Search Tuning: Tweak Your Knowledge Management System’s Engine

Are you tuning in to search tuning?You’ve built a knowledge management system to provide your stakeholders all the information they need to meet your objectives. “If you build it, they will come,” right? Well, you can lead them in, but if they can’t find what they’re looking for, they may just leave. And the more times they fail to find what they want, the more resistant they will be to use your system.

It’s up to you to give your users the best chance of finding the necessary information through search tuning. The knowledge management specialists at StyleMatters can help you build the search functionality you need to ensure that your information is found and your goals are achieved.

What is Search Tuning?

Search tuning is the process of improving your search engine so that it serves your users better and increases your success. According to Matthew Lynch, Senior Consultant at Charteris specializing in e-commerce, “Search tuning enhances the relevancy of results, reduces the number of zero results found and helps identify…changes that may be required.”

Tags, titles, weighting and display of results all figure heavily in search tuning. Tagging your products with metadata (whether you offer indie films, infant clothing or instructional articles) helps users find the kinds of things they are looking for. In searches, titles are often weighted more heavily than regular text. That is, if the user searches for “XYZ,” an item entitled “XYZ” will rank higher in the search results than an item that just says “compatible with XYZ” in its description. Exact matches should be weighted more heavily than near matches.

The display of search results in your knowledge management tool merits a separate discussion. Here we will just note that results must present at least some descriptive text along with titles, and the search query should be redisplayed so the user can alter it as desired. Also, recommending or highlighting certain results allows you to push users toward information you currently feature. 

How Does the Search Engine Know Better Than We Do?

Knowledge management pioneers Peter Drucker and Melvil Dewey (inventor of the Dewey Decimal System used in libraries) worked with systems that were not interactive. With such tools, if the user made a mistake or misunderstood the system rules, the search might well have yielded an incorrect result or no result. But current technology can be taught to find what the user wants, even if it is not what he or she asks for.

Matthew Lynch of Charteris writes that “It is estimated that 20-30% of all search terms used on the Internet contain a mis-spelling or a typo.” For example, a StyleMatters client reported that a user seeking “pay-per-click” searched for “paper click.” The search engine doesn’t have ears, so it doesn’t know that these terms sound the same. It only knows that they are spelled differently.

So how can your knowledge management tool be taught to guide users to the information they want (and you want them) to find—even when they search for it incorrectly? Synonyms and predictive search are useful to this end. If you tell the engine that “vidio” means the same thing (for search purposes) as “video”, or that “trouble shooting” means the same thing as “troubleshooting” (and the same thing as “trouble-shooting”), your users can receive optimal results despite spelling errors or varying word forms.

Predictive search allows you to suggest search options that serve your objectives based on user input. For example, if a user types “ca” in the search field, your search engine might suggest “campaign,” “camera” and “USB cable” as search options. Predictive search options change to match user-entered text.

Start Tuning Your Knowledge Management System’s Search

Tuning starts with learning about actual user behavior. Find out what your users search for most frequently, and what searches have returned no results for them. Optimize the former and fix the latter. Make sure to account for acronyms that are frequently used in your business.

Look at your presentation of results. Do you offer multiple results per page? Can you tell where you will go if you click on a particular item? If possible, ask your users how you can improve their search experiences. The more you know about user behavior and needs, the better you can tune your search engine, your results, and your business.

Alan Biehn currently works as a Technical Writer for StyleMatters in Philadelphia’s tallest building. He holds a BS in Audio Recording from The Ohio State University and recently earned his Masters in Environmental Studies at the University of Pennsylvania. Alan has worked as a marketing writer and editor for nonprofit organizations and ESL professionals, as well as for major commercial clients.

Photo credit

Leave a Comment