By Greg Allan, tradesman and creator of Data First Coaching
Why can a quote with a healthy margin still lose money on the job?
Because a fabricator prices in pieces, a day of welding here, a sheet of plate there, a run of powder coat or galvanising on top, and every piece can look fine on its own while the job as a whole still comes in flat. The quote is built line by line. The result only ever shows up blended together, once the job is finished and the real hours and real material are known.
Most fabrication shops quote structural steel, balustrades and gates the same way: estimate the welding hours, estimate the steel and the offcut, add the subcontracted finishing, and total it up. Each line can carry a sensible margin on paper. But if the welding ran long, the sheet yielded less than expected, or the galvanising invoice came in higher than quoted, that loss does not show up as one obvious mistake. It gets absorbed quietly across the whole job, and the final margin looks worse than every individual line suggested it would.
Why does pricing around how the owner welds and cuts cause the gap?
Because a price built around the owner's own speed in the workshop is a price built on the fastest, most experienced hands in the business, and almost nobody else on the job runs at that pace. When a newer welder or a subbie picks up the work, the hours blow out against an allowance that was never realistic for anyone but the owner.
This is one of the most common ways margin disappears on structural steel and balustrade work. The quote looks fine because the owner mentally timed it against their own welding and cutting speed. Once the job is handed to the team, the material and time allowances are too lean, and that shortfall comes straight off the job's profit, job after job, until the price list is rebuilt around how the business actually runs, not how the owner runs it.
What number actually tells a fabricator whether a job made money?
The number that tells the truth is the gross profit on the finished job, measured against what was actually quoted for labour, steel, offcut and outsourced finishing, not the margin that was assumed at quoting time. This is the number most fabrication shops never look at closely, because it only exists once the job is done and everyone has moved on to the next one.
Steel, aluminium and stainless prices move. Welder rates rise. Powder coat and galvanising turnaround and cost change with demand. A price book built once and left alone quietly bleeds margin on every job priced from it, because the cost side of the equation keeps moving while the price list stands still. The only way to know if a job held its margin is to check the actual numbers against the quote, line by line, after the fact, not to trust that a healthy quote stayed healthy.
How can AI agents help a fabrication workshop price more accurately?
AI agents can take over the non-billable work of pulling actual job costs together against what was quoted, so the owner sees where a job's real margin landed without spending an evening on it themselves. An AI agent is simply software that does a job the way the owner would do it, using the owner's own prices, wording and rules, not a generic template.
For a fabrication workshop, that means an agent built around the shop's own price book can flag when a job's welding hours, offcut rate or galvanising invoice have drifted from what was quoted, before the next similar job goes out the door at the same price. This is data first thinking: starting from what the shop's own jobs actually cost, rather than from memory of how the last job seemed to go. AI agents used this way do not replace the estimating, they remove the guesswork sitting underneath it.
Should a fabricator use one fixed price or a price window on custom work?
For genuinely custom fabrication, where every job differs in size, quantity or complexity, a single fixed price often forces the owner to wear costs that were never theirs to absorb. A price window, built from the shop's own cost data rather than a guess, can protect margin on the jobs that do not fit a standard per-metre or per-item rate.
Picture a fabricator doing custom balustrades and small structural packages, where no two jobs were ever quite the same size or complexity. Pricing per metre or per item never worked, because the next job always had something different in it, an awkward site install, an extra run of welding, a finish that needed more prep. The owner kept wearing those extra costs as if they were mistakes on his part, when they were really just the nature of custom work. Once he stopped giving a single fixed price and started quoting a price window on each job, the custom work stopped quietly eating into the margin that the standard jobs were earning.
How does Data First Coaching approach this?
Data First Coaching starts with the fabrication shop's own job data, through the Trade Business Intelligence, to see where the gap between gross and net actually begins, not where it shows up. That might be a welding rate priced around the owner's own speed, a price book that has not moved with steel and galvanising costs, or one weak line pulling down a job that looked fine everywhere else. Greg Allan built this from running his own trade business, and teaches it as: “one of the biggest mistakes we see people make is they price it around themselves doing it.” The how is worked through inside the Academy, with members, because the Trade Business Intelligence is powered by every member who is open minded, innovative and willing to share what they find.
