Tools / 02
Ohio's
One-Person
Problem
A commercial trade shop in Ohio cannot pull a permit without a licensed person on the paperwork, and the state publishes who those people are. I read all 8,965 active licenses, joined them against a year of Columbus permits and five years of permits across three cities, and asked two questions. How many of these companies are one resignation away from being unable to work? How many of the small ones will still exist in three years?
Ohio license roster pulled 2026-07-14. Columbus permits Jan 1 to Jul 9, 2026. Five-year charts cover 162,058 permits across three cities, 2022 through 2026.
of the companies on Ohio's commercial trade license roster run on exactly one licensed person. That is 5,430 of 5,988 companies. If that person quits, retires, or gets hurt, the company cannot legally pull a permit the next morning.
I count humans, not credentials. Count license cards instead and only 75.7% of firms look solo, because in 896 of these one-person companies that person holds licenses in two or more trades. The same human is legally the electrical department and the HVAC department. One resignation closes both.
Who holds the licenses, county by county
Every active commercial trade license in Ohio, placed in the county where its holder is on record. Darker green means more. Tap a county, or filter to one trade.
Cuyahoga carries the most at 908. Thirty-nine counties have fewer than thirty licensed people in them, total, for every commercial job in the county. Paulding has four.
How deep the bench goes
Licensed people per company, across all 5,988 companies on the roster.
The deepest bench in the state is thirteen licensed people. One company has that. Another has twelve. Below them it falls off a cliff, and 5,430 companies sit on the first bar. Only 449 have even a second licensed person to fall back on.
Will the shop still be here in three years?
Every firm that pulled permits in 2022 across Columbus, Cleveland, and Cincinnati, followed through 2025. The bar is the share still in the permit record three years later, by how much work the firm was doing. Whiskers are 95% confidence intervals.
A shop doing 2 to 4 permits a year had a 17.1% chance of vanishing from the record within three years. A shop doing 25 or more had a 3.1% chance. The confidence intervals do not touch. In this market, being small is the risk factor.
This counts operating firms, meaning at least two active months in 2022. The rawest bucket, firms that pulled exactly one permit ever, lost 48.2% of its members, but that group is full of homeowners and one-off LLCs that were never really businesses, so it stays out of the chart. And leaving the permit record is not the same as going out of business. A firm can move to the suburbs, go residential-only, get acquired, or start subbing under a GC's permit. I can see the leaving. I cannot see the reason.
The market everyone says is booming
162,058 MEP permits across five years of Columbus, Cleveland, and Cincinnati city rolls, compared January through June so the partial 2026 year plays fair. Each line is a city, indexed to its own 2022.
Pooled, the three cities are at -3.8% versus 2022. There is no boom in this data. One growing market, Cincinnati. One flat one, Columbus. One shrinking one, Cleveland, down twelve points in four years. A shop reading national headlines about a construction supercycle is reading about somebody else's town.
City jurisdictions only, so work that moved to the suburbs is invisible here. Cincinnati's feed carries no electrical permits, which is a gap in the source, not a zero. Within each city the comparison holds across all five years.
Private equity is supposed to be consolidating the trades. Not in Ohio's permit record. I measured market concentration for every city and trade combination across all four complete years, and the highest reading anywhere is 425, against a consolidation threshold of 1,500. Concentration rose in 2 of 8 city-trade cells and fell in the rest. The independents still own this market, and the fragility on this page is theirs to fix, not a buyer's to absorb.
When the work actually happens
Each month as a share of its own year's average, averaged across the four complete years, 2022 through 2025. Green is the May-to-August season.
The warm season runs 1.16x the winter floor, and it held in 4 of 4 years. Winter is when the pipeline is thinnest and the time to line up next season's crew, because by June everyone is hiring at once.
The limit of this chart is the single peak month. August topped three of the four years and July topped the other. The half-year pattern holds. Which month wins next year is not something this data can tell you.
Columbus, the year so far
Every Columbus firm that pulled a permit this year and also appears on the state license roster. Across the bottom, how much work it did. Up the side, how many licensed people it has to carry that work. 649 firms, 12,224 permits.
Each dot is one firm. The horizontal scale is square-rooted so the crowd near zero stays readable. The green band is where a single licensed person is carrying a real commercial workload.
What this means, plainly
Your license holder is not an employee. He is the permit. Ohio will not let you pull commercial work without one, and if he walks on a Friday you are not bidding on Monday.
The fix takes months, not weeks. A second licensed person has to pass a state exam and carry the years of experience to sit for it. There is no hiring your way out of it in a hurry.
You are worth more than your paycheck says. Of Ohio's 5,988 licensed companies, 5,430 have one person in your seat, and 14 of the 38 strained shops I checked were advertising for a replacement bench the same week.
Nobody puts that leverage on a job posting. It sits in the state roster, and the roster is public.
The shops in the green band
Anonymized on purpose. These are real firms from the Columbus permit rolls, and I am not going to publish a list of small businesses with their weak point circled.
40 more shops sit in the same band. They are named in the private index, not here. Ask me about a specific one.
The heaviest cards at the top are the ones I trust least. Three firms in this band cleared two hundred permits on a single licensed name, and a shop that size is more likely licensing under a second entity than actually running that thin. Those three are flagged for a manual check. The typical shop here is quieter and harder to explain away, twenty to thirty-five permits carried by one person, and nineteen of the 46 look like that.
I hand-checked 38 strained single-license shops for a live job posting on the day the roster was pulled. 14 were actively recruiting, one of them offering a $5,000 signing bonus for an experienced plumber. These owners know exactly how exposed they are. They are bidding for the fix in public.
Wondering about your own shop
What I'm building next
The question worth answering is whether public signals can tell you a trades firm is about to hire, before the job gets posted. Nobody can answer it today, and that is a data problem, not a modeling one. No historical roster exists anywhere. The Wayback Machine has nothing, the state portal serves only today's snapshot, and the roster's one date column is a renewal date. The time axis has to be started, not recovered.
So I started it. Snapshot one is 2026-07-14, snapshots run weekly from here, and the first real license-change events land 2026-07-21. Before seeing a single event, I wrote down three hypotheses and their due dates. They stand published so anyone can check them later.
Firms carrying the most permits per licensed head today will add a licensed person within 12 months at a higher rate than firms carrying the least.
testable Q3 2027
Growing permit volume predicts a hire in the next 6 months better than high volume alone.
testable Q1 2027
The most permit-loaded Columbus firms right now will add a licensed person within 90 days at an elevated rate.
first read mid-Oct 2026, expect wide error bars
The rules are pre-registered too. Each test runs once, on its due date, with 95% confidence intervals, no peeking week to week and no shortening a window because I got impatient. If a hypothesis dies, it gets published dead. Bookmark this section and hold me to it.
What's in the index
This page is a window cut into a bigger system. Here is what the system holds, counted honestly.
Ohio license roster pulled 2026-07-14. Columbus permits 2026-01-01 to 2026-07-09. The roster is a snapshot, not a feed, and it is dated on purpose.
Method, honestly
Source is Ohio's own commercial license roster, 10,409 rows, of which 8,965 are active. The rest are mid-renewal and left out. A license holder means a distinct named person, so someone carrying both an electrical and a plumbing license counts once as a person and twice as a license. Companies are matched on name after stripping the legal suffix, which is imperfect. A shop that licenses one entity and permits under another can read as thinner than it is, and I would rather say that than pretend the join is clean.
109 licenses have no usable county on the roster and sit out of the map. This-year permits are the City of Columbus electrical, mechanical, and plumbing rolls for 2026-01-01 through 2026-07-09, so a firm working the suburbs looks quieter here than it is. The five-year charts add the Cleveland and Cincinnati city rolls, 2022 through 2026, with year-over-year comparisons held to January through June because 2026 is a partial year. Every proportion on this page carries a Wilson 95% confidence interval, and every five-year figure is recomputed from the raw city files at build time and checked against an independent analysis run. If they disagree, this page refuses to build.
Permit-filing concentration is not license fragility. 546 Columbus firms filed every permit this year under one person's name, which looks like fragility and mostly is not. 55 of them carry two or more licensed people, and one filer is just the office habit of whoever does the paperwork. 276 more never needed a commercial license at all. Only 215 firms survive the join as genuinely single-licensed, and 46 of those carry twenty or more permits. Every strain figure on this page comes from the join, never from the filer name.
The roster's effective-date column is a renewal date, not a licensure date. 98% of the dates on the whole roster fall inside 2023 to 2026, and a firm's holders share dates because they renew together on the company's cycle. Model hiring on that column and you will get a clean result that measures the renewal calendar. Nothing on this page treats it as a hiring signal.
No license holder's name, phone, or email appears in this page's data, and no company is named. That is a line, not an oversight.
Run a trade shop in Ohio and want to know what the state roster says about your bench? Send me the company name. I will read it back to you straight, and if you are three deep and fine, I will tell you that.