Two Private Equity Firms Behind Familiar Library Platforms

I drew the library-tech ownership stack on one page last week. The notebook took me three sessions, a primary-check audit, and a couple of admissions that I had been overstating things. The compressed version is shorter than you'd think.

Two firms recur in this map: KKR and Francisco Partners. This is a selected ownership map, not a market-share estimate for the whole library technology stack. Clarivate and Harris/Constellation, discussed below, have different ownership structures.

The DC Council just passed a bill that would meaningfully limit what those firms can charge libraries. Vote was 13-0, unanimous, on May 5, 2026. Mayor Muriel Bowser signed it on May 28, 2026, making DC the second jurisdiction in the country with such a law, a year after Connecticut. It cleared congressional review and took effect as D.C. Law 26-149 on July 18, 2026. A veto would have failed anyway: a 13-0 council overrides easily, and Bowser, who is not seeking re-election, was in the final year of her term when she signed.

The law’s contracting restrictions are conditional. The board must determine that substantially similar laws exist in at least ten states representing at least 50 million people, publish that determination in the D.C. Register, and wait 60 days. A tally of enacted bills alone does not establish activation.

Here's the picture.

KKR: the publisher and the platform

KKR closed on OverDrive in June 2020, buying it from Rakuten. Terms were never officially disclosed; Marshall Breeding inferred about $775M from Rakuten’s $410M purchase price plus its reported $365.6M gain. Three years later it bought Simon & Schuster for $1.62B from Paramount Global, announced August 2023 and closed that October. Same sponsor: KKR. Two different buyout funds, one firm. The vehicles differ; the owner doesn't.

Every time a library buys a Simon & Schuster ebook through OverDrive, two margin streams flow to the same firm. The publisher margin (whatever S&S retains on the library license) and the platform fee (OverDrive's cut on every transaction). KKR is on both sides of the desk.

Antitrust did look at Simon & Schuster, and did not see this. A federal court blocked Penguin Random House's $2.175B bid for Simon & Schuster in October 2022, after DOJ sued in November 2021. But that was a horizontal deal between two publishers, stopped on a buyer-power theory: DOJ argued the merged firm would pay lower advances to authors in the market for anticipated top-selling books. KKR bought the same publisher a year later and nobody framed it as a merger at all, because on paper it wasn't one. The storefront got blocked. The plumbing got bought.

One thing this is not: KKR did not invent that timer. Simon & Schuster set two-year library terms effective August 1, 2019, a year before KKR owned OverDrive and four before it owned S&S. The license length is the publisher's model. What changed is who collects at both ends of it.

Sari Feldman joined OverDrive's board on July 15, 2020, one month after KKR closed. Past ALA president. Past PLA president. Past Cuyahoga County Public Library executive director. At the time of her appointment, she was serving as a Policy Fellow at the ALA Public Policy and Advocacy Unit. That is common in PE-owned verticals, and it is also the kind of seat that puts a librarian's judgment in the room.

Francisco Partners: the K-12 stack

FP acquired Follett School Solutions in September 2021. Split it into two LLCs: Follett School Solutions LLC (Destiny ILS, Aspen SIS) and Follett Content Solutions LLC (Titlewave). One ownership, two operating entities. The split is documented; why it was drawn that way is not. Breeding notes the two LLCs "continue to be branded together as Follett School Solutions."

In December 2025 the two announced what their release calls "a sales collaboration": Follett customers order digital content through KKR's Sora, OverDrive's school app. Publishers Weekly put it harder three months later, "Follett Cedes Digital Delivery to Sora." Reading the split as what made that possible is my inference, not anyone's stated intent. Two separate PE firms commercially welded across portfolios. No shared directors. No coordination required: the structure does the work.

In the same eighteen months, both firms shipped AI features into their products. Destiny AI (built by Tribe AI on Anthropic Claude). OverDrive's Inspire Me generative recommender in Libby. AI-narrated audiobooks entering the OverDrive catalog through PublishDrive. Marshall Breeding's 2026 Library Perceptions Report, which dropped this week, has the demand-side data: zero libraries mentioned interest in AI technologies in respondent comments. The vendors shipped it anyway.

Building on Breeding

Marshall Breeding's 2026 reports just landed, and the data confirms the structural pattern in his own words and his respondents':

• Clarivate consolidation is in his counts. 786 of 2,220 responses came from libraries on a Clarivate product (Alma, Sierra, Polaris, plus the legacy Ex Libris and Innovative systems), more than any other vendor. That works out to 35% of respondents, which is my division, not his figure. These are self-selected survey responses, not an installed base.

• Post-acquisition pain is documented in respondent quotes. On SirsiDynix under Harris/Constellation (acquired December 2024): "Recent acquisition of SirsiDynix by Harris Corporation has lead to some issues regarding price increases, customer service issues during contract renewal, and the loss of some long time SirsiDynix employees." Another respondent reports a decline in support response time, with most development halted except the "BlueCloud" suite. These are librarians, not analysts.

• Breeding floats one explanation for why fewer public libraries are shopping: interest has declined "possibly due to the lack of compelling alternatives." From 29.3% shopping in 2008 to 9.2% in 2025. He offers a second reading in the same passage, that mature products no longer differ enough to justify the cost and disruption of moving. The series counts libraries considering a change, not migrations completed.

Breeding's data shows the pattern. The financial structure behind the pattern is what this post is about.

The squeeze, with the receipt

DCPL spent $1.62 million on ebooks in FY2025, per its executive director's written testimony to the DC Council. That is 34% of its collections budget. 54% of DCPL's total circulation is now digital: 3.74 million items annually. Average patron wait time for an ebook hold: 46 days.

One library license for David Szalay's Flesh: $59.99. What DCPL pays for a physical copy it keeps: $28.99. Library ebook licenses like this expire on a timer or a checkout cap, depending on the model: two years, or 26 checkouts. So the same book costs DCPL more than twice as much as a license that expires than as a copy it owns.

This is the metered licensing model. It is seven years old in its current Big Five form. Penguin Random House moving from perpetual access to metered, in October 2018, was the inflection point. Most current library directors started their careers in the perpetual-access era. The metered era is going to outlive the institutional memory of any other arrangement, unless something interrupts it.

Matthew Frumin (Ward 3 Councilmember, who introduced B26-0490): "Libraries have no leverage to negotiate in a meaningful way. This is all intended to create leverage."

Richard Reyes-Gavilan (DCPL executive director, on the pricing): "The demand is so great that the price we're paying is not sustainable."

Frumin is the same councilmember who reversed Mayor Bowser's FY26 budget proposal that would have cut DCPL branch hours from 61 a week to 40, five days instead of seven, and cut book purchasing on top of that. His committee put back $6,445,346 for the staffing that holds the hours, and a separate restoration for the collection.

The library-pricing bill is the next round of an ongoing relationship: when the Mayor cut, Frumin restored; when patrons need leverage against publishers, Frumin wrote the law. The 13-0 vote on May 5 was every other DC councilmember signing on.

The enacted law addresses specified licensing restrictions. Its price provision concerns an unconscionable disparity in comparable transactions by like buyers; it is not a blanket consumer-retail cap. The director must report quarterly on state enactments. The law also preserves specified licensing arrangements, including simultaneous-user limits. Its contracting approach does not settle how a court would resolve a challenge.

The forty-year compression

Between January 2022 and December 2025, the library lost: Maryland's reasonable-terms law (struck), Controlled Digital Lending as a model (Hachette v Internet Archive, decision plus 2nd Circuit affirmation; Internet Archive declined to seek Supreme Court review), Baker & Taylor as an ebook-lending competitor to OverDrive (collapsed late 2025), and the federal library measurement instrument (IMLS, nearly eliminated by EO 14238 in March 2025; restored by court order in November 2025).

Just before this window, Clarivate closed ProQuest ($5.3B including repayment of ProQuest debt, December 2021). Inside it: FP closed Follett. KKR closed Simon & Schuster. Vivendi/Bolloré closed Lagardère (Hachette's parent). Harris/Constellation closed SirsiDynix from ICV (December 2024). FP's Follett Content welded to KKR's Sora (December 2025).

Two patterns ran in parallel. The closures and the consolidations are not coordinated: different actors, different motives, different lawyers. They line up because the financialized library faces structural pressure on every alternative path at the same time. This map connects selected consolidations to the library workflows they affect, building on the trade reporting cited below.

That's what the notebook does. The map is on the blog if you want the full version.

Monday-morning actions, this Monday

Check implementation as well as enactment: look for the director’s reports and the board’s published determination. I have not established here that the restrictions have been activated.

Ask your director what percentage of the collections budget goes to ebooks. Ask what happens to license fees on titles whose 26-checkout cap has expired. Ask whether your state has introduced a bill like DC B26-0490 and what your vendor says will happen to availability if it passes. Get that last one in writing.

The Sari Feldman appointment is structural. The December 2025 weld is structural.

None of this is conspiracy. It's just a stack that got built, vendor by vendor, while the trade press tracked one deal at a time. The map is what changes when you draw it all on one page. The 13-0 vote in DC is the first to do that. The next eight states are the ones who finish the picture.

---

Sources

• Vendor ownership and the Perceptions data: Marshall Breeding, Library Perceptions 2026 and librarytechnology.org.
• Sari Feldman's board appointment and ALA role: OverDrive press release, July 15, 2020 (company.overdrive.com) and PR Newswire. The ALA Policy Fellow role is as stated in that release and may not reflect her current position.
• DCPL spending, circulation and wait times: Richard Reyes-Gavilan, Executive Director, DC Public Library, written testimony on B26-0490, Council of the District of Columbia Committee on Human Services, public hearing December 18, 2025, in the 78-page hearing record (memo of Chairperson Frumin, January 6, 2026): "In FY25 the Library spent $1.62 million on ebooks... 34% of our overall collections budget. DCPL circulated 3.74 million digital items in FY25."
• The Reyes-Gavilan and Frumin quotes: The 51st, "Libraries can't afford e-books. D.C. lawmakers have a plan.", January 15, 2026, which rounds the same figures to $1.6 million and "last year."
• The $6,445,346 staffing restoration and the 61-to-40 hours figure: DC Council Committee on Human Services, FY2026 Budget Draft Report, June 2025 (dccouncil.gov). Not from The 51st.
• B26-0490 trigger wording, the vote, the signing, and the law's effective date: Council of the District of Columbia, LIMS record for B26-0490. The certification record bound into the signed act shows the final reading on May 5, 2026 as a voice vote on the consent calendar with all thirteen members marked aye and none absent; the Mayor signed May 28, 2026 (Act A26-0327); Law L26-0149 took effect July 18, 2026. LegiScan, cited here earlier, carries the same dates but blocks most automated readers. Bowser's FY26 library-cut history and the Frumin reversal: WJLA and DC Council records.
• The Penguin Random House / Simon & Schuster block: DOJ, "Justice Department Sues to Block...", November 2, 2021, and the amended memorandum opinion in US v. Bertelsmann, No. 1:21-cv-02886 (D.D.C.), October 31, 2022.
• Clarivate/ProQuest close: Clarivate investor relations, December 1, 2021. Follett School Solutions sale: Francisco Partners release, September 1, 2021.
• New York veto context: Publishers Weekly, December 2021.

If you find an error, tell me. I retracted four reach claims from the working notebook getting here. The discipline is verification before pitching.