Reading Time: 13 minutes

I cancelled Bloomberg Law. It was an agonizing decision-making process. The decision itself was not hard. If anything, it was obvious and overdue. That tension, of having an obvious decision that still needs to be teased out to its full extent, was a professional challenge. It has been a long time since I had found myself in the position of feeling like I needed extensive cover for a decision. It’s the difference between being an executive director of an agency and being a director within an organizational hierarchy. I thought it would be worth walking through my decision-making process for anyone who is a new director or may be facing a similar sort of challenge.

This is not about any particular legal publisher or even publishers in general. We make highly visible, impactful operational decisions in many areas. I will be referring to Bloomberg Law because there are some aspects of their market position, collection role, and license that had an impact. But my goal is to focus on the variety of decision points that I considered in making my decision.

I perhaps should start by saying that, with time, there is a certain amount of expertise and confidence that can come with decision-making. You may have seen the same issue arise in the past so you have a preferred way to resolve it. Or you are more aware of likely outcomes. The longer you are in a role, you have a better understanding about your environmental inputs and how your organization will react.

Instinct should not be ignored, particularly once you are experienced. If the juice is not worth the squeeze, you often know it without a lot of other input. Sometimes instinct will warn you that you are missing something, even if you’re not sure what.

From that perspective, then, I came to a decision very quickly with typical inputs: available data, opinions of colleagues, assessment of impact. I am new to my role, however, and new to the academic law library context. I have not yet created the goodwill or political capital that one might accrue over time in a role. I had no burning platform. It was my first really visible decision and I purposefully put on the brakes, turning my initial inclination to decide into something more like a hypothesis.

Communication

Communication is perhaps the most important component of any decision-making process. Since most decisions impact other people, letting them know what you’re thinking as early as possible is important. A wise librarian once told me that they prepare a list of people who should be contacted in the case of a decision, to be sure not to miss anyone.

Communication can take time, however. I think this sometimes causes it to be both a drag on and an obstacle to decisions. If you are communicating authentically, you need to give people time to answer. It is not hard to find stories of people who were asked for input and then find out their input was window dressing, that a decision had already been made and the input didn’t matter.

Alternatively, decisions are sometimes made without communication because there isn’t time to gather input. This is sometimes necessary; sometimes you have to make the best decision you can with the data and insight you can gather and accept risks. The risk goes up, though, of missing an important factor or constituency. In the case of electronic licenses, usage data can help substantially to identify people impacted.

One good thing about communication is that the impacted will respond immediately if you let them know you are considering a change. Someone who is invested will get in touch with you to explain what their needs are and how the decision will impact them. This feedback need not be determinative. I have made a change in services and heard back from people who were impacted but whose expectations were not reasonable. The changes stayed in place and we guided them to alternative solutions.

Communication continues after the decision too. I describe how I handled it below but you should be thinking about how you will communicate the decision, what the change will be, and how you will ameliorate any impact if you can even before you get to your decision. This falls under the heading of change management but is also just common courtesy. Communication that respects the impacts of your professional expertise will help people accept your decisions.

The Double Cost of Waiting

I am going to start with the cost first because I think that is perhaps the most common driver of law library collection decisions. Do we have the money? Do we need the content at the cost required? I think that’s only half the cost though. The other half is strategic.

As I made clear to the publisher’s representatives, the license was fully funded. This was not a decision made because of funding cuts or budgetary shortfalls. It wasn’t made with any financial pressure from within the organization. If the license made sense to continue, we could have done so.

One thing I try to avoid is panicked decision-making or default decisions. This is common when you were planning on a license but your budget gets cut or there is some other unexpected financial impact. No one likes being forced into a position where money alone is driving the decision-making process. That sort of situation eliminates any opportunity to consider value or impact or other criteria. While I still think libraries should have a funding shortfall clause in their licenses, I don’t want to have to use it except in an emergency.

A chart with dual axes. The left hand axis is cost per full time student. The chart starts at $75 per student in 2022 and ends at $91 per student in 2025. The right axis is total license cost, going from $64,900 in 2022 to $81,895 in 2025.
A chart showing year over year license cost increase for Bloomberg Law based on full time equivalent student count.

There is a strategic impact to cost-based decision-making. If you are in an organization with a scarcity mindset or culture, there can be a perception that the organization lacks the financial wherewithal to be good. That a lack of resources is holding back the organization, is making it less than it could be.

If decisions are only being made when money is tight or being cut, or when roles are being eliminated, or space has been filled, I think this can lead to a scarcity mindset. Decisions only being made when things have to be gotten rid of can create a feeling of loss without an offsetting gain.

This is not theoretical. After I made the decision to cancel Bloomberg Law, I communicated the decision to the faculty and students. One student responded, saying they understood the decision and agreed that there are better uses for law school resources. It reaffirmed in my mind that the organization can see this sort of decision in the same way: are you deciding due to scarcity or are you deciding so that you can redeploy for growth? I think it can impact how the library or the organization feels about itself and its resources.

This balance impacted the timing of this decision. I am confident that no one would have questioned me if I had extended the license, already a year-to-year license, another year. At the same time, I knew that this was a good strategic time. You never know about the future and I was confident that it was better to make this decision while the money existed.

Resource Role

This was a lot harder to get at and can be instinctual in some cases. When it comes to people, you can look at what they have done and what the organization needs. In most cases, there is an opportunity to help direct a role so that, over time, it continues to be well aligned to the organization’s needs.

The same thing goes with collection. We add and we deselect as we see the needs of our organization change. The thing that strikes me about electronic licenses is that the larger they get, the less aligned they can become with the organization’s collection goals. In a sense, they become so big they will come entirely out of alignment.

In the case of Bloomberg Law, I tried to understand where it fit into our collection. This is particularly tricky with the so-called comprehensive legal publisher resources. If they were truly comprehensive, we would only need one of them. They are not, so we have to mind overlap and reach.

I have written about usage data recently and the dearth of it was a critical factor in my decision. After that blog post, the Bloomberg Law representatives were able to get me a bit more data. I was then able to add my own coding to it so that I could determine who was using the platform. This was still woefully short of what I have seen from legal publishers in the past. It seems as though the account representatives are being hamstrung by their sales organizations, which is difficult to understand. Surely the people licensing the Bloomberg terminal are interested in usage data?

For example, I was able to get an activity report (how many transactions a user experienced) and I could get a report of what Bloomberg Law content was being used (X% was secondary, Y% was docket) but the two reports were not connected: I couldn’t tell that User 001’s 230 “activities” were secondary content or newsletters or what.

A pie chart reflecting the number of users who engaged in activity at any time with our Bloomberg Law subscription between January 2023 and November 2024. About 40% had no activity and about 50% had between 1 and 100 activities. The faculty data is broken out and layered on the pie chart slices, showing 10 faculty had never accessed the system and 12 had only accessed it between 1 and 100 times, with 4 of those having only 1 activity in the two year period.
Coded data from a Bloomberg Law usage report on activity by account.

There is an interesting and uncomfortable progression here. In the old days, we would collect print (that’s all there was #old) based on sometimes speculative need. We collected just-in-case. You never knew and better to have it on the shelf. As budgets tightened and electronic resources emerged, we shifted our collection strategy to just-in-time, reducing print footprints and relying on digital and inter-library loan to bridge gaps.

The licensing of multiple large electronic databases that overlap—in this case Thomson Reuters Westlaw, RELX Lexis, and Bloomberg Law—means that we are once again collecting just-in-case. The amount of latent content, the content that exists but goes unused within each system, is probably hard to know. On the other hand, the lack of usage data means that the legal publisher is unable to show how much of their database is unutilized just-in-case content. I tend to assume that it is unused unless there is data to show it is being used, the same as I would with in-house circulation data.

This is the streaming media dilemma. How long do you carry a Hulu or Netflix or YouTube TV subscription without watching it? One survey found that viewers spent 110 hours a year searching for something to watch (about 1% of a year). So perhaps you would say you’d need to watch 200 hours a year or 365 (an hour a day). But if you don’t watch anything on a platform, it might as well not exist, and you would likely consider canceling it.

The Bloomberg Law representatives were very clear about the value of their content. I believe this is both true and unpersuasive without additional data. This was one of the major obstacles for me. I know that, within that platform, there is a lot of rich content. I worked at a large law firm in the 90s during law school and BNA materials were used daily. It is content we would have been entirely certain to collect just-in-case back in the day. I agree with them that their content is excellent. At the same time, excellent unused content sits on a shelf just the same as mediocre unused content does. Information that is unused might as well not exist.

There will certainly be differences of opinion about this but I ended up viewing Bloomberg Law very much as an archival resource similar to HeinOnline. To the extent it was not duplicating other large legal publishing databases, it was mostly providing a deep back catalog of important publications. A specialist lawyer would want discrete parts of that deep catalog but I am not persuaded a law school needs it.

Law school is breadth over depth, focused on competence rather than domain specialization, which will come over time once a lawyer is in practice. If there is specialist demand within a law school curriculum, with a dedicated topical center for example, then I would expect that law school would see usage data to justify the license.

For me, Bloomberg Law is distinguishable from HeinOnline because the latter is largely an archive and space saver. There is a current scholarship need for HeinOnline that Bloomberg Law (and many other subscriptions) lacks. There is a long-term content archive that HeinOnline provides by capturing law journals. It reduces my shelf space demand. Most other legal publisher databases may not retain any digital archival versions of editions, so their orientation to currency can be limiting from a collection maintenance standpoint. You can find currency almost anywhere.

Content Impact

I normally make decisions without broadly surveying the researchers. This is not for lack of interest. It is difficult in a public law library to survey a population that changes so regularly. In this case, I started by getting feedback from the library staff. Again, in light of a usage data gap, we had to rely on anecdata. I followed this up with a survey of faculty that was informed by what data we did have: we knew some faculty used the docket functionality and we knew that a large number probably didn’t use the service at all.

This approach is a pretty cautious one. You aren’t risking much by polling the library staff. They will often see the pitfalls you might not know of: a faculty member using a particular resource, a database licensed for a specific program. If you hit a red flag in this first circle of investigation, you haven’t risked much.

This email survey was useful. If it had shown a much higher faculty usage, either direct use or through assignments during courses, it would have shifted the decision towards maintaining the license. In the end, the faculty use was minimal and their anecdotal feedback confirmed that the data used was largely commodity information: current awareness, dockets, and so on. Usage data on this point would have been hugely valuable because some people forget that they have used something but we use the data we have, not the data we wish for.

The feedback confirmed that we were not dealing with unique content. Current awareness is available from all major legal publishing platforms, as are newsletter and alert customization. One faculty member was concerned about missing content and so I ran a comparison of newsletters, extracting headlines by date, and looking at when topics landed on the same date or not. I found editorial differences in coverage but not based on exclusivity.

I’m not persuaded that Reuters-syndicated news will be substantially different from Bloomberg-syndicated news nor even the content generated by Reed’s Portrait Media d/b/a Law360. Law doesn’t happen in secrecy and many of us have supplementary tools to help us close latency gaps. Most news is not investigatory, it’s revelatory. If an event happens, it will be widely reported, either initially or via syndication.

Brief Tangent: Isolating Headlines from Newsletters

I used the Stylus browser extension to set up some CSS that allowed me to separate email newsletter headlines from the text. For Westlaw Today, it was:

.x_co_searchResults_summary {display:none;}

Then it was just a matter of cutting and pasting headlines into a spreadsheet and doing a bit of tagging to categorize the subject matter. My spreadsheet looked like this, and I could then filter on keywords to find similarities:

Again, this is more work than I would normally do but I valued this faculty member’s input and it was a use case that generated more data. The juice was worth the squeeze!

But I digress.

This decision would have been made harder if faculty were using or assigning content in the back catalog of a product. Something that was unique and not available in an alternative format. I think the number of information products that fit that model are pretty small. It also makes them easier to cancel if the people using it find something else to use or change their information needs. Bloomberg Law’s lack of content segmentation (or, alternatively, their substantial, indivisible content integration) meant that I had only an up or down license choice. There was no way to license just the content that was to be used.

Prestige and Emotion

By now, my decision was headed out into relatively uncharted waters. To be sure, there was not, up to this point, anything that suggested that the initial hypothesis wasn’t accurate. But I hope that I’m modest enough to know that I don’t know. So I checked in with some folks to ask about things that I wouldn’t normally have ever considered. Like prestige.

Do people view a law school as more valuable or higher quality based on the number of licenses they carry? That seems pretty silly to me but what do I know. After talking it over, I decided that it was unlikely that carrying one license or another would be a deciding factor for an applicant, for an alum planning a gift, or even for any of the accreditation bodies or marketing rankings. I do plan to have a discussion around keeping our budget up though because I know that number is widely distributed.

What about “law schools like us”? Peer institution benchmarks are a real morass. First of all, as I explained to the account representatives, our law librarian integration in legal research classes isn’t the same as other places. An offer to have me speak to a law firm director who just want firm-wide license also didn’t provide me with new data. I had already spoken to law firm colleagues. I already knew there was no data to connect access to Bloomberg Law in a law school with hiring success on graduation or as a measure of competence or other employment outcomes. The lack of access to their product isn’t going to be the thing that hinders our graduates from career success.

The emotional argument came up too. I had run into this one before, with a $200,000 license in Canada. A governance representative asked whether, if access to the database resulted in a single person not going to jail, it wasn’t worth spending the money? I mean, sure, we would not want someone to go to jail. But there is no causation between access to legal information and legal outcomes.

This is a long-standing law library problem: how do you show your value when the use of legal information is so attenuated from the outcomes in a given legal issue or other success outcomes. Even in law firms, where the workflow is almost entirely contained within the organization, it can be hard to show that access to a platform led to a particular outcome. It would mean discounting the legal professional’s skill involved in the matter, the fact pattern, the other resources used, etc., etc.

One tension that would be unique to a law school licensing content like Bloomberg Law is its relationship to either law student hiring outcomes or success upon joining a legal organization. As I understood their presentation, Bloomberg Law is now firm-wide in about 40% of the AmLaw 200, with more selective per-seat licenses in most of the rest.

There is a whole discussion about whether it is the role of the law school to fund familiarity with legal research products. Given the relatively small size of so-called “big law”, I am most interested in market penetration by a legal publisher across the entire legal profession. Not only are most lawyers in the largest law firms, they will likely not stay there long even if they get there to begin with. The NLJ 500 for 2024 lists about 174,000 lawyers in the top 350 firms, which includes firms of 115 or more lawyers. The ABA’s 2024 national lawyer population survey lists about 1.3 million active lawyers, so “big law” is about 13%.

The bigger question is whether law schools should be teaching law students how to use products or should they be teaching them how to do legal research? As someone else noted on a MAALL roundtable virtual call (if someone remembers who it was, let me know so I can attribute it), the interfaces of the legal research providers change frequently, certainly more than once every 3 years. If we are teaching interfaces, we are giving law students information that we know will be dated by the time they need to use it.

This brings me to a longer term consideration. How many large legal publisher databases do we need to teach to successfully create legal research competence in new lawyers? We certainly don’t need three (Westlaw, Lexis, Bloomberg Law). I’m not persuaded we even need two. As the law firm market changes and we see practitioners shift to a single provider (or no provider), there seems less rationale for a law school to carry multiple large legal research subscriptions.

As we see accelerated delivery of generative and extractive artificial intelligence tools, far outstripping the ability or reach of law librarians to teach the products, a fallback to legal research literacy seems the only likely outcome if it is not already what a law school does. And if literacy is the competence goal, then it can be done with a single so-called comprehensive tool.

Decision Outcome

I digested all the inputs I could and took as much time as I could to make the decision. In the end, I had a 30 day opt-out clause I had to fulfill. The license renewal arrived about 30 days in advance of that, so I had 30 days to stay or go. I used almost all of that time, sending off the final email just before the deadline due to the Thanksgiving holiday.

I now focused on preparing the communications about the decision. There is no question that I was aware that some people would have their scholarship or other work disrupted. I wrote up a summary similar to this blog post to explain the process. This was as much for me, as a bit of a checklist, but I wanted to have something to share with those who cared for more information.

I was not inviting input into the decision once it was made but even people who have worked in law libraries for decades don’t necessarily know how they run or why we make decisions. I figured, at this point, where I’m an unknown quantity and people may wonder about my decision-making, it couldn’t hurt to lift the hood a bit and let people see how the engine is put together.

Once I had that revised and polished, I sent out the emails to faculty and students that the license would lapse on December 31st. We are fortunate that we have a faculty liaison librarian program that we just spun up, so our librarians could help individual faculty with a need. One positive outcome is that we are building out our own knowledge on our team on how to use alerts and newsletters within Westlaw and Lexis to be better able to support the current awareness needs of our faculty.

I was surprised to get so little feedback about the decision. In a sense, it made me feel like I made a mountain out of a mole hill. But I felt more confident about the decision having really taken the time to be more deliberate, even though I started out very confidently. This was very much how it was for me when I went to Canada from the United States and had to figure out what was the same and what was different, and which skills were transferable and which needed to be adapted.

The tough bottom line for library directors making large or visible decisions is that you have to be able to own your decision. That’s why they hired you. There is a balance between analysis paralysis, which hinders good stewardship, and making poorly informed choices. Most decisions are not forever. I figured that, if it was a mistake, I could always ask for the budget money again and we could restart. In this particular case, I don’t think that will happen. But it can pay to think about what your future strategy will be, to identify replacement products or to re-license or re-purchase (as we used to with print books) after a period of time.

It’s this strategic perspective that should really drive your decision. If you can make the decision when you have the resources, you have time to align it with your strategy. Your strategy may be to reduce your license costs but it doesn’t have to be done as a cut. When I think about this decision, I think that I will consider it successful if, from a strategic perspective, we are still moving forward successfully despite not carrying that license any longer. It’s something that will only be known in hindsight.