Search-dominance and NYPL’s home page redesign

An earlier version of this post appears on the New York Public Library’s blog. I’m highlighting it here as a success in applying generalized web UX research to the specific needs and content of an institution.

Among my first tasks upon taking leadership of the Website Department at New York Public Library in late 2012 was to update its old and ineffective home page. Driving the strategy were a few important points that distinguished NYPL and its web site from many others:

  • a huge proportion of visitors used the home page simply as a pathway to catalog searches, leading to book-checkout transactions.
  • With 88 branch (neighborhood) libraries, and four large research libraries, NYPL was a HUGE institution, with a TON of great events and classes going on all the time, that hardly anybody knew about.

I wanted to support that aggregate search-dominant visitor behavior, but also to make people aware of all the other amazing things the library did (and still does). We needed a great, beautiful home page that didn’t make search any harder, but put those events right in your face.

One of the questions asked in early Web usability studies was “do some Web users prefer search to link-based navigation?” In other words, are some users “search-dominant?” In 1997, the answer was yes: Jakob Nielsen determined that over half of users prefer site search to links. More targeted studies conducted by Jared Spool by 2001, however, came up with a different result: it’s not specific users that are search-dominant, but specific websites. This result acknowledges that on some sites (think of and search is usually the best way to find what you’re looking for, but on others (think of newspaper sites, bank sites) you’re more likely to try the site’s main navigation links before you search.

Twelve years later, and in the absence of continuing research, I knew that many Web users had developed an expectation for which method would work better, depending on what they were looking for. That is, it’s not so much the site you’re on; it’s the information you’re looking for that determines whether you search or navigate. You expect search will work better than navigation when:

  • you’re looking for one specific result, but you don’t know what category it might fall into (Lunar Asparagus at;
  • you’re looking for a category of things, and you expect that the number of items in the category is too many to narrow down via navigation (e.g., portable radios at

Otherwise, you navigate first—then try search if the navigation fails you.

This is where NYPL’s needs and content came in. We had a mix of content, some of it better served by navigation, and some better served by search. Our visitors seemed to understand this. For example, about 80% of home page visits go straight to search, with the goal of finding books—for which search is clearly the best method. And with frequent user testing we’d demonstrated, for example, that visitors almost exclusively used the “Locations” navigation item when looking for a branch or research library. Similarly, other visitors used the navigation to go to the Digital Gallery, read a blog post, or find an event listing. More than 10% of traffic on the home page resulted in a click on the navigation bar.

The Library had prioritized a critical institutional goal for the website—to promote the great breadth of Library services. For example, we hoped book readers would make more use of our classes and programs, and that event registrants would investigate collections items relevant to their subjects of interest. We wanted non-users—people who could take advantage of Library services, but didn’t—to be amazed by what the Library offered to them (which is truly mind-blowing). We wanted to do that without making the catalog search any harder to find or use (in fact, we made it a little better).

Essentially, we needed to pleasantly distract our visitors! We absolutely wanted it to be as easy as possible for you to find what you’re looking for (whether with search or links), but we also wanted visitors to find things they didn’t even know they’d be interested in.

The page we launched did all that. It showcased tons of events, classes, recommended books, performances, workshops, videos, and web features, in almost completely uncategorized fashion (people compared it to Pinterest). Our Communications team–who were in the best position to choose what to promote at any given time–used the back-end to update specific highlights or events.  The result? exploratory behavior (basically, a click on any featured content item) doubled.

Sadly, I’m sorry to say that now, NYPL has gone too far. In the effort to promote those events, they’ve now hidden the search. This is probably a transitional page (see, but you can see how the search has been removed, inelegantly, from the vast, gaping hole it used to fill in today’s site header:


Hiding something that 80% of your visitors come for in order to promote the other 20% is not user-centered design, and the library, of all places, traditionally puts users first. C’est la vie, I guess–we’ll see what they do next.

Leave a Reply

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