Metal Fabrication ERP: Alberta Job Shops

Quick answer: fabrication shops rarely lose money on the jobs they think they lose money on. The leak is almost always drop, unrecorded weld time, or rework absorbed into the original job, and none of those are visible from an accounting package plus spreadsheets. Fixing that is an operations problem before it is a software problem, and it has a limit worth knowing before you spend anything.

Why Alberta, and Why Here

Nisku alone is the largest manufacturing industrial park in Canada, roughly 9,345 acres, home to more than 400 businesses and over 6,000 workers, with module assembly work feeding oilsands extraction, upgraders, refineries, and petrochemical plants. Add Acheson and Leduc and the Edmonton region holds one of the densest concentrations of fabrication, module, and pipe shops on the continent.

Most of those shops are privately held, run somewhere between fifteen and a hundred and fifty people, and quote work from a drawing. This page is written for them.

How the Work Actually Runs

Understanding the software question requires being honest about the operation, because fabrication breaks the assumptions most business systems are built on.

Every job is one-off until proven otherwise. Quoting happens from a drawing or a sketch, and the bill of materials is frequently incomplete until the job is on the floor. A system that demands a complete BOM before work can start will be worked around by week two.

Quotes are not built from parts lists. They are built from weight, cut length, weld inches, and finishing. An estimator carries the logic in their head and a spreadsheet. Any software conversation that does not start by reverse-engineering three real quotes is guessing.

Material comes from two directions. Large sections get bought to the job. Common sizes get drawn from stock. Both paths have to exist, because the shop will use whichever is faster that day.

Work leaves the building mid-job. Galvanising, plating, heat treat, and machining all go outside and come back. The material stays your asset the entire time.

Rework is routine. Treating it as exceptional is how job costs end up systematically understated.

Where the Margin Actually Leaks

This is the part worth reading twice. Ranked by how often each turns out to be the answer when a shop says it is busy and not making money.

	The leak	Why it hides
1	Drop and offcut	The job gets charged for what was theoretically needed instead of what was consumed. Nesting is not modelled, so the difference disappears
2	Weld time	Under-recorded, because welders are welding instead of clocking
3	Rework	Absorbed into the original job with no separate visibility, so the job looks like it cost what it should have
4	Outside processing	Cost lands after the job closed, or on the wrong job entirely
5	Material price movement	Between quote and purchase, on work quoted weeks ahead
6	Consumables, gas, grinding discs	Expensed to overhead and never attributed to anything

Notice what these have in common. Every one is a timing or attribution failure, and not a pricing failure. Shops respond to thin margins by sharpening the quote, which does nothing, because the quote was roughly right and the cost was never measured.

That is also why disconnected tools cause this leak instead of merely failing to fix it. When the estimate lives in a spreadsheet, purchasing lives in email, the shop floor lives on paper travellers, and the invoice lives in an accounting package, there is no point at which the four meet. Nobody is being careless. The information physically has nowhere to reconcile. Our spreadsheets vs an operations platform comparison covers the three points at which that arrangement stops working.

The Honest Limit, Stated Before the Estimate

Job costing is only ever as good as what the floor records.

A shop where welders and fitters record time against an operation gets operation-level costing and can see that the fitting is fine and the welding is over. A shop where the crew records against a job number at the end of the day gets job-level costing and can see only that the job lost money.

Both are legitimate. What is illegitimate is selling the first and delivering the second. If anyone quotes you a system on the promise of accurate job costing without first asking who records weld time and when in their day they record it, they have not thought about your shop.

Decide which shop you are before you sign anything. That decision drives the scope more than any feature does.

What an ERP Genuinely Does Here

Operational reality	The capability	Verified position
Galvanising, heat treat, plating	Subcontracting	Odoo documents subcontracting as engaging a third-party manufacturer within the production flow, with the material staying yours at the subcontractor
Drop and expected yield loss	Scrap and yield on the bill of materials	Scrap is native, reachable through Inventory then Operations then Scrap
Heat numbers surviving to the finished job	Lot tracking on raw material	Native, and it changes how receiving works, so confirm the requirement before designing it
Discovered scope	Project structure instead of a manufacturing order	Map job to manufacturing order where a BOM exists, and to a project where the work is discovered as it proceeds. Decide per shop
Historical job cost for estimating	Job cost history queryable by similar work	This is what makes the estimator faster, and it is the most underrated benefit
Nesting	None	Searching the whole Odoo 19.3 documentation set for nesting returns zero results. Nesting stays external and feeds quantities in

That last row matters commercially. If a vendor implies their platform nests, ask them to run it on your sheet sizes and your part mix in front of you.

Material Certification, and What It Does to Receiving

Shops working to code carry a documentation obligation that shapes the whole system.

A mill test report or material test certificate documents that material meets specification. Certificate types under EN 10204 are commonly summarised by suppliers and inspection bodies as 2.1, 3.1, and 3.2. A 3.1 is validated by the mill’s own authorised inspection representative. A 3.2 is additionally witnessed and countersigned by an independent third party such as SGS, TÜV, BV, DNV, or Lloyds. The operational trap is that a 3.1 is not acceptable where a 3.2 was specified, so the requirement has to be caught at order time instead of at delivery.

Confirm certificate requirements against the standard itself and your contract, because the summaries above are the industry’s shorthand instead of the text of the standard.

Systemically, the consequence is one sentence long. Heat numbers have to survive from receiving through to the finished job. That is lot tracking, and it changes receiving from a quantity check into a quantity and identity check. Shops that skip this discover it during a client audit, which is the expensive way.

Certification of the shop itself is a related question. The CWB Group publishes a searchable directory of companies certified to CSA standards including W47.1, and ASME publishes a certificate holder search with scope and expiry. If you hold either, welder qualification and procedure records are already an obligation, so the system should hold them instead of a filing cabinet. The same expiry-tracking discipline shows up in our procurement readiness checklist, because the buyers asking for your certs are usually the ones running prequalification.

The Five Questions Worth Answering Before Any Demo

These are worth asking yourself, and they are the same ones a competent implementer should ask you.

1. Show me the last three quotes you sent and how you built each number.
2. When a job needs galvanising, what happens to it in your system today?
3. How do you know what a finished job actually cost, and when do you know it?
4. Who records weld time, and when in their day do they record it?
5. Do your customers ask for material certs, and how do you produce one now?

Question four decides your scope. Question three is usually the one that produces an uncomfortable silence.

What to Do First

Take one finished job from last quarter and rebuild its true cost by hand. Material actually consumed including drop, all labour including the rework nobody logged, outside processing, and consumables. Compare that to what you quoted and what you invoiced.

Do it for three jobs and the pattern will be obvious. Most shops find the leak sits in one or two of the six rows above, consistently, and that finding is worth more than any software shortlist because it tells you what the system has to capture.

From there, our Edmonton software planning guide covers sequencing the decision, industrial operations covers the equipment and parts side, and operations software covers the categories. Shops serving producers should also read the northern Alberta oilfield services guide, because invoicing through operator portals adds a constraint this page does not cover.

For shops that conclude Odoo is the likely path, Calgary-based Solvync is one Alberta option for manufacturing configuration. Compare any partner against at least one alternative, and make shop floor recording, outside processing, and material traceability explicit line items in every written scope you receive.

Disclosure: Solvync may have a commercial relationship with the Biztech network operator. Solvync implements Odoo and does not implement nesting or estimating software, so treat that link as one vendor path among several.

Sources

• Odoo subcontracting documentation
• Odoo scrap documentation
• CWB Group certified companies directory
• ASME certificate holder search