Designing for Faceted Search

While I was inundated with conferences a couple of weeks ago, I missed s a nice article by Stephanie Lemieux at User Interface Engineering (a site I recommend in general) entitled “Designing for Faceted Search“. It briefly explains faceted search and offers some usability tips. It’s not quite as comprehensive as my upcoming book, but it’s also free and is somewhat less than 100 pages.

Of course, I’m delighted that she uses a couple of Endeca-powered examples (NCSU Libraries, Buzzilions). She also cites the Financial Times, but links to the (which I believe is powered by FAST, a subsidiary of Microsoft) rather than the recently launched Newssift, which uses Endeca.

Just one quibble: she says that “Just 3 facets with 5 terms each can represent 243 possible combinations.” I suspect she transposed the 3 and the 5. The right number of combinations is 125 = 5^3, since a combination represents 3 independent selections from 5 possible choices.

By Daniel Tunkelang

High-Class Consultant.

5 replies on “Designing for Faceted Search”

I would say that you are both wrong (oooo burn), as in some faceted systems, users would be able to make up to 5 selections in all three facets. such as wanting to see indian OR thai slow-cooker dinner recipes (from the articles chosen theme of example). is that 5!^3? that would be 360 combinations.


I think you mean 2^15 = 32,768 if you’re counting all possible ORs within a facet–unless you mean to count order. Of course, at that point we’re just counting subsets of value, ignoring the fact that they’re in distinct facets.


I found the article okay; but not as deep as I would have like to have seen with real UI/Design recommendations for facets, location, display, disclosure – which it seemed short on.
I did disagree with this assertion: “As users combine facet values, the search engine is really launching a new search based on the selected values”
This is not necessarily true – it really depends on what search engine you are using – some implementations I have seen do not submit another query. One I have actually worked on definitely does not submit another search with selected values.


Fair point. For example, a client-side approach like Exhibit does everything through JavaScript in the browser. And other implementations take a similar approach using Java applets. Of course, one can argue that they do “submit a query”, only that it’s handled locally rather than on the original server.


Good point. I was thinking it depends on the type and size of data your searching and the complexity of the objects that are returned as well as associated assets (like photos in ecommerce). I have a designing for faceted search presentation that focuses just on design/usability factors which does run close to 80 slides – so I realize that you could probably write a good size book on the subject. Maybe I should back away from being too critical.
Btw: Kayak in both alpha and beta versions loaded the entire >300 typical results set into a javascript array which is what allowed for relatively quick sorting and filtering by facets.


Comments are closed.