Content discoverability: serendipity and supporting less experienced users

Share this on social media:

At the Researcher to Reader conference, in London in February, Tom Beyer gave a talk on supporting serendipity in the content discovery process.

This is an important element of platform functionality and navigation that particularly benefits less experienced users. The challenge that publishers and we as platform hosts face, is how can we help those users that arrive at your site, but not at the piece of content that is exactly what they are looking for (if they even know what they are looking for)? You’ve seen it many times before in both scholarly publishing and in other market segments – typically we try to deliver the most relevant, valuable, alternate content through related links.

Related content links are judged on the quality of the links, specifically:

  • Do they make sense in the context of the content that the user is currently on? and
  • Do they truly provide a serendipitous experience?  

Both of these criteria for link quality are important as it is entirely possible to provide extremely related links that are not particularly useful, e.g. in a book the most related links will be to other chapters in the same book but that is ultimately not really what the user is looking for.

There are also many ways of generating the related content links with two main efforts being around document relatedness on the one hand and user behaviour (people who read/downloaded this or read/downloaded that) on the other. Publishers have typically concentrated on document relatedness and have often invested significantly in semantic enrichment technologies to improve the quality of these links.  

Unfortunately, despite the large amount of time and effort that has gone into creating useful related content links, the end result in the platform has often been underwhelming.  More often than not the links are hard to find and decipher. This is fundamentally a UI problem although it may be exacerbated by a lack of necessary metadata.  

So how do we provide more context to help the user quickly understand which links are most likely to provide the results they are looking for? The first solution is to simply improve the UI around the links by increasing their visibility on publication pages and labelling them explicitly, e.g. 'related resources' or 'recommendations'. In addition, many publishers are starting to provide context for users on the reason for the link to be selected. For example, SpringerLink presents related links based on 'similar concepts' whereas SAGE describes providing related links based not only on 'similar concepts' but also managing the relatedness by restricting the size of certain corpora in order to increase unexpected and, hopefully, more interesting discoveries.

Publishers’ 'browse' interfaces frequently suffer many of the same interface problems that related links do. And on top of that, the need to create appropriate taxonomies and ontologies to organise the content is a significant and complex task. In each of these cases sites end up providing lists of content with some minimal organisation and context. And the question then becomes: are less experienced users well served with these interfaces, and if they are not, what could we do better?

I think that ultimately the answer to fully supporting users in serendipitous content discovery is to provide significantly more context and to provide editorially authored guides to the content. This approach is based on availability of subject matter experts to select content, put it  in order, and provide both an overview of the entire set as well as introducing each item and putting it in context. Safari is exploring this idea with their Tutorials product and AAAS is doing something similar with their topic pages. Safari has built an entire user experience around Tutorials to allow users to monitor their progress as they move through each one.  

An obvious objection to this idea is that it requires a commitment to a generation of significant amounts of extra content and the management of a large group of subject matter experts. Of course this content may prove, as in the case of Safari, to be attractive enough that it can generate revenue in its own right. In addition, in the academic world there is already an analogous form of content, the survey or review article, that has many of the same features.  

Unfortunately, for the most part these articles exist primarily within their respective repositories and are not properly surfaced through the user interface to provide the guided path through the content that would be useful to our less experienced users. I can imagine a user interface designed around this kind of content that would provide exactly the guided launch pad that less knowledgeable users require to get their bearings in new subject areas.

The rest of the Researcher to Reader conference provided a rich diversity of points of view on many different topics of interest to the scholarly publishing community. The conference started with a bang with Vitek Tracz from the Science Navigation Group, predicting the death of journals. Although he was quite convincing about how their role as containers was no longer necessary in the online world, he could not deny the fact that journals still provide an imprinteur of quality and, consequently, are important to the university tenure process for researchers.  

Other interesting talks included:

  • ‘Show me the Money’ which addressed the financial implications of Open Access (OA) publishing. Danny Kingsley was especially effective in discussing the administrative costs that universities have to absorb in order to manage the process in the UK where Gold Open Access has become the dominant OA mode; and
  • James Evans, a researcher from University of Chicago, presented on ‘Considering the Sociology of Research’ which examined how research articles are changing under the pressure of online discovery, specifically the increase of writing with longer and more marketing style language, and how researchers themselves are moulding the focus of their work to conform to ‘safe’ areas of study to maximise publication potential.  

The conference also had a set of workshops that allowed delegates a chance to interact and discuss a particular topic in some depth as each workshop met multiple times over the course of the two days. Each workshop leader then briefly reported back to the entire conference on discussions. I opted to participate in the Innovation in Publishing workshop which was ably led by Martha Sedgwick, executive director of product innovation at SAGE Publications.  

The significantly differing views on innovation from the larger publishers ('how to manage innovation without losing focus on their core business') compared with both the small publishers ('how to innovate on a shoe-string') and the technology tool vendors ('innovation is the raison d’être for the enterprise') made for stimulating discussion.  It turns out innovation looks very different from different vantage points.

Tom Beyer is director of platform services at Safari Book Online