Follow the Dollar: How Library Ebook Pricing Actually Works

What to know: A library ebook is usually a license, not a purchase. The price and the term can differ sharply from the consumer edition. The useful question is not "Why does this one title cost so much?" It is: what did our library pay, how long does access last, and how many checkouts did that license deliver?

Three numbers worth pulling from an ebook invoice: the price per license beside the consumer price, the license term (perpetual, checkout-limited, time-limited, or both), and the cost per checkout before expiration.

One Book, Three Prices

Pick up Kristin Hannah's The Women at your local bookstore. It costs $18.

Buy the Kindle edition. $14.99. You own it forever.

Now check what a library paid for the same book in digital form: $60 for a 24-month license. When that term ends, continued access requires another license.

Here's what real libraries are actually paying, documented by Timberland Regional Library in Washington State:

Title	Physical book	Library ebook license	Consumer Kindle
The Women (Kristin Hannah)	$18.00	$60.00 / 24 mo	$14.99
Listen for the Lie (Amy Tintera)	$17.00	$60.00 / 24 mo	$13.99
The #1 Lawyer (James Patterson)	$18.00	$65.00 / 24 mo	$14.99
Toxic Prey (John Sandford)	$19.00	$27.50–$55.00 / 12–24 mo	$14.99
Toxic Prey (audiobook)	N/A	$95.00 / 24 mo	~$20.00

The point of the table is not that every library title follows one formula. It is that a library's access terms are a separate product, with a separate price.

According to a 2025 Computers in Libraries study, the average library ebook license in its sample cost $54.62, compared with an average Kindle price of about $13.66. That comparison is useful as a market snapshot, not as a multiplier for an entire collections budget.

The Markup

The Big Five do not use one uniform model. Before comparing prices, compare the term attached to the price.

Publisher	Model	Typical library price	Consumer price	Multiplier
HarperCollins	26 checkouts	~$40	~$13	3.1x
Hachette	24-month term	$55–$68	~$14	4–5x
Macmillan	24-mo / 52 checkouts	$55–$60	~$14	4x
Penguin Random House	24-month term	$20–$65	~$14	1.5–4.5x
Simon & Schuster	12–24-month term	$59.39	~$14.63	4.1x

Source: ReadersFirst Publisher Price Watch, Good e-Reader pricing breakdown. Source list

Correction, July 30, 2026: an earlier version listed Simon & Schuster at $20–$30 and called it the most reasonable of the Big Five. That was wrong. ReadersFirst's price table puts it at $59.39 as of March 2026. The table above has been corrected.

HarperCollins pioneered the 26-checkout model in 2011. Other publishers use time limits, or a combination of time and checkouts. A higher price does not tell you much on its own; you need the term.

Tip for a board packet: Put one local title next to its consumer price, license term, checkouts, and cost per checkout. Your own invoice is more persuasive than a national average.

What an expiration term changes

When your library buys a physical book for $18, it sits on the shelf until it falls apart, typically lasting 50+ circulations over years of use (a common industry benchmark, though actual lifespan varies by binding and handling). You own it. You can repair it, donate it, sell it at a book sale, lend it to another library.

When a library licenses a $60 ebook for 24 months, continued access after the term requires a renewal or a new license. Some licenses also expire after a checkout cap; some use whichever condition comes first.

Here's the math, laid out plainly:

Ebook license: $60 every two years
Ten years of access: five license periods = $300
Physical paperback: $17, owned permanently, 50+ circulations
This is a comparison of access models, not a universal value judgment: the digital license also offers immediate remote borrowing and can serve a different audience.

One data point illustrates why term matters. OverDrive data cited by Jane Friedman showed Macmillan titles averaging 8.5 checkouts during a two-year license period, against a 52-checkout cap.

And 79% of those licenses expired because of the time limit, not the checkout cap.

In that data set, the time limit was the binding condition. It does not prove the purpose of every license model, but it does show why libraries should track expiration reasons rather than assuming a checkout cap measures actual use.

Source: Jane Friedman, citing OverDrive data on Macmillan lending patterns.

The same market report found that perpetual-access terms represented 34% of Big Five library ebooks in 2019 and 15% in 2024. That is a meaningful shift in the options a library can buy. Source: Words and Money.

Prices can move quickly

ReadersFirst documented the following 2024 library-price increases:

HarperCollins: +15% on library ebooks, +8% on audiobooks, in a single year.
Macmillan: +20% on ebooks in a single year.
Hachette: +4% on ebooks, +20% on audiobooks.

Tip for acquisitions: Compare the same 10 frontlist titles in two years of invoices. Separate price changes from term changes. A lower price can still cost more if the access term shrinks.

A five-minute invoice audit

Pick five titles that had a long waitlist.
Record the license price, term, and checkouts before expiration.
Divide price by checkouts.
Mark licenses that expired by time before reaching their checkout cap.
Bring that one-page result to the next board or budget meeting.

Use neutral language: “We paid $X for Y checkouts over Z months.” That is enough to start a useful conversation about collection strategy, consortium buying, and the terms a library should seek in its next renewal.

Questions worth asking your vendor

Which of our licenses expired by time, and which by checkout count?
Which backlist titles are available with perpetual access?
Can we export title-level price, term, and circulation data for our own analysis?
What changes at renewal: price, term, or both?

Sources

Disclosure: I build library software. To understand how these systems work I took them apart and rebuilt them: L/30, a library system, and MetisLib, a lending co-op. Both are open source, both run as sandboxes, and neither takes real patron data or is sold to anyone.