Search Confusion: Increasing
November 22, 2008
In three separate engagements this week, I had to define some basic terms for the people paying me for my time. The baffler was the meaning of the word “search”. For the first group, search was a find-a-phone number problem. I explained that one could find a phone number by browsing a list of entries, typing a portion of the person’s name, or using a “word wheel” (a truly loathsome term) to browse the entries in the directory. I showed an example of looking for an entry with a geospatially aware mobile device. The Google service worked well. The group understood the options and was able to move forward in the discussion without cross talk about Web search, organic search, and other “types” of search that were not germane to the specific problem this outfit had.
The second conversation was about an end-to-end search situation. The idea was that at one end of an information pipeline were boxes of paper. There were archived files on tapes. At the other end of the pipeline were professionals who had to review the informiaton to determine if there was a product idea nugget that had been lost inadvertently. In the middle of the pipeline were what the group called “inputs”. These “inputs” were Web pages, third party information, and routine office electronic information. The task was to make the information available. “Search” for this group was a business process. The notion of typing a name and getting a result was assumed to be the probelm. The problem was not well defined. Yet the word “search” was used to describe scanning, transforming, indexing, classifying, and making accessible an unknown amount of information. “Search” was a glittering generality.
The third client wanted to eliminate search. The idea was that proprietary content would be blended with Web content from public sources like the state government in Ohio. The idea was to create a report, actually a mash up like a Google Map with real estate listings. The user would be able to type of word or two or just pick from a list of reports on offer. “Search” for this group was a business intelligence function. “Search” really meant, “We don’t want our users to have to search at all. We want to provide answers with a mouse click.”
I concluded that the “search industry” has a big problem. The term “search” is for practical purposes meaningless. Martin White and I in our new study Successful Enterprise Search Managment which will be releaed on November 28, 2008. (More information is here.) define search by its context. The message is that if one talks about search without getting consensus on what the group in a particular context means, the likelihood of a problem is greatly increased.
The three examples underscore the importance of defining one’s meaning of the word search. In reflecting on the procurements that each of the three clients are going to undertake before the end of the year, I thought about the benefits of defining “search” before talking about features and functions:
- Problem definition. With search defined, I surmise that it will be easier to explain what specific problem the informatoin access system has to solve. A good definition nips some types of fuzzy thinking or wishful thinking in the bud. Allowing fuzziness to propagate almost ensures wasted time and effort in the procurement process.
- User needs. The definition of search makes it easy to identify the people to ask about the system’s features and functions. Why ask an attorney about the needs for a chemical structure search system?
- Vendors to tap for bids. The definition, in my opinion, makes it a little easier to narrow the list of prospective vendors. Most vendors have a core competency. Asking a directory database sytem vendor to bid on a business intelligence system is a waste of resources.
- Procurment team efficiency. With a consensus about the meaning of “search”, the procurement team will be able to eliminate the tail chasing in some meetings. No one wants to waste time in meetings where the arguments shift because the definitional foundation is like Jello.
- Lawyers and litigation. With a clear definition of search, the license agreement can be made more precise, maybe include some metrics against which to measure vendor performance. Without a defintion, a contract can be filled with must. When a problem arises, customer and vendor point fingers, end up in court, and only the legal eagles win.
My thought for November 22, 2008: define what you mean by search and get the other person to define search. Agree on a meeting and let the indexing begin.
Stephen Arnold, November 22, 2008
Comments
3 Responses to “Search Confusion: Increasing”
-
- Subscribe to Beyond Search
Feature archive
News archive
-
Nice post. A lot of people–in industry and academia, and on both sides of the negotiating table–confuse search as a problem (i.e., the problem of finding information) with search as a solution (i.e., techniques for retrieving content in response to words entered in a text box). Frustration results when today’s “search” solutions fail to address today’s “search” problems.
Daniel Tunkelang,
Thanks, but this post is a signal of frustration. I am getting fed up with SEO consultants, search consultants, and Enterprise 2.0 consultants. Add to these folks’ blather, the drivel that passes for marketing for some high profile search and content processing vendors. The economy is in severe trouble and double talk, fake enthusiasm for semantic gizmos, and neologisms spawned by a 22 year old Berkeley graduate won’t make things easier for customers who want to solve problems. Plain, accurate descriptions are what we need from vendors, consultants, and addled geese like me.
Stephen Arnold, November 24, 2008
[…] the other day that raised a conundrum: what is *not* search? What do I mean by that? Well, as Stephen Arnold points out in a recent post, “search” can be anything from a “find-a-phone number […]