Search Engine Adaptation for Populations
- Omer Ben Yehuda
- Oct 1, 2006
- 5 min read
Today, we integrate a search engine as a given in every external/internal organizational knowledge service (websites, portals, communities, content/document management systems, etc.). This is the most common and preferred search method.
With high demand comes high supply. Today, there are many types of search engines. Some search the entire document, while others search only titles or keywords. Some use Boolean principles, while others incorporate different search styles. Organizations struggle with the question: Which engine is most efficient?
Before detailing my recommendations, I'll note that these recommendations are in the field of knowledge management and aim to instill principles for efficient use of content and information. For technical or technological suggestions, it's recommended to consult IT experts.
When implementing a search engine in a knowledge service, one shouldn't automatically choose the most sophisticated and complex engine. What determines a search engine's efficiency is its adaptation to the needs of the population it serves. User needs and characteristics should determine the choice of a search engine, not the technology or knowledge provider preferences.
Below are the most well-known search styles and their suitability for target audiences:
Full-Text Search
Search that locates the requested search term in all parts of the text, both in the title and content (Google, for example).
Such searches tend to find many results (depending on the search range) and are therefore suitable for those seeking to find as much information as possible in the requested field.
Populations that might need such a search method are research and academic people studying a particular subject or procurement people looking to compare topics.
This engine assumes that users have the time and patience to browse search results and create synergy between them. This engine is not always efficient for those seeking a quick, precise, and focused search.
Title Search
Search that locates the requested search term only in the title of information pages.
Therefore, it is especially effective in knowledge services that include information such as guidelines, procedures, laws, etc. The more static and unchanging the information, the more efficient this engine will be.
Users become accustomed to existing information pages and can use their names to locate the necessary information. The name of the law/procedure they're looking for (or part of it) usually includes the search term they need.
This is also a convenient method for information updaters who don't need to plant keywords as they are already integrated into the title.
The disadvantage of this engine is its inflexibility - a word that doesn't appear in the title cannot be used as a search term.
Keyword Search
An engine that allows planting appropriate and unique search terms on each information page.
In this way, to help users locate information, one can incorporate associative words from their world and language in addition to words from the title or common words.
For example, an insurance company has a procedure for "collecting payment from a customer." The operational systems indicate that the customer owes by placing a "flashing light" next to their name. Because of this light, users commonly refer to debtors as "flashing lights."
A keyword search engine will allow the procedure to be found using the terms "debt" and "flashing light."
Planting the right words appropriate for users on each information page gives the keyword search engine the potential to display the most accurate, focused information in the shortest time. Therefore, it's suitable for populations lacking time when they must reach specific and relevant information at the highest speed. Such populations include service providers, sales representatives, and employees measured by output.
This search method also has two disadvantages:
The more users there are, the more associative words there are, and there may be an overload of conflicting search terms that create "noise" results when searching for focused information.
Many organizations err by incorporating irrelevant search keywords. For example, on the "payment collection" information page we mentioned earlier, the word "Avi" might refer to the procedure owner.
Moreover, we don't always know what the organization will need tomorrow. We attach an information page, consider its current context, and develop accordingly. However, tomorrow needs to change; the market changes and the indexing is no longer appropriate.
Attribute Search
Search engines allow users to locate full and focused information on a specific topic based on its characteristics (for example, file type, date created) or parts of it (keywords, fields with fixed values).
We already mentioned them in last month's journal and suggested a way to enjoy their benefits.
These engines are paradoxical. On the one hand, there is no more efficient engine for reaching the most focused information the first time (minimum time). Still, on the other hand, users perceive them (unjustly) as lengthy and tedious and do not (to put it mildly) include them among their preferred tools.
The "bad" reputation of advanced search engines stems from two reasons:
For this engine to realize its full potential, fieldwork must be done, including classification and cataloging of each page in the knowledge service and research to determine what parameters are appropriate for users as search parameters. This preparation work and the need to overcome challenges causes many organizations to do "half the job.” As a result, the use of advanced search engines is disappointing and not focused enough. The result is not worth the effort.
An advanced search engine requires the user to be active and think during the search. We humans are "programmed" to conserve energy and tend to go for the "easier" solution, which is the regular search engine where we only need to write a search term (without thinking too much...).
Other advanced search methods require dedicated search engines:
Some expand the search range, for example:
Linguistic search (morphological, Soundex, etc.)
Federated search (search above additional search engines)
Some narrow down the content:
Metric searches (based on the distance between parts of search terms)
Searches incorporating thesaurus (synonyms, hierarchical word trees, etc.)
Context-dependent search (like ClearForest)
Some organize the results described in the "Search Engines" article in the July 2002 journal.
Summary
There are many search methods and diverse engine families. Each has its strengths and weaknesses.
Before "rushing" to use the most sophisticated engine or compromising on an engine automatically supplied with the technological tool, it's recommended to remember that the search engine has significant weight in influencing the friendliness of the knowledge service.
We must adapt the engine to its users and their needs, thus ensuring search efficiency and, more importantly, knowledge efficiency.
Comments