Steel Fabrication Man Hours per Ton
Most structural steel shops land between 15 and 30 man hours per ton for typical building work — closer to 15 for repetitive bolted framing, closer to 30 when moment connections and varied members show up. Miscellaneous metals are a different animal: stairs, rails, and ladders routinely run 60–120+ hours per ton. If someone quotes you a single "industry number," ask what kind of steel they're talking about, because the spread between a bar joist job and a monumental stair is nearly 10x.
Here are the ranges that hold up in practice:
| Work type | Man hours per ton | Notes |
|---|---|---|
| Simple bolted framing (repetitive beams/columns, shear tabs) | 12 – 20 | Long, heavy, repetitive members |
| Moderate structural (mixed members, some moment connections) | 20 – 30 | The most common bucket for building work |
| Complex structural (heavy welding, built-ups, trusses) | 30 – 45 | CJP welds, stiffeners, tight tolerances |
| AESS (architecturally exposed) | 40 – 60+ | Ground welds, appearance criteria |
| Stairs and railings | 60 – 120+ | Light pieces, heavy layout and fit time |
| Ladders, gates, misc metals | 80 – 150+ | Almost all labor, almost no tonnage |
These are shop hours only — fitting, welding, handling, and loading. Detailing, erection, and field work are separate line items (covered in our 2026 fabrication cost breakdown).
What Counts in the Man Hours per Ton Number?
Before you compare your shop's number to anyone else's, agree on what's inside it. The honest version includes every direct shop hour the job consumes:
- Unloading and material handling — receiving, staging, moving stock to the saw and pieces between stations. On heavy work this is 10–15% of total hours; shops with poor crane coverage or bad layout burn far more.
- Cutting and prep — sawing, coping, beveling, drilling or punching.
- Fitting — layout, tacking clips and plates, squaring. Usually the single biggest bucket, often 35–45% of shop hours.
- Welding — actual arc time plus setup and repositioning. Arc-on time is typically only 20–30% of a welder's day; the rest is handling and prep, which is why welding hours don't drop just because you bought a faster machine.
- Cleaning and painting — blast, grind spatter, shop primer. Add 1–3 hours/ton for standard primer, more for special coatings.
- Loading out — sorting by sequence, bundling, securing trucks.
What it should not include: detailing, project management, erection, or rework hours from drawing errors (track those, but separately — burying rework in the productivity number hides the real problem).
The dollar math is simple. At a $75/hr loaded rate, 25 hours/ton is $1,875/ton in shop labor — usually more than the steel itself. Material runs $1,100–$1,600/ton delivered for common shapes (current numbers in our structural steel cost per ton guide), which is exactly why labor hours, not mill price, is where bids are won and lost.
Why Do Stairs and Rails Blow Up Hours per Ton?
Because hours track piece count and complexity, and tonnage tracks weight — and stairs have lots of the first and almost none of the second.
A W24x76 beam, 30 feet long, weighs 2,280 lbs. One saw cut each end, four holes per connection, maybe two shear tabs. Call it 3–4 shop hours for over a ton of steel: roughly 3 hours/ton.
Now take a single flight of steel stairs: two stringers, a dozen pans, closure plates, landing framing, posts, and 40 feet of handrail. Total weight might be 1,500 lbs — but you'll spend 50+ hours on layout, fitting treads square and level, welding out both sides, and grinding everything smooth because people put their hands on it. That's 65+ hours/ton on a good day.
Nothing went wrong in that comparison. It's the same physics on every misc-metals job: small pieces mean handling time per piece stays fixed while pounds per piece collapses. This is why estimating stairs by tonnage-based rules of thumb is the fastest way to lose money in this business — you have to price the pieces and operations, not the pounds. Our miscellaneous metals estimating guide walks through pricing this work by assembly instead.
How Do Connections per Ton Drive Shop Hours?
Two jobs with identical tonnage can be 40% apart in shop hours, and the difference is almost always connections.
A useful way to think about it: every connection is a fixed chunk of fit and weld time, regardless of how heavy the members are. So the real driver is connections per ton:
| Job profile | Connections per ton (approx.) | Effect on hrs/ton |
|---|---|---|
| Heavy industrial, long spans | 1 – 2 | Low — 12–18 hrs/ton |
| Typical office/commercial framing | 2 – 4 | Middle of range — 20–28 hrs/ton |
| Light framing, short infill beams | 4 – 8 | High — 28–40 hrs/ton |
| Moment-frame or braced-frame heavy | 2 – 4 but complex | High — the type costs you, not the count |
Connection type matters as much as count. A standard shear tab is 30–45 minutes of fit and weld. A moment connection with CJP flange welds, stiffeners, and UT inspection can be 8–12 hours by itself. Run the connection schedule before you commit to an hours-per-ton figure — the per-connection dollar ranges in how much do steel connections cost map directly onto shop hours.
Quick tell when reviewing a bid: if the drawings show short spans and lots of infill framing, average member weight drops, connections per ton climbs, and your hrs/ton needs to climb with it. A job full of 8-foot W8s behaves nothing like a job full of 30-foot W24s.
How Do You Sanity-Check an Estimate with Hours per Ton in 5 Minutes?
This is the benchmark's real job. It's too blunt to build an estimate with, but it's excellent for catching an estimate that's gone off the rails. The five-minute check:
- Pull total shop hours and total tons from the estimate. Divide.
- Bucket the job using the table at the top. Typical commercial framing with a few moment frames? Expect 20–30.
- Adjust for the three big movers: average member weight (light average weight pushes the number up), connection type (moment/braced work pushes it up), and finish (AESS or heavy coating prep pushes it up).
- Compare. Inside the adjusted range: move on. Outside it: find out why before the bid goes out.
Worked example: 85-ton office building, mostly bolted framing, six moment frames at the core. Estimate shows 1,690 shop hours = 19.9 hrs/ton. Expected range for this profile is 20–28. You're at the very bottom of the range with moment frames in the mix — that's a flag. Dig in, and you find the estimator carried the moment connections at shear-tab hours. Fixing it adds ~180 hours, roughly $13,500 at $75/hr. That's the check paying for itself on one job.
It works in both directions. An estimate at 38 hrs/ton on simple framing means either something in the scope justifies it (galvanizing prep, small pieces, brutal schedule) or you're about to bid yourself out of the running.
One prerequisite: the tonnage has to be right. Hours per ton divided into a takeoff that missed 15% of the members tells you nothing — garbage tons in, garbage benchmark out.
How Should You Track Your Own Shop's Number Job by Job?
Published ranges get you in the neighborhood. Your own trailing number wins bids, because your 22 hrs/ton on school gymnasiums is a fact, not an industry average.
Minimum viable tracking, no software required:
- Log actual shop hours per job from time cards, coded to the job. Not estimated hours — actuals.
- Log final fabricated tonnage from the shipping record, not the bid tonnage.
- Tag each job with a profile: work type, dominant connection type, average piece weight, finish. Four fields.
- Compute hrs/ton at closeout and keep a running list. After 10–15 jobs you'll have real curves by profile.
Two habits make the data worth having. First, keep rework hours in their own bucket — a job that ran 31 hrs/ton because the drawings changed three times shouldn't drag your framing benchmark up. Second, review the estimate-vs-actual gap on every closeout, not just the losers. A shop that consistently beats its estimate by 20% isn't efficient; it's leaving money on the table on every winning bid.
Shops that do this for a year end up with the most valuable estimating asset there is: a defensible, shop-specific hours curve that no competitor can copy.
FAQ
Is man hours per ton lower on heavy steel?
Yes, substantially. Handling and connection time is roughly fixed per piece, so heavier average member weight spreads those hours over more tons. A job averaging 1,200 lbs/piece might run 15 hrs/ton where the same shop runs 35 hrs/ton on work averaging 300 lbs/piece. Always check average piece weight before applying a benchmark.
Does automation change the number much?
Where it's applied, yes. A beam line with automated drilling and coping can cut prep hours 30–50% on repetitive framing, and shops running them post 10–14 hrs/ton on work that takes a manual shop 18–22. But automation helps least where hours are worst — stairs, rails, and one-off misc work are still mostly skilled hand fitting. Buy the beam line for framing throughput, not to fix your misc-metals margins.
Where do the benchmark numbers come from?
Mostly fabricator cost history — there's no official published standard. AISC surveys, estimating references, and shop consultants circulate similar ranges to the table above, but every credible source says the same thing: use published figures to bracket a bid, use your own job-costed history to price it.
Should I quote customers a man hours per ton rate?
No. It's an internal benchmark, not a pricing unit. Quote lump sum or unit prices built from your actual operations. The moment a customer knows your hrs/ton and loaded rate, every negotiation becomes an argument about your productivity.
Every hours-per-ton check starts with the same input: an accurate ton count. If the takeoff missed members, every ratio downstream of it is wrong. SteelFlo reads your PDF drawings and finds every steel member across 4,500+ sections and 6 international standards, with 95-99% accuracy when paired with your review — so the tonnage you're dividing by is the tonnage that's actually on the job. Get tonnage right first: your first 3 takeoffs are free.