Inside Multifaceted Search

Guided navigation. Faceted navigation. Faceted search. Faceted browsing. Multifaceted search. Whatever you call it, ignore it at your peril.

It first made its appearance in e-Commerce, and it is still a great way to find a product based on its characteristics, but it is starting to make its way into mainstream document search. This talk, given at the Enterprise Search Summit, showed how IBM implemented multifaceted search in 2003 for its ThinkPad line of notebook computers (now sold by Lenovo).

The before picture

In 2002, IBM’s Web site navigation required customers to choose the ThinkPad “series” the customer wanted (such as “T Series” or “R Series”), but many customers did not know the correct series. This kind of navigation is a standard one in the industry, where even “Build Your Own Computer” experiences require you to choose a model or a series to start from. While this can work for some site visitors, it is too challenging for many others.

IBM first focused on modeling its visitor behavior, using the Web Conversion Cycle methodology explained in the book Search Engine Marketing, Inc. so that each step in the conversion process could be modeled and measured. The metrics from 2002 showed there was an opportunity to improve the sales conversion rate by increasing the number of visitors that found the product that matched their needs.

Enter multifaceted search

Multifaceted search, with its ability to show all of the characteristics of a ThinkPad (its price, memory size, disk size, weight, and many more), allows searchers to choose the characteristics most important to them. As soon as they do, the list of ThinkPads matching their criterion remain in the search results and the facets are updated to show only valid choices of what’s left.

So, if the first selection is from the Price facet (“Under $1000,” for example), then the remaining facets are updated based on that choice—perhaps there are no ThinkPads under four pounds for that price, so that choice is removed. Contrast that with the typical “advanced search” where searchers are invited to choose their criteria up front—they ask for a ThinkPad of less than four pounds for under $1000 and get the dreaded, “not found.” When they have specified more than two criteria, they don’t even know which one to back off to get some results. With multifaceted search, they just pick their facets in the order of their importance to them. Invalid combinations are never shown.

When IBM added multifaceted search to its ThinkPad e-Commerce experience, the results were immediate. Conversion rates jumped, with customer surveys revealing that multifaceted search was the biggest reason for their purchases.

Beyond e-Commerce

Multifaceted search is important for more than e-Commerce. IBM is beginning to use it for other kinds of searches, such as for downloads and even some kinds of technical documentation. More and more, multifaceted search is breaking out its e-Commerce mold and becoming part of a general purpose search experience.

Download the complete set of slides for this talk on multifaceted search.