Information Today, Inc. Corporate Site KMWorld CRM Media Streaming Media Faulkner Speech Technology Unisphere/DBTA
Other ITI Websites
American Library Directory Boardwalk Empire Database Trends and Applications DestinationCRM Faulkner Information Services Fulltext Sources Online InfoToday Europe KMWorld Literary Market Place Plexus Publishing Smart Customer Service Speech Technology Streaming Media Streaming Media Europe Streaming Media Producer Unisphere Research

For commercial reprints or PDFs contact Lauri Weiss-Rimler (

Magazines > Computers in Libraries > November 2019

Back Index Forward

Vol. 39 No. 9 — November 2019

Ebook Availability, Pricing, and Licensing: A Study of Three Vendors in the U.S. and Canada
by Michael Blackwell, Catherine Mason, and Micah May

This article reports on a study about ebook title availability in the U.S. and Canada conducted by members of ReadersFirst (, a loosely knit organization of 300 libraries worldwide that is dedicated to making access to ebooks as free and open as print. ReadersFirst has been studying ebook availability in order to establish benchmarks and identify trouble spots. Three market-leading vendors in the public libraries sector were sampled: OverDrive, bibliotheca, and Baker & Taylor. Overall, while availability of ebook titles was close to that of print, licensing and pricing continue to present challenges to public libraries seeking to build deep digital collections. 


The impetus for this study is the work of Rebecca Giblin of Monash University and her Australian colleagues. They examined 546 titles from five library ebook vendors in Australia, and the same titles in Australia, Canada, New Zealand, the U.K., and the U.S. for one vendor.1 Then they completed a 100,000-title study, looking at one vendor’s offerings across the five countries. The group has published two papers so far. The first noted that while “availability was better than anticipated, licensing practices make … infeasible … certain kinds of e-book[s]” and cited “considerable intra-jurisdictional price and licence differences” across vendors.2 The second paper determined that “[b]y setting terms independent of the titles’ value to libraries, publishers may discourage libraries from adding older and less-demanded books.” 3

Members of the ReadersFirst Working Group were intrigued by the results. In this follow-up study, we have compared three major ebook vendors in the U.S. and Canada, hoping to advance the work of the Australian team and give a more detailed view of the state of the library ebook market. The purpose of our study is not to suggest that one vendor is better than another. It provides a snapshot from early 2019 of what digital librarians face when building and maintaining collections. 

The study includes 557 titles for the U.S. and 556 for Canada. Some titles from the initial Australian study were included in our study for some overlap and consistency (for selection criteria for the original study, see Our Methods 2018).4 Some titles that are primarily of cultural significance to Australia have been replaced with Canadian- and U.S.-produced titles, chosen by library selectors from library review sources. Titles of historic interest date back to 1924. Approximately 50% of the titles are from the past 5 years. Titles of interest to children, teens, and adults are also included.

A follow-up to our study, should one ever be conducted, might focus more narrowly on Canadian and U.S. titles, since some lack of availability of titles could be attributed to low demand here for some U.K. or Australian authors—or perhaps a lack of international publishing agreements.

Caveats— Our study has two known limitations. First, data collection took place over 2 months in late 2018 and early 2019, instead of during a preferably shorter window of time. Some variation in licensing models or price could be attributed to querying titles from different vendors at different times. Second, our study does not include every library ebook vendor in the U.S. and Canada, although it does include three market leaders. Consequently, some titles that we list as unavailable might be available from a different vendor or, if available, even at different prices or access terms. However, the number of likely variations is small, and the three vendors are well-established and heavily used. Our study’s findings are a valid snapshot of the sampling time and reflect what most public libraries in the U.S. and Canada can expect to find.


  • All of the titles included in this study were available in both the U.S. and Canada, but not all titles were available as ebooks.
  • Amazon-published books, which are not available to libraries, account for many of the unavailable titles.
  • Overall, prices for ebooks varied widely—from $2.27 to $90.95 in the U.S. and 99 cents to $124.04 in Canada.
  • While ebooks are sometimes cheaper than their print counterparts, they generally cost more.
  • Recently published titles are usually more expensive than older titles, but older titles often cost more than might be expected.
  • Perpetual license titles were generally more expensive than those metered by time period, number of circulations, or a combination of both.
  • Vendor coverage of titles varied more in Canada than in the U.S., and no single vendor had complete coverage.
  • Licensing models complicate access: Short-term metered licenses at a relatively high cost for older titles may make those titles less desirable to acquire, while high-cost perpetual access models may limit access to current bestselling titles.


Title Availability

While all titles could be purchased in print through online vendors—resulting in near 100% availability, even if out of print—we limited print to what was available through library vendors. We privileged high-quality print runs, but some titles were only available as inexpensive paperbacks. 

In the U.S., 479 titles (85.9%) were available in print. In Canada, 525 were available, for 94.4%. One hypothesis regarding this discrepancy is that Canada has an easier time purchasing some of the Australian/New Zealand/U.K. titles in the study because of publishing rights. That hypothesis is borne out by print publishers, such as Puffin, Penguin Group Australia, University of Queensland Press, and Harlequin Australia not publishing some titles in a U.S. edition. Differences in copyright laws might account for other differences.

To buy every available title in print, a U.S. purchaser would pay $5,289.27 (or $11.04 per title). A Canadian librarian would pay $10,704.69 (an average of $20.39). Why do Canadians pay nearly twice the average per title? It isn’t the exchange rate. Converting to U.S. dollars at today’s exchange rate, Canadians would still pay $15.29 per title on average. The chosen vendor is an unlikely factor. Any follow-up study might look across several print vendors and average among them or use the lowest price.

Ebook Availability

For Canada, 468 ebook titles (out of the 556 total) were available, down from 525 in print, which is 57 fewer, or about 11% less. Eighty-eight total titles were unavailable in ebook form. Some of the titles available in print, but not as ebooks, are of greater significance in Australia (for example, a Colleen McCullough title and many children’s titles). Their publishers, perhaps, do not think demand is high enough to license them. Maybe Canada’s copyright laws allow for print access that publishers do not wish to give for licenses. In total, 94.4% titles were available in print, but only 84.2% as ebooks. The availability of physical titles was identical to that found in the Monash e-lending study (94%). Interestingly, ebook availability in our snapshot was greater than that reported in the Monash e-lending study of Canada, which found only 70% of titles available. The difference is perhaps due to the original study using only one vendor for Canada. 

For the U.S., only 10 fewer titles were available in ebook form, 469 down from 479 in print. Print works were available at 85.9% and ebooks at 84.2%. The number of available ebook titles in the U.S. and Canada was nearly identical, even though the actual titles available in each country were different. The number of unavailable titles in both countries was identical: 88. As with the available titles, the unavailable titles vary between the two countries. Unavailable titles show one disturbing trend: In the U.S., fully 40 titles were not available to library ebook readers but were available to Amazon users via Kindle, Audible, or both formats. The availability of Amazon titles in Canada has not been assessed here, but there is reason for concern in this country too. Two more titles were self-published and not available, although the ability of vendors to provide such content in the future is vital given market trends, with smaller publishers and independently published titles occupying a growing amount of what readers consume.

Publication date is somewhat of a factor in availability, but not as much as might be presumed. Forty-seven titles in the study were published before 1970. Of those, only eight were unavailable in the U.S. However, the unavailable titles from before 1990 include six Pulitzer Prize winners or nominees and other significant titles:

  • A Summons to Memphis , Peter Taylor (1986, Pulitzer)
  • Elbow Room , James Allen McPherson (1977, Pulitzer)
  • Guard of Honor , James Gould Cozzins (1948, Pulitzer)
  • Humboldt’s Gift , Saul Bellow (1975, Pulitzer)
  • The Elected Member , Bernice Rubens (1969, Booker Prize)
  • The Way West , A.B. Guthrie (1949, Pulitzer)
  • Tales of the South Pacific , James A. Michener (1947, Pulitzer)

Times and tastes change, and a work once thought to be essential might no longer seem compelling enough to be licensed. Nevertheless, our study suggests that building a collection including significant works from the past may be a challenge, especially (as shall be explored later) if they are available only on time-based license terms. In the U.S., 53 of the 88 unavailable titles (60.2%) were from 2013 or earlier. In Canada, 56 of 88 (63.6%) of titles from 2013 and earlier were not available. Since only 50.1% of all titles date from before that time, we conclude that older titles are less likely to be available in ebook form, including titles of significant literary merit. 

General/literary fiction is usually less likely to find a place. In the U.S., for example, 49 adult fiction titles were not available—28 of them (57.1%) could be described as general or literary fiction, compared to 45.3% of all adult fiction titles. Also less likely to be available were children’s works, especially those with pictures. In the U.S., 23.9% (21) of the unavailable titles were for children. Perhaps the lack of print’s tactile quality in ebook form limits demand and so, licensing. Some of the children’s selections were primarily of U.K. or Australian interest and perhaps thus less likely to be licensed. Interestingly, YA titles were more likely to be available when compared to other titles. Only seven of 60 total YA titles were available in the U.S. (11.6%) compared to the 15.8% of total books not available. Poetry and drama were only available at 50% in our titles sampled. One might hypothesize that they may be more difficult to acquire. But, our overall sample of them was small. 

Differences exist in what is available in the two counties, although each had 88 titles not available. Thirty-two titles were different, with 16 each not available in one county but present in the other. Canadians may not enjoy Oscar Hijuelos’ The Mambo Kings Play Songs of Love or Iris Murdoch’s The Sea, The Sea (although interestingly, they have access to another Murdoch title). They can, however, enjoy Saul Bellow’s Humboldt’s Gift. It seems unusual that Americans cannot read a Pulitzer Prize-winning novel by one of their fellow Americans, but perhaps it is because Saul Bellow was Canadian American. Availability seems unconnected to potential demand, with publisher decisions driving licensing. 

While availability of ebook titles may lag a bit in Canada when compared to print, having a mean of 85% of titles available for both counties in ebook form when compared to print is encouraging.


Ebook vs. Print Availability

Greater availability of titles means that building a popular collection of high-demand ebooks is more possible than perhaps it was even 5 years ago. Building a robust ebook collection is roughly comparable to building print collections, if only availability is considered. Many older, but still popular and/or significant, titles are available digitally and can provide collection depth. However, collecting ebook titles of less popular interest may be a challenge, especially in poetry, drama, and literary fiction.

Amazon is now a major contender in digital content. Unfortunately, Amazon does not allow libraries to purchase titles from its growing collection of unique holdings, even though some of them make it onto various bestseller lists. This is often due to exclusive licensing, which occurs in ebook and digital audio. Since Amazon has exclusive rights to a variety of titles, libraries’ ability to curate and preserve in digital format is being challenged.


Buying one copy of all U.S. ebook titles would cost $16,597 (at an average of $35.54 per title), which is more than three times the cost per title for print. For Canada, one copy of each would cost $17,745, at an average cost of $37.92 per title—which is better when compared to print costs for the Canadian titles, but still nearly double the cost of print titles.

Ebooks in the study sometimes cost less than print books—and sometimes far more. The least expensive U.S. title in the study cost $2.27. The most expensive cost $90.95. In Canada, prices ranged from 99 cents to $124.04. Bestsellers generally occupy a high middle ground well above the average. A surprising occurrence is that less popular titles are often not lower in price. The Executioner’s Song (1979) by Norman Mailer costs $14.02 in the U.S. in print, but $75 for the ebook from all three U.S. vendors. An older publication date is no guarantee of a low price: Of the 47 titles published before 1970, 24 are priced at $50 or more in the U.S.—well above the average cost of $35.54. The Autobiography of Mark Twain, Volume 1 is priced at $90.95 for a one copy/one user perpetual license. It contains Twain’s uncensored autobiography in its entirety and exactly as he left it, but little else differentiates it from the free version available in Project Gutenberg. The cost is likely to make it more desirable for academic than public libraries. There is a correlation between cost, currency, and popularity: Bestsellers from the past 5 or so years generally fetch $55 to $78, with one copy/one user titles being pricier than metered access. 

It is no surprise that ebooks are more expensive than print books for libraries. How much more overall can perhaps be determined by cost-per-use per title. Although calculating such figures is tricky, a few examples suggest ebooks are, in practice, even more expensive than might be supposed. The first person to check out a print book might damage or never return it. Ebooks are immune to such mishaps; publishers sometimes justify a higher ebook cost for libraries than for consumers by arguing that digital never wears out. So unlike with print, libraries don’t have to buy more (although that argument is largely invalid as more publishers move to metered licenses). However, a hardback book might last 100 circulations and often does. In that scenario, costs may be pennies per use. In both formats, some titles may never circulate well and are costly on a per-use basis.

In general, however, ebooks seem less likely to offer as good an ROI. They may offer low cost if a title is inexpensive and available on a one copy/one user perpetual license. The aforementioned The Executioner’s Song is available in such a license. Its initial cost of $75 nevertheless makes it seem less likely to equal a print investment. A print version of the title at $14 could easily last 5 years or more. If it circulated 14 times, its cost would be $1 per use. Would this older ebook title likely circulate 75 times in 5 years, or even 10 times in most libraries? 

Metered licenses complicate the matter even more. For example, The Collected Stories of Katherine Anne Porter costs $10.95 in print in the U.S. All three U.S. ebook vendors sell it for $40 on a 24-month license. This title is culturally significant. If it had a guarantee of 52 checkouts in 24 months, costing 77 cents per use, many librarians would likely purchase it. If it circulates 10 times in its license, its cost-per-use is $4 for the ebook. For print, the cost would be about $1.10, and the book would still be available. Many librarians might balk at purchasing either of these digital titles, thinking them unlikely to have high enough circulation to justify the cost. Many titles that are available may not, as the E-lending Project group has averred, be accessible under publisher terms. Moreover, with short-term licenses, can we even guarantee titles will be available long term?


Ebook vs. Print Costs

While we might never expect library digital cost-per-title to be as low as print, the metered models’ higher title costs suggest that cost-per-use for digital on average is certainly higher—likely much higher—than print. Even if cost is only about $1.13 per digital use (on, say, a   52-circ license for a popular title at $59—not uncommon) rather than pennies for print, it is clear that cost may be a barrier to delivering library service efficiently.  

Even high-demand titles on 2-year/52-circ models are unlikely to return as well as print, although at least they are likely to get the full use that their license allows. Our ability to afford older titles of significant merit, or even newer but perhaps less-popular, works is likely limited. Are current publisher practices squeezing libraries into a popular-title-only collection model, even when we might wish to provide and conserve more in digital?


Substantial cost differences exist across vendors per title. The U.S. has more uniformity than Canada. Many prices are the same for all three U.S. vendors. Some variation nevertheless exists: $87 for a title versus $55, $78 versus $55, and a high of $81 versus $11.56. In Canada, prices per title vary across vendors much more often and perhaps even more in degree: $124.04 versus $12.95 and $107.97 versus $21.72. Why, for Charlene Harris’ Dead in the Family, do two U.S. vendors charge $55 when the third charges $9.99? There is clear pattern that the prices vary—but why?

Are publishers giving discounts that not all vendors offer? Is it different vendors acquiring access to books at different times and never updating their cost data? Do publishers quote different prices to different vendors? The issue seems well worth exploring, but it would require querying the vendors. Without such information, pricing can seem capricious, unconnected from a work’s age or the current demand for it. 

Many bestselling authors’ titles are offered at a standard high rate, but other titles may vary considerably. In some instances, publishers are not being consistent in their rates, even for bestselling authors. For example, David Baldacci’s Memory Man (2015) from Hachette is $84 for all vendors in both countries. From the same publisher, Nicholas Sparks’ Safe Haven (2010) comes in at $78, while Donna Tartt’s The Goldfinch (2013) rings up at $90. Why the difference? All of them are one copy/one user licenses, which explains why they are much more expensive than John Grisham’s Camino Island (Random House, 2017) at $27, with its 24-month metered license. In our findings, cost variations seemed more frequent across vendors in Canada than in the U.S. In the latter, only 11 titles had at least a $5 variance across vendors. In the former, the figure was 69 titles. The difference is striking. 


Cost Variations Among Vendors

The price for the same title can vary substantially from vendor to vendor. Older titles don’t necessarily have low pricing. Collection librarians must ask their vendors questions. For example, when a publisher offers a sale, does the vendor advertise it and lower prices to match? What are the costs compared to publisher list prices?  

Vendors may charge more (or less) than the publisher asking price. Libraries that can draw from more than one vendor may achieve lower costs, with the obvious drawback of having to shop for bargains and then direct users to different vendor platforms to discover titles.   A vendor, perhaps a nonprofit one, that always guarantees offering the publisher sales price might have a market advantage.


In Canada, the number of titles offered by the three vendors was 331, 403, and 440. Obviously, one vendor had a bigger selection than the other two and one far less. Vendor offerings varied less widely in the U.S.: 428, 426, and 451. Significantly, no single vendor offered all titles in this study. 


Coverage Variation Among Vendors

Having more than one vendor is not only advantageous in terms of cost, but it also provides the best chance to offer the most titles. In both the U.S. and Canada, using a combination of vendors two and three will get most available titles. In both countries, however, one must use all three to get 100% of the titles. Using the offerings of just vendors two and three will only leave out one or two titles in either country; however, extrapolating our title sample to a much larger sample may reveal considerable gaps in coverage if a library uses only two of the three vendors.  

The need for an aggregated platform that offers access across vendors, such as Library Simplified’s app, may be beneficial for ease of use and prevent content deserts in which titles languish due to users choosing one favorite platform. Having multiple vendors will require more work for collection staffers. Smaller libraries with limited resources may be forced to use only one vendor and do the best they can.


Of the 469 available U.S. titles, 159 (34%) were offered in a one copy/one user model. Six titles had major discrepancies (1.3%) across vendors—when one vendor listed a title as one copy/one user but one or two other vendors listed it as metered. In Canada, 136 titles out of 468 were one copy/one user (29%), but 18 discrepancies (3.8%) existed. These discrepancies might have been due to Penguin Random House (PRH) moving from one copy/one user to metered access on titles during our data-gathering period. Although not an issue for U.S. titles, in Canada, five of 18 titles were from PRH; vendors were possibly in the process of changing license terms as we collected data. Vendor one reported more titles from Scholastic as one copy/one user than did the other two. The reason for these variances is unclear and worth considering, although again finding an answer would require querying vendors and publishers. 

Especially noteworthy, however, is the difference between the U.S. and Canada, with metered access being 5% higher in Canada. Differences in reported license models across vendors in each country are still a less significant occurrence than in the original e-lending study of five vendors in Australia (in which 22% of the titles had discrepancies). Those differences, however, are likely due to one of the vendors in that study not having changed its offerings after a big shift in one publisher’s license model. 

The one copy/one user model, once the dominant example, seems to be waning, with 66% and 71% of the titles in the U.S. and Canada, respectively, offered currently with a meter of some kind. Metered models offer advantages due to the almost universal lower price points, allowing for more copies to be purchased when demand is high. Unless a publisher offers a buyback of titles (apparently, this practice sometimes happens with one copy/one user titles, but librarians don’t always know about it), meeting patron demand for high-interest titles in the one copy/one user model can be a challenge for the opposite reason. A library may have to buy many copies of the same title, which may go unused when the demand ebbs. For popular works, metered licenses are often likely to have their terms completely fulfilled, thus providing a good ROI. Having only metered access means constant re-evaluation and rebuying and possibly a complete loss of access if a title is ever discontinued. For less popular titles, moreover, a metered license can be detrimental, with the possibility of it expiring before the maximum number of circulations is fulfilled or time runs out with few checkouts.

Some variations in metered licensing terms exist across vendors. For example, in Canada, George R.R. Martin’s A Dance with Dragons was offered by one vendor at 24 months or 52 circs, whichever comes first, while the other two offered it at only 24 months. East of Time by Jacob Rosenberg was offered by one vendor at 36 circs/24 months and another at 52 circs/24 months. Why the difference? Thirteen such discrepancies existed in Canada—perhaps the most significant being when one vendor licenses for 12 months and another for 24 months at the same price. Only one such discrepancy appeared in the U.S., with two vendors listing a title at 24 months and another at 12. Variation in metered licensing across vendors would not seem to be a significant factor in either country, but Canadian selectors with multiple platforms to choose from might wish to compare. 

The actual license terms can be problematic. In the U.S., 196 titles were licensed by time only—12 months (27 titles) or 24 months (169 titles) were the most common terms. A mixed license (24 months or 52 circulations, whichever comes first) accounted for 64 titles. Thirty-nine titles were available for 26 circulations without a time limit (Canada has no such licenses). In Canada, titles were licensed thusly at 12 months (24), 24 months (180), 24 months or 36 circs (19—the U.S. has no such licenses), and 24 months or 52 circs (39). If demand is not high in a license period, cost-per-use can be high. Such titles also require frequent relicensing. Librarians may be hesitant to buy tiles on a 1-year or even 2-year license unless they are of guaranteed popularity. A 26-circ-per-title model at least gives the U.S. a guarantee of cost-per-use and is perhaps a license worthy of greater adoption. Both publishers and libraries would know that a book that costs, for example, $52 to license for 26 circs would eventually cost $2 per use. It would be up to the ordering librarian to decide if that cost-per-use was acceptable. This alternative seems better than never being certain about cost-per-use, as is the case with time-bound metered models.


The Impact of Licensing Models

While our study shows little variation in licensing models across the vendors, the increasing dominance of the metered license, especially those based only on time periods and not number of circulations, presents issues for trying to build rich collections of long-term significance in a time when we are increasingly challenged to meet customer demands for content in digital formats.  

The challenge grows when one considers the high number of non-bestselling titles that are available in metered access only. Of titles not officially listed as bestsellers, but that are popular enough that they might perform well within the restrictions of their metered licenses, Canada had 197 and the U.S. had 146. Will such books obtained in metered licenses be cost-effective? Titles licensed by time and not number of circulations, especially if the time period is only 1 year, seem particularly problematic. Reordering of these titles, even if they perform well within the constraints of their license terms, seems inevitable.  

The last 2 years have, as a recent Monash update study has established, seen at least the Big Five move further away from the one copy/one user perpetual model. Out of the 48,794 Big Five ebooks the team studied for one vendor in the U.S. in 2017, 16,500 (33.8%) were available in one copy/one user perpetual licenses. In 2019, that number had declined to 199 (0.4%).5  In our study, bestseller Stephen King’s works are on 12-month licenses. If publishers offered a variety of license models at point of sale, libraries could best use their budgets optimally while offering publishers and authors revenue.

Offering titles simultaneously in metered (cost-per-use or time/circ) and one copy/one user models would be a good start. One copy/one user models might be slightly more expensive to reflect their permanence in the collection. Many smaller publishers are offering a variety of licensing models at the point of sale. Large publishers might find that they sell more to libraries if they adopt a similar variety of options.  

The recent decision by Macmillan to embargo sales to libraries suggests that at least some publishers view library sales as problematic.6  It is noteworthy that it offered—although under terms that libraries find onerous—multiple license models. It is to be hoped that advocacy from libraries and negotiations with publishers can result in the implementation of multiple models under favorable terms for both.


Libraries in the U.S. and Canada have nearly the same access to titles in ebooks (at least for those sampled) as they do in print, with Canada having slightly richer print offerings. Libraries are nevertheless challenged when trying to offer a broad and deep digital collections that would match print holdings. Our study, while differing from the original five-vendor Australian study in the number of discrepancies across vendors, validates its conclusions that availability does not equal access, due to restrictive combinations of price, licensing models, and relative demand for titles. 

Costs for ebooks, especially cost-per-use, are likely to be much higher than for print, even for older titles. Library budgets do not go as far in the digital realm. Variations in availability of titles, pricing, and licensing terms exist across vendors. Libraries that have the resources (materials budget, employee time) to use more than one vendor may be able to realize cost-savings or more favorable access terms and might certainly build a broader collection. They will face challenges in keeping up with the workload, and the need for improved and standardized APIs (as with NISO’s forthcoming FASTEN project)7 to streamline cross-platform access—or for an app such as Library Simplified8 to improve discovery across platforms—has never been clearer. 

No single digital content vendor will ever be able to offer all the titles that library customers seek. Less-well-funded or smaller libraries may face greater challenges. Unless they belong to a consortium, they may not have the employees or funding to look at multiple platforms. The current trend to move toward metered licenses, and even toward embargos on library sales,9 with the cost of less-popular, older items will present challenges to all libraries. Are we being forced by business models and costs into boutique ebook collections, offering only titles that are likely to earn their keep and forever reordering the most popular authors?

With the unlikelihood of legal action from governments to extend fair use or other copyright advantages into the digital realm or to legislate prices10 (although now called upon by library associations in both countries), a good first step forward would be for libraries, vendors, and publishers to agree on multiple license models that will allow libraries to make best use of their funds while still respecting authors’ rights, publishers, and vendors. Some smaller publishers already offer such models. It is strongly recommended that large, prominent publishers also offer multiple license models. If libraries could at least make better use of their funding through multiple models, balancing the needs of library readers, publishers, and authors might be more easily accomplished.

While the increasing number of titles that are exclusive to Amazon and not available through libraries is troubling, this trend does point to a possible new direction. Since it is unlikely an agreement with Amazon is forthcoming and since large publishers may offer somewhat cost prohibitive options, libraries are encouraged to take advantage of the burgeoning independent and small-publisher ebook market to broaden their collections.

No matter what else these data show, the need for librarians to talk with their vendors is clear. Some questions to pose are as follows: What are the reasons for cost variations? Do the vendors get deals or news of buybacks (although buybacks don’t exist for metered titles)? Do they share this news and allow libraries to take advantage? Perhaps above all, with some large publishers making license changes that do not necessarily help libraries and are seemingly unwilling to talk with libraries, how might the vendors join with libraries to advocate for a more robust digital experience?

While our study is suggestive, it would be interesting to test its findings with a larger sample. A study that extends our title list to include all vendors providing ebooks to public libraries—including DPLA Exchange, Ingram, and ProQuest—would be a first step. The 100,000-title Monash study conducted for one vendor in the U.S. and Canada could be extended to our three vendors or all vendors in the two countries. Narrowing that list to titles originally published in the U.S. and Canada might give the truest picture of how print and ebooks compare in availability. A truer picture still might result by comparing offerings from many print vendors, as well as focusing on Amazon. A broader study would have to be coordinated with many libraries, and getting access to all the vendors, with agreements from each, would be a challenge—not to mention the work of gathering and analyzing such a large dataset. The larger sample would, however, provide evidence of the challenges libraries face when building digital collections. We believe our findings would be validated. 

A smaller study that focused solely on titles for children and teens, or on different genres across our three vendors or many more vendors, could provide a more definitive look at the offerings for that particular class of readers, which appears to be underserved with ebooks. 

There is always more work to do. Let’s get to it. 


ReadersFirst ( is a loose international group of libraries that banded together in 2012 and 2013 to advocate for a better library digital content experience. The ReadersFirst Working Group, with some 30 members from Canadian and U.S. libraries and nonprofits such as the Digital Public Library of America (DPLA) and LYRASIS, undertakes projects pertinent to the group’s mission.  

Contributors to the U.S. and Canada study were Michael Santangelo of BookOps, who collected data on two U.S. vendors; Catherine Mason of Columbus Metropolitan Library, who worked with another U.S. vendor; Trina Jacobs of St. Mary’s County Library, who surveyed print titles in the U.S.; and Susan Caron and her team from Toronto Public Library, Nancy Peel from Windsor Public Library, and Ann Archer and Monique Brûlé from Ottawa, who got data from print vendors and three digital content vendors in Canada. Full disclosure: St. Mary’s Public Library allowed access to the Monash e-lending project for one vendor for U.S. data.  

The two papers cited at that start of this article were also produced by members of the ReadersFirst Working Group. We thank Rebecca Giblin and her team for their seminal research.


1. See the work of Rebecca Giblin and her colleagues at

2. Giblin, Rebecca, et al. 2018. “Available—But Not Accessible? Investigating Publisher E-lending Licensing Practices.” Sydney Law School Research Paper No. 19/20.

3. Giblin, Rebecca, et al. 2019. “What Can 100,000 Books Tell Us About the International Public Library E-lending Landscape?” Sydney Law School Research Paper No. 19/21.

4. For the methodology of title selection in the e-lending project, see Our Methods.

5. Giblin, Rebecca. Emails to the authors. Aug. 1, 2019.

6. Albanese, Andrew. “After Tor Experiment, Macmillan Expands Embargo on Library E-books.” Publishers Weekly July 25, 2019.

7. Flexible API STandard for E-content NISO (FASTEN).

8. Frequently Asked Questions. Library Simplified.

9. Feldman, Sari. “Libraries Must Draw the Line on E-books.” Publishers Weekly July 12, 2019.

10. Albanese, Andrew. “ALA 2019: ALA Council Votes to Take E-Book Issues to the Public, Congress.” Publishers Weekly.

Michael Blackwell ( is director of St. Mary’s County Library and project and communications coordinator for ReadersFirst. A member of the ALA Digital Content Working Group, he has frequently presented and published on public library ebook matters.

Catherine Mason ( is catalog and serials manager and digital downloads administrator for Columbus Metropolitan Library. She has extensive experience in ordering and licensing titles for a large public library consortium. 

Micah May ( is ebooks consultant for the Digital Public Library of America (DPLA). Before joining DPLA, he was director of strategy and then senior director of business development and innovation at the New York Public Library (NYPL). At NYPL, he built capacity for innovation and led many groundbreaking projects (including the creation of the NYPL Labs, Open Ebooks, the MyLibraryNYC partnership with New York City schools, and the open source SimplyE reading app).