The MacArthur Foundation, among whose fundable ‘causes’ is credibility in online information, is funding the start up costs of the Reference Extract project, that the New York Times characterized as “Google if built by librarians”. Michael Eisenberg, emeritus Dean of the UW iSchool, is leading this effort involving David Lankes of Syracuse University, and a team at OCLC led by Jeff Penka.
I spent a day last week among a dozen or so invited advisors brainstorming the direction and development of this system. The group included librarians, technologists, entrepreneurs, venture capitalists, and researchers, and the diversity of their reflections mirrored the variety of their backgrounds.
The effort is intended to augment, rather than displace, search capacity on the open Web, adding the credibility of librarianship to the mix of ‘special sauces’ that distinguish commercial search engine relevance ranking today.
Librarians have been known to turn up their collective noses at open search, impugning the quality of search results, and a plethora of online dross bolsters the position. But it is mostly good enough for most of us, most of the time. Will reference-librarian links improve that quality?
It might be more useful to think of librarian-vetted links as another cut on relevance ranking… useful primarily in circumstances when ‘good enough’ is not quite good enough. Still, such results are more likely to live comfortably within existing search environments, rather than in competition with them.
Business models comprised a significant part of the day’s discussion. Early in the discussion, one of the venture capitalists in the group posed the questions asked of any new venture:
• Who are the buyers?
• What are the compelling reasons to buy?
• Does the approach enjoy a sustainable competitive advantage?
It is commonplace for buyers and payers for a given service to not be the same people, and this will almost certainly be the case in this venture. It is hard to imagine a search service garnering by-the-drink or subscription support in today’s ubiquitous search environment. This leaves advertising models (generally unappealing to public institutions such as libraries), foundation support (not a strong contender for long term sustainability), and the perennial favorite, subsuming the operational costs (losing money on every transaction, and making it up on volume).
The latter model we know well, and there are reasons it may be particularly effective in the present case. But why should the library community take on yet-another cost in a time of flat or declining support? Because future library use will depend over the long run on transferring brand equity from the physical to the digital world.
One of the UW librarians present at the meeting evoked a statistic related to their implementation of WorldCat Local, of which UW has been an early adopter. Inter library loan fulfillment doubled almost immediately after launch – users found things and asked for them… twice as often as before the launch. It is hard to ignore a doubling of a basic service demand that translates into more satisfied patrons.
Thus, it is not outlandish to imagine substantial return on a reference/search/relevance-ranking system that amplifies the value of mediated search – generating results that enjoy the imprimatur of an information professional. A successful service will raise the profile of what librarians have to offer web users, and return substantial dividends to libraries, helping to achieve credibility at web-scale that we have enjoyed in the physical world for a century and more. And the costs of such a system will be substantially mitigated by mining existing reference workflows, not creating new ones. Another example of making library data work harder.
Librarians enjoy a degree of public trust that is rare among any group of professionals. As is evident in the OCLC report Perceptions of Libraries and Information Resources, this trust translates to a brand equity that can be captured in a single word: books. The promise of RefEx is, in part, to extend the brand equity of physical information formats (books) to digital information credibility on the Web. No single project will accomplish this, however well it succeeds. But to fail to effect this transformation over time is to allow decay of public trust in proportion to the decline of the impact of print relative to digital systems.
So, the value proposition for imbuing online information with greater ‘credibility ‘ is to convert the brand-equity of traditional book-bound trust into its digital equivalent – a brand-promise for managed, evaluated information stores that can be counted upon to meet needs when just-good-enough… isn’t. This is a promise librarians must keep.
George Washington's profile (Mt. Rushmore) from the flanks of the monument. Taken in October on our trip west from Ohio to Seattle.