A new white paper from publisher SAGE, “Expecting the Unexpected: Serendipity, Discovery, and the Scholarly Research Process”.

Serendipity is considered mainly in the context of discovery via automated content-recommendation systems, since the research (a survey and a literature review) was done in the context of the making of the new SAGE Recommends system.

So the report’s not really about serendipity in the wild frontier of academic keyword search on the open Web. There are some interesting observations, however…

  “Serendipitous discovery should be of particular interest to information providers precisely because there is so little precedent; there is still tremendous scope for individual organizations to bring their own priorities and values to bear on how they recommend or otherwise help researchers discover their content.”

  “If discovery is too exacting or too precise, it can end up reinforcing habits rather than exposing students and researchers to new information, sharply limiting the researcher’s view of the world of information. … We might even suggest that there is room for errors and luck in recommendation systems; a serendipitous system that does not include some element of chance is hardly serendipitous at all.”

  “… based on our research, it appears that approaches to encourage serendipity that do not place the content front and centre might encounter problems.” [i.e.: academic searchers want recommendations based on the actual content, rather than on the behaviour or tastes of other system users]

  “The less exciting, but equally as important, corollary to discovery is delivery, or access: providing the patron with the material once they have found it. Given that “the researcher’s discovery-to-access workflow is [already] much more difficult than it should be” (Schonfeld, 2015 $ paywall), improving discovery before solving the challenges of infrastructure and access is perhaps kicking the can down the road. This is not to say that there is no value to tools and solutions that promote discovery within an isolated silo, but their potential is limited until publishers, libraries, and discovery vendors make interoperability a priority.”