How often do roll-ups actually fail?
Often enough that the burden of proof sits with the acquirer. The published estimates vary by method and vintage and should not be averaged: more than two-thirds of roll-up strategies are reported to create no value for investors (Harvard Business Review research, carried by DealRoom, updated June 2026); 30 to 40 percent of private equity roll-ups underperform their initial expectations (CT Acquisitions, Private Equity Roll-Up Strategy guide, published June 2026); and only 30 percent of acquisitions achieve the combination benefits they were underwritten against, with 83 percent of practitioners in failed transactions citing integration execution as the primary cause (Bain and Company data from 2024, cited by Acquisition Stars, March 2026, labelled as historical context).
The most useful finding is not a failure rate at all. In one dataset of buy-and-build programmes, the cohort with two or more add-ons produced the lowest returns despite the fastest revenue growth (BCG and HHL Leipzig data, cited by CapitalPad, June 2026). Growth and return came apart in exactly the cohort where the strategy was pursued hardest.
That is worth sitting with, because it inverts the usual defence. The problem with an underperforming roll-up is rarely that it failed to grow. It grew. It simply spent more on growing than the growth was worth, in cash, in management attention, and in the leverage it consumed.
None of this is an argument against consolidation. Add-ons were roughly three-quarters of US sponsor buyouts in the second quarter of 2026 (PitchBook, Q2 2026 US PE Breakdown, 6 July 2026), and the strategy dominates the market because the arithmetic works when it is executed. The point is that execution, not arithmetic, is the variable.
How much of the return really comes from multiple arbitrage?
Less than the pitch implies, and it arrives only at the exit. The spread is real: average enterprise value to EBITDA runs from 5.9 times at ten to twenty-five million dollars of enterprise value up to 10.0 times at one hundred to two hundred and fifty million (GF Data, published 27 January 2026, covering the first nine months of 2025 and used as a labelled baseline). A platform buying at the bottom of that ladder and selling at the top captures the difference.
But the capture is contingent on three things happening in sequence. The acquired businesses have to be bought at the small-company multiple, they have to be genuinely combined into something a buyer will underwrite as one company, and the exit multiple has to still be there years later. Fail the second and the buyer applies a sum-of-the-parts view rather than a platform view, which returns the seller to the multiple they started with.
The market-level version of the same risk is currently live. Entry multiples reached a record 11.8 times EBITDA in 2025 (McKinsey Global Private Markets Report 2026, February 2026, labelled as measuring 2025), and the entry-to-exit spread compressed from roughly four turns across 2010 to 2020 to about 1.6 turns from 2020 (iCapital, February 2026). A programme underwritten on multiple expansion is underwritten on a narrowing quantity.
The honest description of a roll-up is therefore an operating strategy financed by an arbitrage, not an arbitrage decorated with operations. When the operating half is under-resourced, the arbitrage half does not survive on its own.
“Every roll-up I have seen fail was still growing on the day it stopped working. The revenue chart was the thing everyone pointed at, and underneath it were four accounting systems, three sales methods, two founders who had been promised autonomy, and a credit facility with no headroom left for the deal that was supposed to fix it.”
What are the specific failure modes?
They recur in a short list: overpaying for early acquisitions, systems fragmentation, retained founder autonomy, cultural mismatch, insufficient management depth, and a pace of acquisition faster than the balance sheet or the organisation can absorb (CT Acquisitions, June 2026).
Overpaying early is the most expensive because it is compounding. The first two acquisitions set the internal reference price for everything after, and a platform that has paid seven times for its first add-on struggles to explain to the next seller why five is the right number. The discipline has to exist before there is any pressure to deploy.
Retained autonomy is the most common, because it is agreed for good reasons. Founders are asked to stay, they want to keep running their business their way, and the acquirer wants continuity. The result is a group of companies under one owner rather than one company, and a buyer at exit will price it that way.
The pace failure is the one that is fully visible in advance. Total debt for issuers below ten million dollars of EBITDA now clears at 2.50 to 3.25 times, down from 2.50 to 4.00 times in July 2025 (SPP Capital Partners, Market At A Glance, July 2026). A programme that plans six acquisitions on leverage capacity sufficient for three has already decided how it ends.
| Failure mode | What it looks like early | Diagnostic |
|---|---|---|
| Overpaying for early add-ons | The first two acquisitions clear well above the size-band average and set the internal reference price | Compare paid multiples across every acquisition in order, not on average |
| Systems fragmentation | Consolidated reporting is assembled in a spreadsheet each month from separate systems | Ask for a consolidated monthly profit and loss on one chart of accounts, produced in a normal close cycle |
| Retained founder autonomy | Acquired businesses keep their own pricing, sales method and brand indefinitely | Ask what changed operationally in each acquired company in the twelve months after closing |
| Pace beyond capacity | The next acquisition is signed while the previous one is still being absorbed | Test remaining leverage headroom and integration capacity against the stated acquisition plan |
What predicts the outcome before it happens?
Four diagnostics, all of which can be run after the second acquisition rather than after the sixth. Whether the acquired businesses report on one chart of accounts. Whether one person owns the integration plan and has capacity for the next one. Whether the acquisition price discipline has held from the first deal to the most recent. And whether the platform's remaining leverage headroom supports the plan as written.
The reporting test is the one that produces the fastest answer. If the platform cannot produce a consolidated monthly profit and loss on a single basis within a normal close cycle, the businesses have not been combined, whatever the organisation chart says. That is also precisely what a buyer's quality of earnings exercise will discover later, at a worse moment.
The customer test comes second. Combination benefits that depend on selling one company's product to another company's customers should be visible in the customer data within a few quarters. If they are asserted for three years and never appear in the numbers, they are not going to.
There is a market reason to run these now rather than later. Diligence is lengthening, with 73 percent of senior US investment bank executives expecting greater complexity over the next twelve to twenty-four months and 51 percent calling technology diligence the single most burdensome element of the review (SRS Acquiom and Mergermarket, published 23 February 2026). A platform whose systems were never combined is walking into the workstream that is hardest to survive.
What does this mean if you are selling into a roll-up?
That the acquirer's integration record is part of your diligence, particularly if any of your consideration is deferred or rolled. A seller taking cash at close carries none of this risk. A seller rolling equity into the platform carries all of it, and is underwriting a strategy rather than a business.
The questions are specific and answerable in one meeting. How many acquisitions has the platform completed, and are they on one accounting system. Who runs integration, and what are they working on now. What is the remaining headroom under the credit facility. What happened to the principals of the last two acquired businesses, and are they still there.
A platform with good answers is a genuinely strong counterparty and often the best available home for a smaller business, because the combination is real and the seller's proceeds compound inside something that works. A platform with poor answers is offering the same headline number with a different risk profile attached.
The asymmetry is worth stating plainly. Rolled equity in a well-integrated platform is the most attractive outcome in the lower middle market. Rolled equity in an unintegrated one is a minority position in a group of companies that a future buyer will price separately.
As of August 2026
Sources: Harvard Business Review research carried by DealRoom, updated June 2026, for more than two-thirds of roll-up strategies creating no value for investors; CT Acquisitions, Private Equity Roll-Up Strategy guide, published June 2026, for 30 to 40 percent of roll-ups underperforming initial expectations and for the failure-mode categories; Bain and Company data from 2024, cited by Acquisition Stars, March 2026, for 30 percent of acquisitions achieving underwritten combination benefits and 83 percent of practitioners citing integration execution, used as labelled historical context; BCG and HHL Leipzig data, cited by CapitalPad, June 2026, for the cohort with two or more add-ons earning the lowest returns despite the fastest revenue growth; PitchBook, Q2 2026 US PE Breakdown, 6 July 2026, as at 30 June 2026, for add-ons at roughly three-quarters of US sponsor buyouts; GF Data, published 27 January 2026, covering the first nine months of 2025, for size-band enterprise value to EBITDA averages, used as a labelled baseline; McKinsey and Company, Global Private Markets Report 2026, February 2026, for the record 11.8 times median buyout entry multiple in 2025; iCapital, February 2026, for entry-to-exit multiple spread compression from about four turns across 2010 to 2020 to about 1.6 turns from 2020; SPP Capital Partners, Market At A Glance, July 2026, for total leverage bands below ten million dollars of EBITDA and the year-on-year contraction; SRS Acquiom and Mergermarket, M&A due diligence study 2026, published 23 February 2026, surveying 150 senior US investment bank executives, for diligence complexity and for technology diligence being called the most burdensome element. The diagnostics and the rolled-equity questions are drawn from our own mandate practice. A companion article on this site covers what makes an acquisition cadence repeatable.