hiker with binocularsWhen it comes to seeking and retrieving scholarly content, databases like PubMed, Web of Science, and Scopus are well established go-to resources. In the last 5 years, mainstream tools like Google Scholar and ResearchGate disrupted learned publishing and academic library supply chains – right alongside the dawning of web-scale library discovery services like Summon and EDS.

In this rapidly evolving market, it’s great to see that there is already a new generation of cutting edge search tools, challenging the status quo and offering information users new avenues to the research literature. There are three new services in the content discovery marketplace that stand out and deserve our collective attention. They share some key qualities, and yet each offers a truly unique solution to specific use cases and common challenges of today’s scholarly and professional content discovery experience.

  1. Yewno is a VC-funded start-up, gaining notably positive response from academic librarians in its first year. Powered by an “inference engine” using computational analysis and a concept map, it intends to supplement institutional discovery, adding visualized browse options to search results pages to expose relationships between ideas, authors, and publications.
  2. ScienceOpen, launched in 2013, is a privately funded independent start-up that hopes to compliment open-web discovery. Their feature-rich site draws users from mainstream search engines to more precisely search and filter content using their index built by a web of citation-level relationships, referring readers to publisher or aggregator platforms for full-text access.
  3. TrendMD is tackling discovery outside of search altogether, instead enabling links to related articles across publishers within the context of reading a full-text item on a content platform. Launched in 2014, TrendMD is a Y Combinator startup and serves as a classic “onward journey tool” that aims to generate relevant recommendations serendipitously, thereby growing usage.

All three of these new discovery tools aim to offer greater precision in content discovery across the world of relevant resources. They know that comprehensive search is important, but not all users care to drink from the Google firehose. These new services are focused on sharpening researchers’ tools for filtering the noise of information overload from web-scale search – something the current giants of mainstream and institutional discovery software have not perfected.

These new services aim to increase the contextual presentation of links to full-text content, but are going about in unique ways. TrendMD is focused on recommending related content to users reading or evaluating full-text publications, and making cross-publisher linkages appealing to oft-competitive content providers. ScienceOpen wants to harness user surfing the open web and lead them to a promised land with sharper search tools alongside a host of other research tools. Yewno is working to capture institutional users and offer them another way to explore resources subscribed by their library, supporting both educational and research aims.

Remarkably, all three new entrants are leveraging the latest in semantic technologies, text/data mining, and machine learning to iteratively improve as both indexed content and user activity grow. This will likely be a key differentiator, even in the face of the mighty forces behind Google Scholar. However, like their established competitors, the three are focused on information discovery and evaluation, leaving access controls and download authentication rights to the publisher and library systems to manage. As the longest running, TrendMD is already demonstrating measurable impacts on product usage – metrics which will vary slightly for Yewno, given their focus on the institutional library market.

These new tools have taken user-centered approaches to their development of scholarly discovery software. Each organization is born of and rooted in the researcher experience, which sets them apart from many of their competitors. Each one grew out of a passionate wish to address specific research use cases and make immediate improvements to users’ ability to quickly and accurately zero in on the most relevant material at the point-of-need.

Lack of relevance and context is a common pain point for advanced users like graduate students, instructors, and practitioners; those who we might call the “power users” of academic discovery services. To a degree, these new services are somewhat biased toward serving these power users, who are more likely to care about search precision and use filtering features (studies show younger users avoid search filters, as demonstrated by work from Deirdre Costello, among others). This likely points to these three services not being any direct threat to established web-scale library discovery services, which aim to serve a diverse group of institutional information seekers.

TrendMD and ScienceOpen were founded with a focus on medicine and life sciences, which they’re both working hard to extend across educational fields to appeal to multidisciplinary users at all levels. Yewno started out with a broad discipline focus, aiming not only to enhance the discovery experience, but also deliver deeper answers and develop the critical thinking skills to their users. Therefore, these three newbies have quite a hill yet to climb, in their hopes to cover a critical mass of published literature and address the full scope of academic discovery use cases. Comprehensive resource coverage and breadth of publisher participation will be a key challenge for these new services, which continues to plague all of their current competitors. This aim for both breadth and depth could put them at risk of losing some of the magic of their targeted discovery solutions that were launched to improve specific elements of the research discovery experience.

These are exciting times to be in the business of scholarly content discovery – and these three new services represent a host of researcher-centric innovations on the horizon, promising to keep the market lively for years to come.

Lettie Y. Conrad

Lettie Y. Conrad

Lettie Y. Conrad, Ph.D., is an independent researcher and consultant, leveraging a variety of R&D methods to drive human-centric product strategy and evidence-based decisions. Lettie's specialties sit at the intersection of information experience and digital product design. She currently serves as Product Experience Architect for LibLynx, Senior Product Advisor for DeepDyve, and a part-time lecturer for San Jose State's School of Information. Lettie is also an active volunteer with the Society for Scholarly Publishing and the Association for Information Science and Technology, among others.


14 Thoughts on "The Latest in Search: New Services in the Content Discovery Marketplace"

Welcome, Lettie! This is a very useful overview of these new discovery tools, yet I can’t help wondering if they will be of any help at all to humanists wanting to search book content, not just scientists looking for article content. One of the obstacles to the advancement of open access in monograph publishing is the primitive state of discovery as it applies to humanities literature, especially in book form.

And, related to Sandy’s comment, the enduring power of print, which in the arts and social science monographic space globally continues to dominate by a factor of (at least) four to one, with significant signs in some territories that even a very modest rate of p- to e- transition is beginning to slow down. But somehow I don’t think that venture capitalists (and others) are terribly excited by this enduring legacy technology, even if all the available evidence suggests that it remains (by some very considerable distance) the preferred scholarly choice for long-form reading. So I fear the current impasse will continue.

Welcome Lettie likewise!

The big problem with books is that most of these methods rely on full text and there is way too much text in a book. Nor do books have the standard structure that science articles have. A worst case is collected essays, especially by different authors.

For books one would probably need fundamentally different approaches than for articles, especially in the humanities. Sounds like fun.

Great points, all around, David, Richard & Sandy!

I totally agree that arts, humanities, and some social sciences are excluded from some STM-leaning discovery innovations — both in terms of content coverage and in their sensitivity to variations across researcher experiences. That’s what makes the user-centered books discovery research by folks like Anna Faherty so relevant for HSS publishers to read and deeply understand, https://academicbookfuture.files.wordpress.com/2016/06/faherty_academic-book-discovery-full-report.pdf.

And many thanks for the warm welcomes!


Nice to see new approaches and now, with access to earlier literature, it should be relatively easy to see how they work with real historical cases where the correctness and importance of a discovery is hardly in doubt.

The paradigm case in the biological sciences is Mendel’s discovery, reported in 1866, of what we now call genes. There were a few citations in the literature, but the penny did not drop until 1900 when three continental botanists found he had anticipated them (Kilroy was here!). Even then, it took six exhausting years for William Bateson to spread acceptance in the English-speaking world.

Another is the Hering-Butler description of heredity in terms of information-transfer in the 1870s. By a tenuous chain this led to a book by physicist Erwin Schrodinger in the 1940s, and many regard this as the introduction point of information concepts. With Mendel the delay was 35 years. With Hering-Butler it was 70 years.

And, to get really picky, there is the case of Patrick Matthew whose 1831 discovery of the principle of natural selection was belatedly acknowledged by Darwin (to whom we today attribute natural selection), although many who had corresponded with Darwin had long been cognisant of Matthew’s work.

“Powered by an ‘inference engine’ using computational analysis and a concept map, it intends to supplement institutional discovery, adding visualized browse options to search results pages to expose relationships between ideas, authors, and publications.”

In the days before articles, books, etc. became “content delivery systems,” this sort of thing was called “thinking.”

One question: what does “develop the critical thinking skills to their users” mean? Is there a typo?

The stuff going on in these search systems is math and mapping, not thinking. Mind you it can certainly generate new thinking on the user’s part, just as reading a map can. But there is no magic here, just some relatively cool algorithms.

Actually, I did mean to say that Yewno actually cares about and is designed with critical thinking and information literacy of end-users in mind. As resident TSK chef @jillmwo noted this summer, Yewno is focused on ideas and context, driven by semantic tech. https://scholarlykitchen.sspnet.org/2016/07/13/have-you-looked-at-this-yewno/

Also, very well said, David – I agree the work going on within an inference engine is computational, whereas human cognition brings the magic 😉

Regarding visualization of science, what is really needed is 3D visualization, preferably with what is called “fly-thru navigation.” Most visualizations fail because they are a flat 2D projection of what is actually a robust 3D structure. These projections are very messy, as well as incorrectly placing distant things close together, because one is far behind the other in the projection view. We worked on this a bit at DOE OSTI but to my knowledge they never got to a prototype. It is still waiting to be done, like a video game.

Comments are closed.