Enterprise Search: A Problem of Relevance to the Users

January 23, 2015

I enjoy email from those who read my for fee columns. I received an interesting comment from Australia about desktop search.


In a nutshell, the writer read one of my analyses of software intended for a single user looking for information on his local hard drives. The bigger the hard drives, the greater the likelihood, the user will operate in squirrel mode. The idea is that it is easier to save everything because “you never know.” Right, one doesn’t.

Here’s the passage I found interesting:

My concern is that with the very volatile environment where I saw my last mini OpenVMS environment now virtually consigned to the near-legacy basket and many other viable engines disappearing from Desktop search that there is another look required at the current computing environment.

I referred this person to Gaviri Search, which I use to examine email, and Effective File Search, which is useful for looking in specific directories. These suggestions sidestepped the larger issue:

There is no fast, easy to use, stable, and helpful way to look for information on a couple of terabytes of local storage. The files are a mixed bag: Excels, PowerPoints, image and text embedded PDFs, proprietary file formats like Framemaker, images, music, etc.

Such this problem was in the old days and such this problem is today. I don’t have a quick and easy fix. But these are single user problems, not an enterprise scale problem.

An hour after I read the email about my column, I received one of those frequent LinkedIn updates. The title of the thread to which LinkedIn wished to call my attention was/is: “What would you guess is behind a drop in query activity?”


I was enticed by the word “guess.” Most assume that the specialist discussion threads on LinkedIn attract the birds with the brightest plumage, not the YouTube commenter crowd.

I navigated to the provided link which may require that you become a member of LinkedIn and then appeal for admission to the colorful feather discussion for “Enterprise Search Professionals.”

The situation is that a company’s enterprise search engine is not being used by its authorized users. There was a shopping list of ideas for generating traffic to the search system. The reason is that the company spent money, invested human resources, and assumed that a new search system would deliver a benefit that the accountants could quantify.

What was fascinating was the response of the LinkedIn enterprise search professionals. The suggestions for improving the enterprise search engine included:

  • Asking for more information about usage? (Interesting but the operative fact is that traffic is low and evident to the expert initiating the thread.)
  • A thought that the user interface and “global navigation” might be an issue.
  • The idea that an “external factor” was the cause of the traffic drop. (Intriguing because I would include the search for a personal search system described in the email about my desktop search column as an “external factor.” The employee looking for a personal search solution was making lone wolf noises to me.)
  • An former English major’s insight that traffic drops when quality declines. I was hoping for a quote from a guy like Aristotle who said, “Quality is not an act, it is a habit.” The expert referenced “social software.”
  • My tongue in cheek suggestion that the search system required search engine optimization. The question sparked sturm und drang about enterprise search as something different from the crass Web site marketing hoopla.
  • A comment about the need for users to understand the vocabulary required to get information from an index of content and “search friendly” pages. (I am not sure what a search friendly page is, however? Is it what an employee creates, an interface, or a canned, training wheels “report”?)

Let’s step back. The email about desktop search and this collection of statements about lack of usage strike me as different sides of the same information access coin.

A single user wants a system he or she controls. The objective is to locate information needed to do a job, answer a business question, or look up a fact, among other “enterprise” look ups. The single user is looking for a solution because the alternative does not meet his or her needs. Thus, enterprise search is not working. The user is willing to go off the reservation in order to locate needed information. This type of action causes nervous stomachs among security wonks. Accountants get queasy when the costs of individual systems and support are mixed into the budgets. Other users may want access to a similar system or to the single user’s system. In short, lots of challenges flourish when the enterprise solution is non existent, does not work to the satisfaction of a power user, or is ignored by Gen X, Millennials, and Gen Y employees who prefer Facebook or a WhatsApp query to a pal.

The squeal from the squeezed enterprise search professional wanting a “guess” is significant for three reasons:

First, users in this case seem to be finding information without making the official enterprise search system the hub of findability. This is a problem in an organization that wants proof that an investment is making a contribution. Without solid return on investment metrics, enterprise search managers have to rely on usage figures. With low usage, the bean counter formulates a hypothesis: Enterprise search is not worth the money. That’s one source of the squeals: Money.

Second, users ARE getting information. The methods are likely to include the work around my email correspondent is exploring. Others just use Google, social media, or whatever method works for a particular individual. The result is that information collection reverts to the systems and methods of everything EXCEPT the enterprise search system. If an organization wants to be organized, the actual information collection process is essentially disorganized. Whether the disorganization works or not is irrelevant. A Dilbert boss wants organization. Thus, the organization becomes even looser than it may already be. Good for some businesses, not so good for other businesses.

Third, the way in which enterprise search is deployed in my experience boils down to a few basic steps: Find a software solution. Index the content that seems logical to index. Make the system available to users. The problem is that delivering what users want is routinely outside the capabilities, resources, and knowledge of those responsible for the information access system. Procurement teams turn to vendors who deliver what has been widely available for decades: keyword search with a few easily deployed bells and whistles. Vendors continue to sell software that performs the types of operations provided by IBM STAIRS IIII in 1970.

But increasingly enterprise workers want access to behind the firewall information and to YouTube, Facebook, Snapchart, WhatsApp, and Flickr or Google image search.

There is a mismatch between what mainstream vendors of keyword search sell and sell really hard and what users are doing in their every day work life.

Net net: Enterprise search vendors are selling bias ply tires in a run flat world.

No wonder the squeals are surfacing in naked rawness on LinkedIn’s enterprise search forum. The butcher is already at work in the pig pen.

It makes sense that open source solutions, good enough search, and whatever finding tool is included with a database systems or Windows are fertile ground for work arounds.

I expect more howls and “guesses.”

Guessing won’t work. The solutions may be next generation information access solutions. Today’s enterprise search vendors, “experts,” and search managers have to work overtime to implement NGIA solutions.

That’s difficult. That’s why the pathetic request for “guesses” exposes the shallowness of traditional enterprise search solutions.

Stephen E Arnold, January 23, 2015


One Response to “Enterprise Search: A Problem of Relevance to the Users”

  1. harveyxclt.mywapblog.com on January 24th, 2015 12:45 am


    Enterprise Search: A Problem of Relevance to the Users : Stephen E. Arnold @ Beyond Search

  • Archives

  • Recent Posts

  • Meta