A well-functioning efficient market is characterized by the ability of participants to choose to enter into transactions informed by equal access to information. To decrease market efficiency, it is entirely common for firms to change the parameters of “choice” by trying to “lock in” their customers. Everything from affinity programs, such as frequent flyer miles, to pernicious tying strategies that elicit antitrust scrutiny are forms of lock-in. For the emerging category of research workflow products, providers can be expected to pursue, where they are not already doing so, a variety of forms of lock-in.

A Prisoner - Locked in to Jail
Joseph Wright of Derby, The Prisoner, Yale Center for British Art, Paul Mellon Collection.

The Big Deal is a classic case of lock-in. It was a brilliant strategy, so tempting to libraries, which received vastly more content than ever before at what was at first a minor price increment accompanied by predictable pricing. It was also quite clearly, at least to some librarians with strategic foresight, a terrible strategy for managing a collections budget over time. Yet within just a few years, one after another academic library moved to the Big Deal model of bundled content licensing.

The Big Deal was a brilliant strategy for publishers because universities moved to this new model through a choice they typically made of their own volition. Since then, academia’s unhappiness with the predicted effects of the Big Deal has been channeled at commercial publishers who are seen to have exploited their position. Academia finds itself no less locked in than had it been forced into the model kicking and screaming.

In recent years, Holtzbrinck’s Digital Science has developed an impressive portfolio of tools that support researcher workflow, as well as university research management. So has Elsevier. Clarivate is building an interesting portfolio as well. Much remains unsettled. But there are substantial opportunities for these providers to integrate individual products and services and thereby over time lock individual researchers, and their universities, into these tools. Both Digital Science and Elsevier have denied that their strategies will result in an end-to-end lock-in for researchers across every part of their research lifecycle, but there are other forms of lock-in as well that we should expect to see develop over time.

What follows is a taxonomy of lock-in. For each of several types of lock-in, an example is provided, not as a prediction but simply as an illustration. Note that some cases will lock individuals in, while others may lock universities in, and in some cases there are elements of both.

Exclusive benefits for mutual customers. Providing exclusive benefits for mutual customers of a single provider’s many products is a form of lock-in. A year and a half ago, Elsevier and the University of Florida announced a collaboration that would over time link from an institutional repository to the publisher’s version of record for authorized users while enabling others to access an author-accepted manuscript version that the publisher directly populated into the repository. This arrangement was seen by the University as advancing its compliance obligations while improving access. If such an arrangement going forward were offered exclusively by a publisher like Elsevier to mutual customers of its journals and its institutional repository platform Digital Commons, this would be a substantial boon to customers of both products. Although ScienceDirect journals and Digital Commons would each continue to work independently for customers of only one or the other product, together they would add substantial value that would also serve to lock libraries in to Elsevier scholarly communications services.

Service interdependency. Developing discrete services so that they are interdependent, or preferentially functional, with one another is a form of lock-in. Research information services like Converis, Symplectic Elements, and Pure depend on a variety of data sources. Among these data sources is information about research citation and publication impact. The parent companies of two of these services also provide citation databases (Web of Science and Scopus). If Converis or Pure had a technical dependency or licensing terms such that the university also had to maintain a license to the same provider’s citation database, that could lock universities and their libraries into one provider’s research evaluation services.

Data portability and reusability limitations. Providing additional value for data in context such that data cannot be reused with the same functionality elsewhere is a form of lock-in. Scientists are utilizing workflow tools such as Tetrascience, Hivebench, Figshare, Mendeley, and Overleaf to design and manage their laboratory experiments, gather and organize data, work collaboratively, and write up their findings. Whether these tools become as integrated within each of the parent company workflow portfolios as I expect they will, the data that they are accumulating about scientific research and collaboration are of great value to researchers (and others). How exactly does data portability work and how will it evolve over time? One can download one’s Twitter archive fairly readily, for example, but only as static content, far less valuable than in the platform environment for which it is intended. It is possible that over time standards will be developed making workflow service data not only portable but also readily reusable in other systems, but I am skeptical. Unless that comes to pass, limitations on data portability and reusability will constitute a major form of lock-in.

Institutional sales and product bundling. Institutionalizing services that are today already used by researchers is a form of lock-in. We are already beginning to see bundled institutional sales of selected workflow products, such as repository and research information platforms that already have institutional customers individually like Figshare for Institutions and Symplectic Elements. Over time, other kinds of integrations will increase the logic of offering, though not necessarily requiring, service bundles. Providers will be inclined to offer larger and larger bundles moving up to the complete scientific workflow, an extremely powerful way to reach an institutional customer. Universities might come to license a single end-to-end solution from a given provider, such that there are institutions that provide — and support — the complete “Elsevier workflow” for example to their researchers. Even if individual researchers can still elect to utilize alternatives, they would have to use free services or pay for them from research funds. The draw of using the institutionally-provided infrastructure from a single provider would constitute a powerful form of lock-in.

In offering this depiction of where we may expect lock-in to develop for researcher workflow products, I want to emphasize that as the portfolios of research workflow products grow, so do the benefits of integrating them and thereby the opportunities to lock in researchers and their universities. And let there be no mistake: lock-in can be developed even when alternatives to a product or portfolio are available in a marketplace. As we know from airline deregulation leading almost immediately to the development of frequent flyer plans, competition does not prevent lock-in but, rather, can incentivize it.

Those who look to journals, each of which is its own copyright-enabled monopoly, as the only kind of lock-in do not fully appreciate how individual researchers and their universities can freely choose products into which they will become locked in. Academia through its libraries pursued the temptations of the Big Deal volitionally, and consumers choose loyalty for an airline over ticket price. Can we not expect research workflow providers to find ways to lock in researchers and universities to individual products and an overall portfolio? And can we not expect individual researchers, laboratory groups, and universities to choose of their own accord researcher workflow products, connections, and benefits that yield such lock in?

Roger C. Schonfeld

Roger C. Schonfeld

Roger C. Schonfeld is the vice president of organizational strategy for ITHAKA and of Ithaka S+R’s libraries, scholarly communication, and museums program. Roger leads a team of subject matter and methodological experts and analysts who conduct research and provide advisory services to drive evidence-based innovation and leadership among libraries, publishers, and museums to foster research, learning, and preservation. He serves as a Board Member for the Center for Research Libraries. Previously, Roger was a research associate at The Andrew W. Mellon Foundation.


22 Thoughts on "Workflow Lock-in: A Taxonomy"

This post relies on the scary term “lock in” to cast everything — providing value, creating new and better solutions, consistent execution, smart innovation, or disciplined customer insight — in terms of creating shackles or prisons for users. I don’t think that’s fair or helpful.

Next thing, you’re going to tell me that professional and scholarly societies create lock-in by bundling membership, journals, education, and meetings into a reasonably priced package.

I’m loyal to a particular car brand. Is it because I’m “locked in” or is it because they make cars that are consistently reliable, well-designed, stylish, and comfortable?

As a loyal customer, I’m also able to use this against the car company in many ways, you might say, making my jailer my servant. I’ve received free service, large discounts, and excellent sales terms over the years. The same goes for any service where repeat business or rolling subscription terms are important to the model.

I think a more useful framework for this discussion would be “switching costs.” It’s well-known that businesses try to evaluate and create switching costs for customers, as they want to know how well they’re doing at creating multi-year value and avoiding customers selecting a competitor. That’s different from lock-in. Switching costs that are sufficiently high may make customers feel locked in, but they are only switching costs — whether that means the cost of building a CRM of their own, of migrating data, and so forth. Switching costs can be calculated, and run against the perceived benefits of the switch. Publishing platforms and their customers do this dance all the time.

Back to the car loyalty, both participants (seller and buyer) incur switching costs. The seller risks incurring high customer acquisition costs if they lose a loyal customer (returning customers are far more valuable, and it’s cheaper to keep a customer than to find a new one), so they will do a certain amount (concede on price, expand service offerings) to keep a loyal customer.

I think if you had cast this in the more mutual and malleable terms of switching costs, and how various businesses create higher or lower switching costs (by bundling services, adding unique data or elements, or having patented or unique market offerings), that would have been a more interesting and less rhetorically-driven post.

A sidenote: the tired assertion that journals are copyright-driven monopolies could be retired now, I think. If you look at author behavior, they publish in multiple venues with multiple publishers, and certainly don’t seem to be constrained by their prior publication choices or outcomes. A reader might retrieve a paper from JSTOR or Ovid or PMC or another aggregator and not from the publisher, so even the decision of where to get a paper is not constrained necessarily by outlet. If there is no constraint on choice except for single decisions (which paper to download, where to submit), how is that a monopoly?

It’s great to see the curtain being drawn back to reveal how many different ways lock-in comes about. Too many times we don’t realize lock-in has occurred (the means by which it is occurring) until we find ourselves in the position of being locked-in (the end result). High switching costs are indeed one means to lock-in … a highly recommended in the business and marketing literature as a matter of fact! As an individual researcher at a university, I myself find myself the victim of the institutional sales lock-in effect often as well. Having this taxonomy, Roger, has already been useful today in thinking about when I’ve locked-in myself because I’ve been unwilling to suffer the switching costs and cases in which I am locked-in by institutional practices because there are different escape strategies for the different situations.

Thanks Lisa. I’m thrilled to hear this is useful. There are probably other forms of lock-in as well that deserve consideration. I’m happy if I’ve helped spark a discussion about them. Who knows: perhaps if we are all a little more mindful about these issues, it can strengthen the hand of the consumer.

Kent, I appreciate your help in getting the new year off to a vigorous start here at the Kitchen! But I think your reasoning is a bit confused.

It is absolutely fine to keep buying cars from the same manufacturer because you love their quality and pricing: that is the essence of choice I am hoping to see maintained in the scholarly publishing and workflow marketplace. It changes the parameters of your choice, though, if every time you buy eggs you earn points that lower the monthly cost of your car’s lease from a single manufacturer. It may not be illegal — and it may or may not be bad for the auto marketplace overall — but it certainly affects your choices.

As to switching costs: They are an important issue here, and captured in some parts of my piece, but I am concerned that we explore other kinds of tactics here as well.

While not all of these tactics are necessarily bad, it is during this phase of market development for a new and emerging categories of products that we should give closer attention to which of this tactics are beneficial to all, which are troubling and deserve opposition even if they are in another sense legitimate, and which should be cat as illegitimate. I hope you can see that there is a spectrum here.

Businesses increase switching costs all the time with tangential offers, eggs and cars notwithstanding. If I get a better deal from an e-commerce store because I use a certain airline, that increases switching costs if I choose to fly another airline. If I get a better deal on a credit card because I choose a certain hotel chain, same thing. Increasing switching costs is what we’re talking about.

The problem with casting this as “lock in” is that is looks unilateral instead of bilateral, like the customer is beholden to the company and never the other way around. Companies run a lot of risks in scaling service businesses, and starting them is usually the most expensive part.

The reason businesses want to increase switching costs is because customer acquisition is expensive, so to earn back those costs, most customers need to be retained for a number of years to make it work. For instance, it may cost a company $10,000 to acquire a customer who pays $2,000 per year for a service. They will need to keep that customer for a long time to make that customer profitable. If there are no switching costs — that is, nothing much lost if the customer cancels — the business model fails. This goes for membership organizations, car dealers, grocery stores, banks, airlines, aggregators, and so forth. Customer acquisition is expensive, and switching costs are embedded in the value proposition of almost every company. And they are bilateral — the company risks a lot and continues to refine its model to maintain switching costs.

One of the low-priced journals I worked for was capable of running complex financial analyses due to the brilliance of the team. The customer acquisition costs for a $99 subscription could run over $1,000, meaning the journal needed to keep customers for 15 years to make money (given the production costs and margins). Membership societies often themselves have high customer acquisition costs, and work hard to impose switching costs (accreditation, certification, etc., are all associated with switching costs). To keep members, subscribers, or customers, firms will often concede on pricing or expand service or add offerings, as their switching costs (cost of acquiring new customers compared with keeping current ones) can be very high, as well.

I don’t think my reasoning is confused. I see in this post an over-reliance on a single rhetorical trick (labeling everything with the scare term “lock in” and using a prisoner analogy), and that over-reliance made it difficult to rationally address the full scope of what switching costs mean to a business, why they are important, how they are bilateral, and how they matter to workflow options and opportunities.

I’m sorry you dislike the rhetoric. But I made clear that many forms of lock-in, as I defined the term, are volitional by the customer. And that, in some cases, this is precisely the problem.

Is “customer acquisition” a term commonly used in the business world? I find that disturbing; people aren’t for “acquiring”, and I think its existence reinforces the jailer/prisoner framing rather than refuting it.

It makes me wonder how much of the semantics/rhetorical debate centers around phrases designed to let us ignore what happens to the people on the other side of the transaction.

“Customer acquisition costs” and “customer acquisition” are common marketing terms. They are used to make the process more objective. When you acquire or get a customer, it’s a happy thing, and an obligation to serve. There isn’t anything pejorative about it. The customer expects to be served and to get value for money, and if the supplier doesn’t meet expectations, the customer drops them or demands a refund. All of these things factor in. The relationship is bilateral, and mutually beneficial if it’s a happy relationship.

Calling ‘adding more value than your competitor’ a lock-in strategy is not very helpful.
Apparently the commercial providers are better able to provide and support scholarly workflow tools.

Open source variants have emerged for some components, and at a product level these might even be on par with the commercial offerings. But to consistenly support and nurture these tools is a very different, and less glamourous, job. It is something you see all over the FOSS (free and open source software) world: people will put effort in creating great ideas and great products such as Linux, but when you need someone to put together the pieces and make it easy to work with, you have to pay people such as Red Hat to do the boring work.
So you either pay in recognition and kudos, or you pay in cash.

The real question is: why does the scientific and library community value these tools so little that they fail to provide proper recognition (read: career prospects) to the people creating and supporting open source workflow tools – or the people working on Open Access publishing for that matter.
The budget for academic recognition should be far more accessible than the cold hard cash needed to buy tool licenses and support contracts.

I do not think that “adding more value than your competitor” is necessarily a kind of lock-in. That is not a quote from my piece, nor is it a part of my argument. There are many ways to add value without limiting choice. It is the exclusive and limiting arrangements that I refer to in this piece that may be troubling.

I strongly agree with you that we have a problem in academia with inadequately recognizing the contributions of people, and organizations, in support or enablement roles. While the library is often the victim of this dynamic, I don’t don’t that in others or perpetuates academic hierarchies. I am disappointed if this is limiting our ability to create a more competitive marketplace for workflow tools. I would welcome the chance to learn more if you are willing to share, here or privately.

The quotation marks were intended as a summary of the practices you describe, not to indicate a direct quotation. I apologize for the lazy punctuation 🙂

I do think that bundle deals &c. are adding real value both to vendors and to customers. If they didn’t, no one would enter into such agreements. But of course librarians and scientists need to be aware that these deals do indeed increase switching costs. I think mostly they are aware of that, as they were aware of the effect of the Big Deals.

Community-driven tools and solutions are inherently more democratic, and can solve some really thorny challenges through the power of a global network of contributors. The philosophy of openness as an organizing principle has real power. But commercial concerns have a laser focus on innovation +and deep pool of resources that allow them to create some powerful tools.

High switching costs for customers are the ultimate goal of any software development enterprise–open source or commercial. It takes a lot of time and effort to create and maintain tools that people love to use, and keeping a customer from trying something else is a major component of the business model. That’s why integrations between tools like Microsoft Office or Salesforce have been so successful for so long–they just work better together than apart and they solve business workflow problems. In the commercial business world we called that strategy brilliant, but we quietly vilified those companies under our breath for creating that dependence, a dependence that we blithely “locked” ourselves into.

I don’t see anything inherently bad about committing to tools that solve your particular business challenges. That’s why over here at COS we’ve built OSF to utilize add-ons and remain highly modular–we want you to be able to work how you work. “Lock-in” becomes uniquely problematic where control over and rights to research outputs and data are concerned. As a researcher, I’d have a real problem continuing to support any tools that did not allow me to retain copyright and complete control over how, where and at what cost my work is published, accessed, or re-used.

A definition of “lock in” from Wikipedia below. It seems to be commonly accepted that “switching costs” = “lock in”

“In economics, vendor lock-in, also known as proprietary lock-in or customer lock-in, makes a customer dependent on a vendor for products and services, unable to use another vendor without substantial switching costs.”

Thanks John. I still haven’t found any reason to utilize a different term for the basket of behaviors I describe here.

I’m still struggling a little with the semantics, mostly due to the jailer/prisoner relationship that is implied. Is the term only useful when the user is dissatisfied with the service or the price? What if I build an internal system at my company that would be costly/difficult to move off of — have I locked myself in?

The benefit of phrasing these experiences in the terms of switching costs rather than lock-in is that switching costs are variable and changeable. The switching costs vary depending on circumstances and how they are created (and how long they endure). For example, switching costs for music players were relatively low until iTunes, which created huge switching costs for users of non-Apple music players. Spotify changed that, and lowered the switching costs, and that put Apple on the defensive.

Phrasing every switching cost as “lock in” makes them all appear to be nefarious and immutable. They aren’t. They are variable and mutable. This is why I think that while the overall idea of this post is fine and laudable, the rhetoric is too brittle and doesn’t allow for some subtlety that framing all this as switching costs would have naturally cultivated.

Kent, I am grateful for your kind words about the ideas inside my post. I appreciate you turning the commentary back towards the considerations I analyzed in the piece — which has been praised by everyone from scientists to librarians to an equity analyst — rather than getting stuck in a semantic debate of no consequence.

In an earlier reply Roger you referred to these all being volitional by the customer. Perhaps there’s a germ to a solution in the semantic/rhetorical issue there.

A spectrum whereby there is volition by the customer and a parallel spectrum where there is violation of the customer. I have these as a parallel as they aren’t mutually exclusive. Nor are they immutable.

A relationship can start as being highly volitional but over time become more violational. So there is benefit to customers as Kent states, but I find it becomes increasingly harder to exert those benefits as the relationship increases as they vendors want to give less away over time. This manifests as longer times in home queues (because our calls are so important…) etc.

Better minds than mine can see if there’s more in this, but it seems the balance is always shifting anad what was once volition comes over time to be violation. And that’s when the grievances might start.

Indeed. I believe that is almost precisely the point I was making in invoking the Big Deal. “Academia finds itself no less locked in than had it been forced into the model kicking and screaming.”

A cursory search of a business database on the terms “lock-in” and “switching costs” tells me two things: one, a significant number of articles about “lock-in” suggests it’s a term of art within the marketing field (though an idiomatic one that doesn’t seem to warrant its being a subject term), and two, that the terms aren’t interchangeable: “switching costs” seems to focus on a marketing process, and “lock-in” as a potential *result* (for consumers) of establishing high switching costs, if I’m understanding it right. As Roger is enumerating effects on consumers, it seems like the right term to use. If there’s a semantic/rhetorical issue here, (jailer/jailed), it seems to exist within the marketing field, not just in Roger’s post.

“Lock in” is essentially code for “very high switching costs, so high that switching is too painful or costly or impractical to realistically consider.” Sometimes, the switching costs are driven by the customer’s situation (I may want to switch insurance providers, but the days after I get into a car accident is not when my switching costs will be lowest).

Switching costs occur on a spectrum, with lock-in at the extreme end of it. I think putting these costs on a spectrum, and acknowledging the dynamics on both sides, is more enlightening than labeling all switching cost scenarios as “lock in.”

I think it is overly limiting that the discussion over the semantics of “lock in” is focused in switching costs, or any costs specifically, because we are not in the for-profit business aiming to drive up our bottom lines or ROI. In the marketing examples, that is exactly what they are trying to do; for libraries, I hope not.

The auto industry example is a poor comparison because the auto industry has relative equality across the competition, which is not the case in the library publishing and vendor space. Going from Toyota to Ford will get you relatively the same product. There is nowhere close to the same relative equality across library publishers, vendors, and open source projects, and where there are similar products, it is prudent (I would suggest required) that one use a project management reality triangle (scope, resources, time) to fully understand the lock in.

Scope is where I the most disparity in the list above, and for the main players in this space, there simply is not enough relative equality of scope between the players in this space to believe it is even feasible to switch between offerings and simply drive away (metaphorical pun intended). And I include open source offerings. For the near future, any decision we make in this space (including choosing to NOT implement something), we must acknowledge we are locking ourselves into that decision for a period of time that is indeterminate today.

Resources has to include the staff it takes to use the products, as well as the staff it takes for testing, implementation, training, etc. And while staff do cost money, in nearly every budget conversation I’ve had while in higher education, we rarely equate staff time to money / cost like we would in comparing the cost difference between an EBSCO product vs. a ProQuest product or X package vs. Y package. This is exactly where open source projects in libraries has created a lock-in, and we need to discuss this fact a lot more. I make that last statement as a long-time leader and proponent of multiple open source projects, yet see the daunting future of sustaining these projects in rather uncertain conditions.

Time is rarely on our side and it is immutable. For many services, the length of time it would take to transition from Product A to Product B is greater than the capacity the staffing of our libraries can afford given the plethora of other workflows and services we must maintain. Our propensity to avoid ending any local service or saying no to good ideas in combination with the varying cycles of research and scholarly activities leaves little good times to make these changes. Taking the appropriate amount of time to transition methodically and carefully, such as 12-18 months, is a half-life for some products and a period some technologies have vastly improved or been outright replaced.

Lock in is indeed the right way to discuss this. While I recognize the emphasis the imprisonment metaphor places on this discussion, the lock in isn’t always so dire, but instead are constrained by the reality triangle and need to adapt to living within our means and these constraints. Keeping this broader perspective can help us call out predatory practices of vendors to exploit these constraints, but we also have to recognize the lock in that naturally exists and the lock in we ourselves have created.

Comments are closed.