Evaluating Data Subscriptions Amid Funding Cuts
Looming changes to research funding
There has been no shortage of headlines around cuts coming to academic research funding in recent weeks. Researchers and librarians alike find themselves in an uncertain period, with some Universities cautiously freezing budgets and others waiting to see what research will be impacted. Estimates of predicted cuts range from tens to hundreds of millions of dollars for Universities.
For anyone who has been in this space long, one thing is certain, this is nothing new. Any seasoned librarian has faced budget cuts before and will again.
That doesn’t make it any less frustrating, especially for those of you who champion research for a living. While budgets may fluctuate, the mandate to support research never does.
One line item subject to scrutiny is always data subscriptions. Their recurring nature and lack of visibility into utility make it difficult to evaluate the return on the investment.
After talking with hundreds of data vendors to evaluate their appetite for working with research libraries (we’ve even compiled a market map of those that do!), I’ve seen so much variety in the offerings of different data vendors that I figured I’d share some (potentially) non-obvious things to consider when evaluating the value of a data subscription.
Maximizing the value of your data subscriptions
There are plenty of ways to evaluate a data subscription. In times of budget cuts, it’s easy to get distracted by price, but what you’re really trying to maximize is value. These are some of the attributes I’ve noticed that the best providers include with their offerings to increase value.
Transparency
Of course, price is a critical element. But the last thing you want to discover is that you’re overpaying for a product. How can you ensure you’re getting the best price possible without transparent pricing? When evaluating a subscription, ask the vendor how they set the price. Can they prove that others are paying the same? Is it a negotiated rate or based on a transparent formula they’ve shared?
Remember, these cuts are potentially hurting vendors too! If they don’t have transparent pricing, there is probably wiggle room for negotiation at your next renewal since they won’t want to churn a customer. It’s also helpful to have an understanding of how prices will change over the years so you’re not caught off guard at future renewals.
Before I started Dewey and was interviewing librarians about data licensing, I wrote down this feedback that has stuck with me ever since. “I don’t want a good deal or a bad deal. I just want the same deal that everyone else has.”
Pricing shouldn’t be a black box.
Freshness
If you’re paying a recurring fee, it should have recurring value.
Are you paying for a subscription, but the data is stale? You may want to consider switching that license to a one-off payment.
A subscription should provide access to the full, longitudinal history of a dataset. However, researchers should also be able to use the datasets to study current events (like the impact of federal budget cuts for example!).
If you are paying for a subscription, the data should reflect that with regular refreshes. It should not be the same experience as a one-time data dump. The best data vendors update their data at least on a monthly basis, if not weekly or daily. There is value in enabling your researcher to access the latest available data.
Ease
To quote an overused expression, time is money. Have you received feedback from faculty and students on the difficulty downloading and using the data they need? Do they have options to download the data? Or are they forced to jump through hoops? For example, if the data is regularly updated, you want it to be easy for the researcher to quickly pull the latest data updates.
If they’re having trouble, some amount of work will likely fall back on you to manage these challenges.
Part of your evaluation can be how the vendors deliver the data to you or your researchers. Does it require the library to manage the provisioning to researchers, or do researchers have direct access to modern data connections and download options?
Scalability
Another important attribute is what we can call scale or reach. As the librarian, you’re often responsible for licensing resources that can be used campus-wide or at least department-wide. But oftentimes, vendors put unnecessary restrictions on usage that limit the overall utility of the data.
Most commonly, vendors try a per-seat model, so only a certain number of researchers can use the data or tool at a given time. Similarly, some put limits on the number of papers that can be published over a certain time period. These seem like nonsense to me. The vendors that want to provide the most value should maximize the usage and research that can be conducted with a subscription.
Similarly, data access restrictions like API throttles or terminal-only access can significantly reduce the utility of a subscription if they make the data less accessible.
Usage
Providers should be willing to share regular, detailed usage statistics across campus. Even better is when the data usage can be tied to published research resulting from your investment.
This helps calculate a value per interaction or value per paper/citation metric to compare data subscriptions. You can imagine how ease of use and scalability are going to impact usage.
Common metrics vendors should be able to provide include the number of users, number and type of datasets downloaded, rows or file size downloaded, and information on which user is having these interactions.
Support
Is there a human to connect with if you have a problem? Seems like an obvious thing, but how many vendors send you to a nameless help email address? It is inevitable that questions will come up as researchers dive into datasets. Your vendors should have an easily accessible contact to interact with.
Where should users go to get data questions asked? Does the vendor provide clear documentation for the data, including data dictionaries, summary stats, and FAQs? Is there a forum or mechanism for them to ask questions?
For many data vendors, academics are their “cheapest” customers, so they often provide the fewest resources to support them.
[Bonus] Compounding value
One thing that often makes data subscriptions hard to swallow is that it often feels like most of the inherent value is the first year of the license. This is because you receive all of the historical data with the original license, and the dataset often doesn’t update that regularly (see “Freshness”). While this isn’t necessarily true since you need to remember you’re licensing permitted use to the data with a subscription, not purchasing it, the best vendors clearly increase the value of the subscription over time.
Are you getting new datasets automatically added to your subscription? Or does your price increase with every addition? Take note of how the offering changes with time. The cost per dataset, attribute, publication, etc., should go down over time, not up.
Wrapping Up
If you had a look at all active data licenses at your institution, how would they score against these attributes? It may be helpful to create a scorecard and rate each license across these attributes on a scale from 1 to 5.
Any given attribute for a given vendor is solvable by itself, but when scaled across dozens of subscriptions, it can create a huge, hidden workload and costs for librarians to manage. Especially if hiring budgets freeze up, librarians are going to need to find ways to get more leverage.
Good luck! I’d love to get your feedback on what I missed or got wrong.
We think about these challenges on a daily basis at Dewey and are solely focused on maximizing the utility of our subscription for librarians. I’m always interested in connecting with anyone who wants to help shape the future of academic data licensing.
About Dewey
At Dewey, we have bundled our datasets into a single subscription, eliminating the need for separate, expensive licenses for individual data providers (some up to $100k for just one dataset!). We manage all data provider communications and licensing, saving institutions valuable time.
Through the Dewey platform, researchers can easily browse a diverse range of datasets across disciplines like real estate, finance, GIS, workforce trends, and much more. Key benefits of the Dewey subscription model include:
- Access to a constantly expanding library of datasets, with new data added regularly at no additional cost
- Access to Dewey Docs, a centralized documentation center for quick answers to important data questions
- Seamless integration with University-wide systems, like integrating single sign-on (SSO) and building internal resources to help educate university researchers on the data we have available (example)
We have seen many Universities unlock data across via a department ($25k), school ($35k), or university license ($50k). Increasingly, faculty are collaborating across campus groups to split costs and unlock the value of just one subscription for even more students and faculty. We love seeing exponentially more research projects supported with every subscription tier.