure based on a proven theory that is predictive of the way individuals think and process information and then act consistently. No other system is predictive."
The example she offered indicated that not all Facebook users access the system consistently. A recent Xyte study revealed that 27 percent of the population had never logged into Facebook and another 20 percent only log in once or more a month.