Why Roll-Ups Fail: Research on Serial Acquirers

What counts as a roll-up, and what the studies measure

A roll-up buys many small companies in a fragmented industry and combines them into one larger business. A buy-and-build program is the private equity version: a platform company plus a sequence of add-on acquisitions. A serial acquirer is broader still: any company that buys frequently, whether or not the targets are small or in one industry.

The research on these strategies is often summarized as a single claim ("most roll-ups fail" or "serial acquirers win"). Neither claim survives a look at what the studies measured. They differ on four dimensions:

There is no single "roll-up failure rate" in this evidence, and none is offered here. Each statistic below is tied to its study, sample and date. Where the article draws a conclusion that goes beyond a source, it is labelled as an inference.

None of these studies measures private-equity-backed buy-and-build platforms directly. Most track public companies, because public stock prices are observable. That gap matters when applying the findings to a lower middle market program, and it is the main reason to treat them as signals about mechanisms rather than as base rates.

The case for serial acquirers: what the favorable evidence says

Buying private targets: Fuller, Netter and Stegemoller (2002)

Kathleen Fuller, Jeffry Netter and Mike Stegemoller studied "shareholder returns for firms that acquired five or more public, private, and/or subsidiary targets within a short time period." Because the same bidder buys different kinds of targets, the design lets them compare deal types while holding the buyer constant. The abstract reports that "bidder shareholders gain when buying a private firm or subsidiary but lose when purchasing a public firm," and that "the return is greater the larger the target and if the bidder offers stock." The authors read this as "consistent with a liquidity discount, and tax and control effects" ^[1].

This is the study most often cited in support of roll-ups, and the reason is clear: roll-ups buy private companies. Two cautions apply. The outcome is the bidder's stock return around the announcement, not the later operating result of the combined business. And the acquirers are public companies, not sponsor-backed platforms. The finding supports a narrower claim: when a public serial acquirer announced a private-company purchase, its shareholders, on average, reacted favorably.

Frequent acquirers in Bain's data

Bain & Company's April 8, 2024 brief "How Companies Got So Good at M&A" states: "Frequent acquirers have a 130% advantage in shareholder returns over non-acquirers—it was 57% in 2000–2010." The article defines a frequent acquirer as a company doing at least one deal per year and says "Hyper-acquirers (companies that do five or more deals a year)" earn "an additional 20% boost" ^[2].

How to read it:

• It is a consulting firm's own analysis, presented in a short article. The page does not describe the sample, the return measure or the statistical method in detail, and it does not display a publication date.
• It compares acquirers with non-acquirers. It does not tell you whether a given acquisition program created value, or what happened to companies that tried frequent acquisition and stopped.
• The same article recalls that, around 2004, "most academic studies found that 70% of all mergers failed," and reports that "Today, executives report that close to 70% of deals are successful." Both are characterizations, one of a literature and one of executive self-reports. They show how much headline "failure rates" depend on who is measuring and how.

Bain's explanation for the gap is experience: the article calls it "the No. 1 predictor of a successful acquirer" and credits better diligence and integration practice. That is the consulting firm's interpretation of its data, and it is compatible with the next sections, which show the same mechanism failing when experience is thin or overconfidence sets in.

The 1990s roll-up wave: what went wrong

The best-documented evidence on roll-ups as a distinct strategy comes from the U.S. roll-up IPO wave of the mid-to-late 1990s. Both sources below are old, and the market they describe (roll-ups financed through public equity) differs from today's sponsor-financed programs.

Roll-up IPOs: a Purdue dissertation

A Purdue Ph.D. dissertation by Muge Tiryakioglu, "Industry consolidating IPOs," examines "58 firms that conducted a roll-up IPO over the 1994–1999 period and 55 build-up IPOs that went public between 1983 and 1999." The repository page shows no completion year; it is identified here by its ProQuest number, AAI3105033. Its abstract reports ^[3]:

• The long-run stock price performance of roll-up IPOs "significantly underperform[s] several market benchmarks, including build-up IPOs, in the three years after going public," while build-ups were not significantly different from the market.
• "The poor stock price performance of roll-ups is not accompanied by poor accounting performance": operating performance of roll-ups improved significantly in the year after the IPO.
• Founding companies and sponsors "decrease their ownership stakes significantly" after the IPO.
• Sample companies pursued "aggressive acquisition programs" after going public; "all post-announcement stock returns are significantly negative whereas pre-announcement returns are positive," which the author reads as acquisitions timed to periods of superior performance.
• The conclusion: roll-up IPOs "have failed to deliver on their promise of creating shareholder value through consolidation and that their performance is not representative of industry consolidators in general."

The split between stock returns and accounting performance is the useful detail. Operations improved while shareholders lost. One reading (an inference, not the author's claim) is that the problem was the price investors paid for expected consolidation gains, not an absence of gains.

"Strategic Rollups" in strategy+business (2000)

An April 2000 strategy+business article by three Booz Allen Hamilton consultants, Paul Kocourek, Steven Chung and Matthew McKenna, states: "close to half of all rollups that we studied have lost 50 percent or more of their market capitalization since June 1998" ^[4]. The article does not set out its sample or method in the text read for this review, so the figure is best treated as a contemporary practitioner observation about one market cycle.

The article's diagnosis is more durable than its statistic. The authors write that "the problem is not the concept but the execution," and that in "the vast majority of rollups we've examined, the problems center on" the integration stage. They list five problems that "cause rollups to unwind":

1. Deal-makers dominate management, with too little capacity to consolidate the business.
2. Owners remain independent, often because earn-outs keep acquired units separate so performance can be measured, which "undercuts integration."
3. A sense of urgency is lost among former owner-operators.
4. Stakeholders have competing agendas: investors, management and former owners pursue different initiatives.
5. Information systems are hard to integrate across dozens of small companies.

The article also observes that many roll-ups were rewarded by the stock market "until they reached roughly $500 million in revenues," when investors began to "probe deeper." That is the authors' observation about public roll-ups in the late 1990s, not a general threshold.

Repeat deals and overconfidence

A different strand of research asks whether experience makes acquirers better or merely more confident. Matthew Billett and Yiming Qian examined "the history of mergers and acquisitions made by individual CEOs." Their abstract reports three findings ^[5]:

1. "CEOs' first deals exhibit zero announcement effects while their subsequent deals exhibit negative announcement effects."
2. Acquisition likelihood rises with the performance of earlier deals, but "previous positive performance does not curb the negative wealth effects associated with subsequent deals."
3. CEOs' net purchases of their own company's stock are greater before subsequent deals than before first deals.

The authors interpret this as "consistent with self-attribution bias leading to overconfidence": managers credit themselves for early success and buy more. They also find that the market anticipates future deals from a CEO's history.

This does not contradict Bain's experience argument so much as qualify it. Repetition can build skill or build confidence, and announcement returns alone cannot tell which. For a buy-and-build program, the practical point (an inference) is that the add-ons after a successful first few deals deserve the same discipline as the first, and possibly more.

Integration and platform failure drivers

Read together, the sources point to a few recurring mechanisms. This section is synthesis; the sources support the individual observations, and the grouping is this article's.

Illustrative example (fictional). Harborline Capital forms a commercial HVAC services platform, Keystone Air, and closes 12 add-ons in 18 months. Each deal passes diligence on its own. The integration team is three people. By month 12, four acquired companies still run their own dispatch software, two former owners are on earn-outs tied to their unit's stand-alone profit and resist sharing technicians, and month-end close takes three weeks because charts of accounts differ. Management cannot say which branches are profitable after allocations. The deal team, encouraged by the first five deals, has six more letters of intent outstanding. Nothing in this example is unusual, and nothing in it shows up in any single deal's diligence. It appears only when the program is viewed as a whole, which is the level at which the strategy+business authors located most failures.

Regulatory risk is part of roll-up risk

Antitrust exposure is a program-level risk for the same reason. The 2023 Merger Guidelines issued by the FTC and DOJ include Guideline 8, which states: "A firm that engages in an anticompetitive pattern or strategy of multiple acquisitions in the same or related business lines may violate Section 7" ^[6]. Small add-ons below the reporting thresholds can still be examined as part of a series.

The leading recent example is the FTC's September 2023 suit against U.S. Anesthesia Partners and Welsh Carson, which alleged a roll-up of anesthesiology practices in Texas; the release describes allegations, and the case "will be decided by the court" ^[7]. The FTC later finalized an order with Welsh Carson that requires notice of specified future acquisitions in anesthesia and other hospital-based physician practices ^[8]. In April 2026 the FTC announced an agreement in principle with USAP whose terms were confidential at the time ^[9]. On May 26, 2026 the court stayed the case for 180 days, through November 18, 2026, while the settlement is implemented; as of October 28, 2026 the public docket showed no proposed final order ^[10]. No court has ruled on the roll-up allegations.

Thresholds, the HSR form and the status of these matters change; check the current agency pages before relying on any of them. This is a description of public materials, not legal advice.

A roll-up fit screen

The evidence suggests testing a roll-up thesis at the program level before testing any single target. The questions below turn the mechanisms above into a screen. Answers should cite evidence and carry a status (verified, unverified, open). They can sit alongside a firm's other screening criteria.

Illustrative scoring (fictional). A team screens a thesis for regional pest control. It scores each question 0 (weak or unknown), 1 (partial) or 2 (strong evidence):

Total 10 of 20. The score itself means little. The zeros are the point: the two weakest answers (add-on pricing and integration capacity) are exactly where the 1990s evidence located losses.

What origination teams can do about it

The research has one consistent lesson for the people building target lists: program risk comes from decisions that look reasonable one at a time. Origination can make those decisions more visible.

• Write down the thesis evidence. For each add-on thesis, record the fragmentation evidence, the assumed value source and the pricing evidence, with sources and dates.
• Log every deal against the program. Keep closed, abandoned and pending deals in one place so integration load and local share can be seen in total.
• Keep gaps open. If share, pricing or owner intent cannot be supported, say so in the record rather than filling the field with an estimate nobody can trace.
• Rescreen when assumptions fail. If add-on prices rise, integration slips or the regulatory picture changes, re-run the screen on the remaining list; see re-screening targets when criteria change.

Docket's research agents answer a client's screening criteria as typed values, each quoting a retained source page, and leave unanswerable questions visibly open.