One key difficulty of online discovery environments is the assumption that the user needs to see all of the information the system has to present - that there might be something in the entire haystack, the entire universe of information we are providing access to, that the user wants. A simple search, or even a complex one, will return tens if not hundreds of thousands of results in most discovery systems, be it a database or a “discovery layer.” While these results are rankable, facetable, and “browseable” we are intimately familiar with concerns that anything on the second page of results or even below-the-fold are essentially invisible to users. We see this approach everywhere, not just in libraries - searches returning results from all categories of things, resulting in a low precision result set due, at least in part, to the sheer enormity of the catalog being queried. Think searching for products at Amazon, or Target, or Home Depot, or Google Shopping.
One alternative to this approach is something we’re seeing in some innovative commercial browsing environments - don’t give the user everything, just give them what you think they might want. While this approach may seem counter-intuitive or counterproductive (wouldn’t this restrict the discovery potential?) I think this is a direction worth considering. Can we provide discovery systems/tools that engage the researcher from the outset on topics and items we know they are interested in? Rather than starting the discovery process with a search bar and nothing else (a la Google), we can present both a search option and a browse option directed at the researcher’s interests.
Last summer I was wandering around in New York and came across a curious sight. Kate Spade had put up a virtual storefront for their Saturday line. In the storefront - a smattering of actual things you could buy and a touchscreen catalog that allowed you to buy that stuff. But only the stuff in the window. This wasn’t everything offered by Kate Spade’s Saturday line, but the display wasn’t meant to show everything. It was meant to draw the user into the catalog and to engage with the content. Sometimes engagement is worth more than exposure.
So what could this actually look like in libraries? There are certainly many options for personalizing search results and experiences based on previous searches, information from the current search session, or profiles built by the users themselves. As a researcher in the ecological sciences, maybe I don’t want to see all of the databases my library has to offer related to the dramatic arts. If I’m almost always looking for conference proceedings, why must I sift through newspaper articles and trade publications in my result sets?
Maybe it is time to consider providing users with less stuff to look at. Rather than giving our users tools to dig through an enormous pile of information, maybe we should give them a smaller pile.