Private equity should buy EHS leadership from someone who has been on the buy side, because the constraint isn’t safety knowledge — it’s the hold-period clock. A consultant delivers findings. An operator delivers a sequenced plan that fits a five-year model, survives three add-ons, and doesn’t need to be rebuilt before exit. Those are different jobs, and most EHS firms only know how to do the first one.
What does private equity actually get wrong about EHS?
The conventional treatment is that EHS is a diligence line item. You order the Phase I, you pull the OSHA logs and the loss runs, someone writes a red-flag memo, and the deal team files it. Post-close, if the numbers look ugly, you hire a consultant to fix “compliance.”
That sequence has a hidden assumption in it: that safety exposure is a stock of problems you can inspect and remediate. It isn’t. It’s a flow, produced continuously by how the business is designed and run. You can remediate every finding in a gap assessment and be back where you started in eighteen months, because nothing about the operating system changed.
Here is the counter-position in one sentence: safety performance in a PE portfolio is an output of operating design on a hold-period clock, and the person who fixes it has to think like a sponsor, not like a compliance vendor.
I’m not making that argument from the outside. I’ve been part of 19 acquisitions. I scaled my own holding company to $20 million. I worked inside a home services roll-up and I’ve served as fractional COO to small manufacturers, home services businesses, real estate operators, and agencies. I have sat on both sides — the one asking what the exposure does to the model, and the one who has to go make the number real.
Why doesn’t traditional EHS consulting work in a PE-backed company?
Three reasons, and they’re structural, not a matter of consultant quality.
1. Consultants price by deliverable. Sponsors buy by clock.
A consulting engagement is scoped to produce an artifact — an assessment, a program set, a training matrix. It ends when the artifact is delivered. But a sponsor doesn’t need an artifact. A sponsor needs a specific level of capability standing on its own feet by a specific date, because the date is set by the model, not by the safety department.
That reframes the entire engagement. The question isn’t “what’s wrong here.” It’s “what has to be true at month 18, what has to be true at month 36, and what is the shortest path that doesn’t collapse when we stop pushing.”
2. Consultants describe exposure. Operators quantify it.
A red-flag memo tells you there are open citations and a high experience mod. That’s description. What a deal team can use is the mechanism.
Take the experience modification rate. It’s not a safety score — it’s a financial instrument that reprices your workers’ comp premium. Rating uses a rolling multi-year window of payroll and loss data that lags the current policy period, and the formula weights the primary portion of each claim more heavily than the excess portion. The practical translation: claim frequency moves your mod harder than claim severity, and today’s improvement doesn’t show up in premium for years.
If you’re buying a business whose mod is going to keep climbing for two more rating periods on losses that already happened, that’s not a safety finding. That’s a modeled cost you should be pricing at the LOI, along with collateral requirements on the comp program, which move with loss history and can quietly consume working capital.
(Verify with your broker before publishing any specific split point, credibility, or state-fund variation — the mechanics differ by state and by rating bureau.)
3. Consultants build for one site. Roll-ups need replication.
This is where the operator background stops being a nice biographical detail and starts being the whole point. In a roll-up, you are not solving a safety problem nine times. You are building one system that can absorb the tenth acquisition in ninety days without a heroic effort.
That means standard work, not custom programs. One incident taxonomy so the portfolio data is additive rather than nine incompatible datasets. One set of leading indicators reported the same way at every site so the operating partner can compare Site 3 to Site 7 without a translation layer. A defined integration sequence with owners and dates that the GM of the newly acquired business can execute while also learning your ERP.
Nine excellent site-specific programs are worth less to a sponsor than one adequate system that replicates. Consultants are trained to produce the former.
What does the EHS Maturity Ladder look like across a hold period?
At FractionalEHS we use a five-stage ladder to diagnose where a business actually sits versus where leadership believes it sits. Mapped to a hold, it looks like this:
Reactive — safety happens after incidents. Records are incomplete, the loss runs tell a story nobody at the top has read. Most lower-middle-market targets are here at close, including ones that present well.
Compliant — written programs exist and match the operation. Training is current and documented. Citation exposure is managed. This is the floor, not the goal, and it is achievable in the first 6 to 12 months.
Managed — leading indicators are tracked and acted on, corrective actions close on schedule, and someone with authority reviews the data on a cadence. This is where insurance economics start to move. Target this by month 18 to 24 so the loss history has time to season before you go to market.
Integrated — safety lives inside the operating system. It’s on the tiered daily management board, in standard work, in capital review. No separate safety meeting, because it’s already in the meetings you were having. Realistic target by exit for a well-run platform.
Self-Correcting — the system finds and fixes its own drift without the corporate function pushing. Rare. Do not underwrite to it.
The diagnostic value is in the gap. I have never walked into an acquired business where the leadership team’s self-assessment and the ladder position agreed. They believed Managed; the log showed Reactive with good documentation.
Isn’t this what our broker’s loss control team already does for free?
This is the strongest objection, and it deserves a straight answer.
Carrier and broker loss control services are real and often good. If you aren’t using them, use them. Their engineers know your class codes, they see hundreds of comparable operations, and the service costs you nothing incremental.
But understand what it is. Loss control is a retention and underwriting service. Its output is a recommendation report to the insured, and its author has no decision rights inside your portfolio company, no seat in the operating review, and no accountability for whether anything actually changes. It identifies. It does not lead.
That distinction is the entire diagnosis I keep landing on: the Strategy–Execution Gap. Companies rarely have a safety knowledge problem. Almost everything they need to know is already documented, already recommended, already sitting in a report someone paid for. What they have is a translation problem between strategy and the floor — nobody with enough seniority to set priorities, allocate resources, and hold a GM accountable is actually working on it.
You do not close that gap with another report. You close it with a senior leader who can sit across from a plant manager who is pushing back, and who has the operating credibility to be taken seriously when they do.
The second objection — “we’ll just hire a corporate EHS director post-close” — is worth its own piece. Short version: you’ll spend four to six months recruiting, you’ll pay a market rate that assumes a full-time seat, and for the first year you’ll have one person doing setup work that doesn’t require them. The fractional model exists because the strategic work and the volume work are different jobs, and only one of them needs a senior person.
What should a deal team do differently on the next transaction?
Concrete, and startable this week:
Pull five years of loss runs, not the summary. Look at frequency by claim type and by site, not total incurred. Frequency tells you about the operating system; a single large claim may tell you about luck.
Ask for the OSHA 300 logs and the 300A summaries for the same period, and reconcile them against the loss runs. Where they disagree, you’ve learned something about the culture that no interview will surface.
In an asset deal, don’t assume open citations and abatement obligations disappear at close. Successor liability doctrines exist and their application varies. Ask counsel specifically, in writing.
Put a maturity target and a date in the 100-day plan, with a named owner who is not the plant manager. “Improve safety” is not a plan. “Compliant by month 9, Managed by month 20, owned by the COO” is.
And ask the person you’re about to hire for this work one question: what’s the sequence, and what happens at month 12 when the easy wins are gone? If they answer with a deliverable list, you’re buying a document.
