Factor

Menu

Posts

When “Broken” Search is a Trailing Indicator of Deeper Issues

One of my professors in college had a saying that I often find useful in our practice here at Factor. He used to say “The issue… is rarely the issue.”

We saw this recently in three projects where the stated issue was “search is broken.“ Naturally that meant the organization’s search engine needed to be fixed or replaced.  

But after digging into each project it became clear that “search” was really a trailing indicator of more systemic problems. Fixing or replacing the search engine was not sufficient to address the underlying goals (even though it was necessary in some cases).  Each project morphed into a knowledge management, content strategy, or information modeling initiative where the search engine was really the final consumer of the work that needed to be done.

Good search results are topical, timely, and targeted.  

They are also the end result of a long conversation that spans technology, content strategy, information modeling, business goals, user research, and governance.  Providing users the ability to type a few words in a search box is not the start of a search process, but is really the final step of many different work streams, capabilities, technologies, and information sources.  It is also essential to remember that when a user types those words in a search box it is because the information they are looking for is not available or accessible to them in the existing experience. A good search or KM experience will both meet and anticipate the users needs.

For results to be topical the search engine/search experience must have the “right” content. The content being searched must have the information that the information seeker is looking for.  If the information seeker is looking for a set of brake pads for their car, searching an academic database of automotive history articles will be of little use to the information seeker no matter the richness, accuracy, or quality of the corpus, even though this corpus will include information about cars and brake pads.  

Given that no search experience can be all things to all people the domain and scope of the search corpus (and thus the search results) must match the needs of the searcher.   Walking into a library to find a replacement set of brake pads will be just as useless as walking into a car parts store and looking for a textbook about designing braking systems.  Both locations provide a rich and comprehensive search experience but neither will have the “right” content necessary to support the search context.

Context can impact the scope of the content or information, it can also impact the language of the information, the accessibility of the information, how the information is being presented, the technical level of the information, etc.  Thus understanding user needs and then having a well defined content strategy can be an essential component of a “search project”

A robust search or information sharing experience provides the right information in a timely manner.  In a rich information environment this means returning search results quickly. It also includes presenting content dynamically during the information seeking journey.  For example, in an e-commerce experience this can include showing people products or content related to a product they are viewing. In an intranet sales portal it may include dynamic links to pricing and marketing materials related to a product, customer, or campaign.  Ideally, this information is presented to users before they even have to use a search box at all.

Meeting these information needs can be inferred by an understanding of the information seeking behavior well known to librarians, but often ignored by people designing digital information systems.  Anyone working on a search or knowledge management project would be well served by reading the information gathering research pioneered by Marcia Bates. Of particular value in her work is the understanding that people’s information needs change as they move through the information seeking process. Meeting these needs and the goals of the organization providing the information requires a robust information model that meets a wide set of needs and goals.

Delivering topical information requires an understanding of the information domain.  Timely information requires an understanding of the search experience itself. Providing targeted information requires an understanding of the information seeker and their needs and goals during the information journey.  Users have different goals, access information in different ways, speak different languages, have different levels of domain information, and come with different cultural perspectives. It is essential that the search or knowledge transfer experience afford and address these needs.

A company may make it easy for users to find product specifications on their website.  But if the specifications are provided only as PDFs and in English, they will be very difficult for their Spanish speaking mobile users to use that information no matter how good the search results are.  If research shows that the vast majority of their current and desired users read the product specifications on a full screen and are fluent in English, then this is not a problem. However, it is essential that this context be understood and designed for at the start of a knowledge management project rather than trying to bolt it on at the end.  

Understanding how to target the information is one of the most challenging parts of a search project.  One big reason for this is that most search projects that we have seen are approached as a technical problem and are managed as such.  Understanding the user needs requires a fundamentally different set of skills than those required to implement a piece of technology.

User research, competitor analysis, journey mapping, etc are all necessary, but their value is often overlooked while their cost is always scrutinized.  Whenever I am asked how much these efforts will cost I always remember a major search and CMS re-platforming project I was working on that was going to cost an organization a few million dollars and take considerable resources over the course of a year and a half.  The implementation of these systems went well, there was a new user interface for the end users and new workflows for the content creators adding content to the CMS, search results were delivered quickly. However, there had been little attention to the needs of the searchers.  While everyone was able to claim success on the launch date, the project ultimately failed because people were not able to access the right information in the right format. So, the cost of not spending about 6% of the overall budget resulted in a huge financial and (more important) opportunity cost.

Building a truly rich information experience absolutely requires the right technology.  It also requires the content/information, experience, context, and understanding of the user goals (and user journey).  This is true for e-commerce, intranet portals, educational or government sites. Each of our three clients have a definite need to “fix” their search experience, however, a true fix will require more than a new search engine.  As your organization struggles with search, it may be wise to ask yourself if, “the search issue is really the issue”.

Comments

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.