Difficulty is a moat, and the market has started paying for it

What is the boring business thesis?

That the things which make a business unpleasant to run are the same things that keep competitors out of it. Heavy equipment, regulatory licensing, filthy work, unsociable hours, difficult customers and working capital that never comes back all deter entrants. Deterred entrants are a moat, and moats are what an acquirer underwrites.

The thesis has always been available and has usually been unfashionable, because a business that is hard to run is also hard to describe in a way that sounds exciting. What changed in 2026 is that the pricing evidence caught up with the argument.

The clearest single datapoint is a reversal. Middle-market manufacturing multiples improved to 7.2x from 6.6x while business services eased to 7.0x from 7.3x, and in the smallest cohort of one to twenty-five million dollars of transaction value, manufacturing firmed to 5.8x while business services eased to 5.5x (Mercer Capital, Middle Market Transaction Update Summer 2026, published July 2026, on GF Data figures as of Q1 2026). Heavy overtook asset-light. That had not happened for years.

The wider version of the same move is the cohort PitchBook labels HALO, meaning heavy assets with low obsolescence: energy, materials, industrials and environmental services. It took 31.2% of US private equity deal value in Q1 2026 against a steady share of roughly 14% across 2016 to 2024 (PitchBook via Blue River Financial Group, Blue River Brief June 2026, published July 2026). Capital did not drift toward these businesses. It relocated.

Why did capital rotate toward these businesses now?

Because the credit market reassessed the alternative. Software's share of broadly syndicated loan issuance fell to 8.6% year to date in 2026 from 17.6% in 2025, its lowest since 2013, and Q2 software private equity deal value fell to $10.7 billion, down 65.7% year over year, cutting software to roughly 6% of private equity deal value from 13.3% in Q1 (PitchBook LCD as of 30 June 2026 on the loan share, and PitchBook data via the Blue River Brief on deal value).

The read from the analyst who published the loan figures is that this is a structural credit story rather than a cyclical one (Sikich, Q2 2026 Credit Market Update, 13 July 2026). Lenders are not waiting out a bad quarter in software. They are re-rating a category whose defensibility they no longer trust, and the money has to go somewhere.

Where it went is toward assets whose replacement cost is visible and whose obsolescence risk is low. A permitted site, a certified process, a fleet, a furnace and a route map are all things a competitor would have to spend years and real money to duplicate. That is a different kind of durability from a contract that renews annually because switching is inconvenient.

One correction is worth carrying, because the wrong figure has circulated widely. The software loan share is 8.6% year to date, not 8.8%; the higher number propagated through several secondary write-ups before being reconciled against the data owner. The direction is not in dispute and the magnitude barely changes, but a figure quoted in a diligence memo should be the right one.

What does the premium actually look like?

It looks like more bidders, higher clearing multiples and rising deal sizes in exactly the sub-sectors nobody wanted to talk about three years ago. Precision manufacturing deal volume rose 19.6% year over year to sixty-one transactions in Q1 2026, and the average sector multiple ticked up more than half a turn to 10.1x EV/EBITDA across 2023 to Q1 2026 from 9.6x across 2020 to 2022 (Capstone Partners, Precision Manufacturing Market Update, 6 May 2026).

The public market is making the same statement. Capstone's precision manufacturing index traded at 17.1x EV to trailing EBITDA as of 31 March 2026, above the Dow Jones Industrial Average at 16.0x (same source). A premium to the broad index is the market pricing a sector as structurally advantaged rather than late-cycle, and it is the sort of thing that used to be reserved for software.

Building products produced the strongest combined move in the sector: year-to-date 2026 volume rebounded 28.2% to 182 transactions, multiples averaged 10.6x EV/EBITDA across 2025 to year-to-date 2026 against 9.4x in 2024, and median deal size jumped to $411.6 million in 2025 from $185 million in 2024 (Capstone Partners, Building Products Market Update, 30 June 2026). Volume and multiple moved together, which is unusual and is what a genuine re-rating looks like.

Environmental services is the cleanest expression of all of it. June 2026 printed thirty-eight US transactions, up 23% year over year, against a decline of roughly 17% in overall US deal count, with listed peers at a median 16.0x trailing EV/EBITDA as of 2 July 2026 (PitchBook data via Blue River Brief June 2026). Waste, remediation and compliance work is about as unglamorous as the middle market gets, and it is beating the market by roughly forty points.

Does this apply to every unglamorous business?

No, and the counter-example sits inside the same industrial sector. Packaging recorded fifty-nine transactions year to date in 2026, a drop of 21.3% year over year, explicitly against a 12.1% rise in the broader industrials industry, with private strategic deals down 29.7% and sponsor-owned deals down 26.6% (Capstone Partners, Packaging Market Update, 27 July 2026). That is a thirty-three point spread against the sector average.

The reason matters more than the number. Packaging is heavy, capital intensive and operationally difficult, and it fails the test anyway because its demand is consumer-exposed rather than structural. Difficulty is not sufficient on its own. It has to sit on top of demand that does not evaporate.

The reshoring theme carries a similar warning. US manufacturing construction spending peaked at $239 billion in June 2024 and has since declined 21%, and excluding electronics, spending rose only 5.6% since tariffs began, which the analyst describes as lower than would be expected for a boom (IoT Analytics, Industrial Macro Pulse May 2026, published 12 May 2026). If a buyer's thesis on a business is a reshoring tailwind, the macro data does not yet support it. The specific customer programme might; the theme does not.

There is also a scale effect that no amount of operational difficulty overcomes. Platform businesses transacted at 7.6x against 6.5x for add-ons, the widest spread in the series, and size bands ran roughly 9x to 11x for enterprise values of one hundred to five hundred million dollars against roughly 7x for ten to fifty million (Mercer Capital on GF Data, as of Q1 2026). A hard-to-run business at the small end is still a small business.

What does a buyer underwrite in a hard business?

Three things, in this order: whether the difficulty is genuinely a barrier or just an inefficiency, what the working capital and capex actually consume, and whether the operating knowledge sits in a system or in the owner's head. The third is where most processes fail.

The first is a diligence exercise with a clear answer. A permit, a certification, a specification approval, a bonding line or a customer qualification is a barrier because a competitor cannot buy one quickly. Low margin caused by poor pricing is not a barrier, it is an unfixed problem, and the buyer will treat it as an improvement opportunity for which it is not paying you.

The second is where the price gets negotiated in practice rather than in principle. Businesses of this kind consume cash on the way up, and a buyer models maintenance capex and the working capital swing through a full cycle, not through your best year. Bring thirty-six months of clean monthly financial statements and a defensible normalized working capital picture; the prevailing standard for this market is at least that, with a quality of earnings report commissioned early rather than in reaction to a buyer's findings (Capstone Partners, Capital Markets Update, 4 June 2026).

The third is the one an owner can fix before going to market and almost never does. If the reason the plant runs is that you know which machine drifts and which customer will accept a substitution, then the moat is you, and a buyer prices that as key-man risk rather than as a barrier to entry. Written procedures, a second layer of management and a documented maintenance history convert personal knowledge into transferable knowledge, which is the difference between an interesting business and a sellable one.

One thing this piece deliberately does not do is put a multiple on scrap and recycling specifically, which was the example that prompted it. No publisher issues a current middle-market transaction multiple series for that sector, and the nearest honest proxies are the environmental services and precision manufacturing figures above. Quoting a number that does not exist would undo the point of the article.