Two years ago, reflecting on the sale of Mendeley to Elsevier, Joe Esposito asked the question, “When Is a Feature a Product, and a Product a Business?” That question has nagged at me ever since. After all, if you could assemble enough features together, you might have a product, and if you gathered together enough of those products, pretty soon you might find yourself with a business. Reflecting on the steady stream of investments being made by some of the most sophisticated content providers, I think we are well advised to examine their approaches under Joe’s framework. Last week, I wrote about how these content providers, such as Elsevier and EBSCO, have worked to find an engine of growth beyond licensing content products to libraries. Although they have different businesses and are pursuing different strategies, I would like to explore one element that they seem to have in common, which is an emphasis on workflow.
Workflows are neither features, nor products, nor businesses. They are sets of activities that can be understood as a process and to some degree systematized, and for my purposes they fall into two categories. Researcher workflows focus on the individual or collaborative research process, including everything from undergraduate paper writing to the most leading edge laboratory research. University business processes are another kind of workflow, and these include everything from the work of the library to acquire and provide content to the work of the research office to establish a research strategy and support researchers in securing grants from, and complying with the requirements of, funders. As growth in the licensed content businesses begins to stall, sophisticated content providers have noticed that there may be both defensive value, as well as entirely new areas of growth, in supporting research workflows and university business processes.
One wry observer has suggested that “workflow is the new content.” Two examples will help to illustrate the shift from content to workflows.
Elsevier has been building up a rich array of research workflow products. Hivebench is an electronic laboratory notebook, used collaboratively by scientific researchers trying to design and organize experiments, share and protect data, and move findings along to patent or publication. Mendeley has grown into a tool for organizing and reading publications, storing and sharing datasets, and connecting with potential collaborators and job opportunities. Over time, it is reasonable to anticipate that the lines of demarcation between these separate products will shift, if they remain separate products at all, as Elsevier’s researcher workflow business continues to develop.
Cobbling together individual products, Elsevier is building a set of research workflow tools, and it is now bulking them up both through internal development and acquisitions. Hivebench was originally marketed to biologists and Mendeley to scientists, and in part to expand the scope of its user community, Elsevier acquired SSRN, the preprint service. Last week’s announced acquisition of Plum Analytics will add features to Mendeley. The suite of tools continues to mature. Among those missing, foreseen by one observer as a near-term direction for Elsevier, is manuscript submission and management.
Through this expanding suite of tools, Elsevier will find that it has developed the means to lock-in scientists to a research workflow, no less powerful than the strength of the lock-in libraries have felt to “big deal” bundles. The lock-in comes from varied sources across different parts of the workflow. The switching costs associated with the collaboration environments in Hivebench and Mendeley derive from a network effect with one’s research peers and collaborators. One’s data and notes are stored and deposited in systems that, even if they allow ready export, are difficult if not impossible to utilize outside the original environment. One’s activity data can drive a level of personalization that may be missing in a new environment. And of course the goal is to sell at least some of these tools in at least some cases institutionally, making a transition to an alternative environment that much more complex.
Notwithstanding these forms of lock-in, for Elsevier these research workflow tools do not need to constitute a business of their own. Activity data from some of these tools are also powering its Pure and SciVal products, which are marketed not to individual researchers at all but to universities to support key business processes of their research management apparatus.
Ultimately, Elsevier’s user acquisition and monetization strategy here is as sophisticated as anything we have seen in scholarly publishing to date. Open access advocates might be concerned about some of these directions, but my sense is that many of these scientists and librarians remain largely focused on trying to compete with, or at least influence, scientific publishing. Building businesses that support, and potentially monetize, researcher workflow is a very different animal. While the Center for Open Science and the SHARE initiative are trying to offer up counterweights, there is little evidence that the open access community as a whole is engaged with Elsevier’s transformation. Springer Nature’s sibling Digital Science is probably Elsevier’s foremost competitor in this space, albeit with a different investment and integration model.
Whereas Elsevier (and Digital Science) are principally focused on the researcher workflow of creating new scholarship, EBSCO has turned its attention to the research workflows associated with accessing the scholarly literature and other library resources. Today for far too many users, the research workflow associated with finding and accessing the scholarly literature through libraries and content providers is far too difficult. These workflows regularly require scholars and students to move across multiple products from multiple businesses, posing a variety of stumbling blocks that impede their progress. A simple Google search finds tremendous amounts of the literature, whether open access or simply illicit. Sci-Hub is said to attract users because a single interface provides access to essentially the entire scholarly literature, even if it is altogether illegitimate. If a content provider could build a single interface that seamlessly provided for the full discovery, access, reading, and citation experience, that would not only go a long way to combating piracy. It would also provide many of the most widely appreciated functions of the academic library.
Two decades ago, EBSCO was a subscription agent, which helped libraries subscribe to the periodical literature, and it offered a set of abstracting and indexing services that it made available online. Over time, it built one of the top aggregators of the scholarly literature, EBSCOHost, and from there it has systematically developed or acquired a variety of tools ultimately designed to bring students and scholars to the literature they need for their research and coursework.
Just as Mendeley is in many ways the heat of Elsevier’s researcher workflow, the interface and dashboard that grows to provide access to various features and services, so the EBSCO Discovery Service (EDS) is at the heart of EBSCO’s. EDS is a discovery starting point – a search engine right now, although I expect it to expand its discovery ambitions over time – for a researcher seeking content of almost any type from nearly any source. The researcher can then move through a variety of middleware that I won’t get into here, provided by EBSCO or a competitor depending on the library’s choice, to move to a content provider site such as EBSCOHost or any number of others, for fulfillment. In a research institution, it is impossible to imagine any single organization providing all the information resources needed to meet user needs. But in a smaller or medium sized institution, EBSCO’s own discovery and content products are closer to providing a complete, and therefore seamless, solution. In this respect, I will be interested to see whether EBSCO will offer products in support of reference (such as the chat products where OCLC has a business), one of the few components missing from offering a fairly complete researcher workflow.
There is another component to EBSCO’s investments, which support the business processes of running a library. Academic libraries have struggled to acquire and manage print and digital collections in an integrated and efficient manner, moving away from “just in case” towards new demand-driven models that enable acquisition “just in time,” and acquisitions and management tools that address these issues are in demand. EBSCO purchased YBP, which along with its subscription agent and its EBSCOHost platform offers up the possibility of creating much more sophisticated fulfillment and delivery models. Linked up with an appropriate library collections management infrastructure, this could be most powerful. Unlike its competitor ProQuest, EBSCO has not acquired such a system of its own but rather is sponsoring the creation of an open source library platform named FOLIO. While FOLIO is touted as open, it will be important to understand the ways that the full suite of EBSCO products interacts over time, and what kinds of switching costs if any begin to accumulate.
As growth in the licensed content businesses begins to stall, sophisticated content providers have noticed that there may be both defensive value, as well as entirely new areas of growth, in supporting research workflows and university business processes.
The EBSCO tool suite is designed to move researchers more efficiently from discovery to fulfillment, providing libraries with the tools to manage this process and increase the flexibility of the models they are utilizing. Over time, and especially for a smaller library with less extensive research collections, EBSCO is offering all of the ingredients of the smaller academic library’s collections and many of its services. ProQuest is pursuing much the same strategy, except that its library system is already mature and in wide adoption and it offers information literacy and citation management tools. This is a highly competitive arena with little in the way of content exclusivity that has helped to make Elsevier so successful.
Maturing Workflow Businesses
In these examples, both Elsevier and EBSCO are building out the pieces of businesses that will support a researcher workflow and that will support university business processes. For both companies, there are clear complementarities and even overlaps between the researcher workflow and the business process, which if well executed will be mutually reinforcing. The basic benefits to researchers and to their universities in these investments are to my eye without question..
For these companies, the researcher workflows are especially important because they will generate a far more robust direct relationship with the end-user — the student or scholar. This is already proving to be a boon for analytics and personalization. Over time, the provision of researcher workflows might even result in direct to consumer business models, offering either a complete solution for unaffiliated users, added value sales to users with institutional affiliations, or ultimately a complete solution. As for the business process products, some of them are already finding substantial sales on campus beyond the library, a trend worth watching carefully.
Workflow business are valuable to customers and users, but they also can generate risks. Ithaka S+R’s study on the research practices of religious studies scholars, released just yesterday, found that even in an environment without complete research workflow solutions, scholars are readily if unintentionally locked into particular tools and find it too hard for them to switch even when a better tool emerges. Analytics and personalization generate further switching costs for users and serve to bolster those who control the activity data and algorithms that underlie them. As these businesses mature, it will be important to be prepared to identify any conflicts of interest that may emerge, as well as additional forms of lock-in that may develop. As these businesses bulk up, and offer something so big and complete that their usage becomes almost inevitable, there is a real possibility that one in each category emerges as successful, just as we appear to have but one search engine and one global social network.
The companies building these businesses are themselves taking real risk, since they are by no means without competition. I focused in this piece on examples from Elsevier and EBSCO, but Springer Nature / Digital Science and ProQuest are no less active in their investments. And competition is not just with peers. As the recent acquisition of Meta demonstrated vividly, other parties are interested in providing for, or disrupting, certain parts of the researcher workflow. The consumerization of technology will likely only serve to draw other companies, some of them far larger, into at least adjacent spaces, in ways that can perhaps be co-opted but otherwise will prove to be competitive.
None of the companies has built a fully mature business here, in terms of researcher workflow or business processes, and all remain in investment mode without any doubt. When businesses cobble together features into products and products into businesses, entrepreneurs take note. There are all manner of startups in Silicon Valley, making features as seemingly trivial as personalized emoji, hoping only to be bought out as features for a bigger company’s product strategy. Entrepreneurs would be well advised to look at the missing pieces in the researcher workflow and business process businesses I’ve reviewed in this piece and build features and products accordingly.