What quotation software for fabricators actually needs to do
Most software sold to manufacturers is built for standard products, and it fails on made-to-order work for one structural reason. Here is what to test on a vendor demo.
There is a confusion at the heart of this market that costs fabricators a lot of wasted evaluation time, and it is worth naming in the first paragraph.
Almost all quoting software assumes a catalogue. You have products; products have prices; quoting is choosing products, applying a discount and generating a document. That model is correct for a distributor, a dealer, a reseller — and it is a perfectly good business to build software for.
Made-to-order fabrication does not work like that. You do not have a price for a work table. You have a price for this work table, at 72 by 30 by 34, in 304, with an undershelf and a tray-type top, for this customer. Every job is a different size, so every price is computed rather than looked up.
A shop that trials catalogue software on fabrication work concludes, quite reasonably, that software does not help with this. The software was answering a different question.
The distinction that matters
Look at where the price comes from.
Catalogue tools hold a price against a product. Variation is handled with options that each add a fixed amount. It works well when the number of real variants is small and known.
Design tools hold geometry. You draw the job in 2D or 3D and the price falls out of what you drew. Excellent when the drawing has to happen anyway; expensive when it does not, because you have made drawing a prerequisite for quoting and most enquiries never become orders.
Formula-driven tools hold the arithmetic. Dimensions and a grade go in; your own rules about allowances, rates, gauges, labour and margin turn them into a price. No drawing, no catalogue of sizes.
For fabrication the third is the right shape, and it is the least common of the three.
Six things to test on any demo
Vendor demos are built to flatter the product. These are the questions that get past that, and each one has a bad answer worth listening for.
1. Change a dimension and watch what happens
Ask them to price a table at 72 by 30, then at 84 by 30, then at 84 by 36. Watch whether the price moves proportionately to material or in steps.
If it steps, the system is holding sizes as variants and somebody has to create every new size. That is a catalogue wearing a costume, and you will spend your life maintaining it.
2. Change the material grade
Same job, 202 instead of 304. Does the price move, and does it move by the right amount?
On a real job the answer is a big number — around a 30 per cent drop from 304 to 202 on a typical table — and it should come from a different steel actually being used in the calculation, not from a percentage applied to a total. Ask which. If the answer is vague, it is a percentage.
3. Ask where a number came from
Point at any line and ask why it is that number. You want to see the working: this much sheet, at this rate, plus this labour, plus this allowance.
If nobody can show you, then when a customer challenges a price, or when your own margin looks wrong, you have no way to find out. A pricing system you cannot audit is a pricing system you cannot trust, and eventually somebody starts keeping a spreadsheet alongside it — at which point you are paying for both.
4. Ask who can change a formula
The answer you want is: your estimator, in the interface, after training.
The answer you often get is: raise a ticket, and it will be done in the next release. That is a support contract dressed as a feature, and it means your pricing changes at the vendor’s speed rather than the steel market’s.
Ask specifically what happens when steel moves 8 per cent on a Tuesday.
5. Ask what the shop floor gets
The quotation already knows what the job consumes. If the system cannot produce a bill of materials and a cutting list from it, somebody in your shop is going to rebuild that by hand from the same information — and the two will disagree eventually.
This is also the honest test of whether the software really understands the job or is just generating documents.
6. Ask what happens to a quotation you have already sent
When a rate changes, does last month’s quotation change with it?
The right answer is no. A document you have sent should not silently rewrite itself; there should be a revision, with its own number and its own history, so that what you sent in March is still what you sent in March when the argument happens in July.
What you do not need
Worth saying, because feature lists are long and most of them are not for you.
A CAD package, unless you already draw everything. Drawing to quote is a large cost imposed on every enquiry, and most enquiries do not become orders.
A full ERP. If the pitch is inventory, production scheduling, purchase requisitions and a general ledger, you are being sold a two-year implementation to solve a two-week problem. Quoting is the bottleneck; start there.
A manufacturer catalogue, unless a lot of what you sell is equipment other people make. If you do both — resell some, fabricate the rest — you probably want both kinds of tool, and that is fine.
When a spreadsheet is still the right answer
If one person quotes, volume is low, and no two jobs share arithmetic, stay where you are. The overhead of any system is real and you would be paying it for nothing.
The threshold is not really volume. It is more than one person quoting, or the same products going out at different sizes repeatedly. The first means your pricing assumptions need to exist somewhere other than in a head. The second means you are retyping arithmetic that has not changed, and every retype is a chance to be wrong.
The one-line version
If the demo cannot show you a job repriced from its dimensions, repriced again from its grade, and explained line by line — it is a catalogue tool, and you make things to order.