5 numbers 8 suppliers, 12 prices
The same part, created five times across five plants, bought from eight suppliers at twelve prices. Nobody is at fault: no two sites share a codification, and nothing compares parts across them.
Cold forming · Direct procurement
The part costs half a cent. The part numbers cost millions.
A bolt is not hard to price. Ten thousand bolts, opened as five references across five plants, bought from eight suppliers at twelve prices, is. Parsio matches parts on what they are rather than on what they are called, collapses the duplicates, and prices what is left. Unlock 3 to 15% untapped savings.
Scope
Screws & bolts Nuts & washers Ball studs Guide pins Rivets & studs Cold formed chassis parts Thread rolling Heat treatment Surface coating & plating Screws & bolts Nuts & washers Ball studs Guide pins Rivets & studs Cold formed chassis parts Thread rolling Heat treatment Surface coating & plating
Why cold forming is different
The part is simple. The portfolio is not.
On a welded assembly the cost sits in the process. On a cold formed part it does not, and treating the two the same is why this family gets skipped. Here the piece is trivial and the count is the problem.
Half a cent and a hundred thousand of them
The unit price was never the problem. The problem is the count: part numbers to quote, to keep alive in the ERP, to negotiate one at a time. Complexity here is arithmetic, not engineering.
B and C parts is where it has been sitting
Everyone works their A parts and their A suppliers. The tail is thousands of references worth a few cents each, and no team has ever been resourced to look at it. That is exactly why it is still there.
The die not the machine, sets the price
A cold formed part is a set of tools: stations, dies, punches, amortised over the volume that runs through them. Two identical parts at different volumes are genuinely different prices, so consolidating references changes the cost underneath.
Savings plays
Six ways the model pays for itself.
Find the part you already buy, five times over
Cluster on diameter, length, grade, thread, head and coating instead of on part number. The same bolt opened independently at five plants stops being five references with five prices.
Put the volume behind one die
When references merge, the tooling amortises over more parts. The saving is not only what you negotiate, it is what the die stops costing per piece.
Bring the B and C parts into scope
A finding worth two hundred euros is not worth a buyer's afternoon. Ten thousand of them are worth a quarter. Agents screen the whole tail in one pass and only surface what clears a threshold you set.
Price every part twice
A bottom-up should-cost from the actual route, next to a predicted price regressed across your own portfolio. The gap between them is the signal.
Challenge the callout that leaves the process
One tight tolerance or one finish can push a part off the header and onto a machine, at a different order of piece price. Show which ones do that, and what they cost across every reference carrying them.
Match the part to the shop
Headers by station count and wire capacity, what is heat treated in-house, what goes out for plating and how far. Two formers with the same catalogue are good at different parts.
How it works
From five codifications to one portfolio.
Read every file
Drawings, including the scans from before CAD. 3D files. ERP transactions from every site, price lists, supplier quotes. Parsio shows the fill rate per property before it shows a result.
Reconcile the codifications
Each plant invented its own article codes and its own descriptions, so one part lives under several numbers. Parsio matches on what the part is (diameter, length, grade, thread, head, coating) rather than on what it is called.
Apply your rules
Your engineers write them in plain language: "below M6 these two grades are interchangeable"; "zinc flake and zinc nickel are not, outside the engine bay". Versioned, so the rule outlives the engineer.
Compute both prices
Wire, heading stations, thread rolling, heat treatment, coating, amortised tooling. Should-cost against predicted price, on every reference. Where they diverge, something is going on.
Customer outcomes
Outcomes, not features.
Save costs 8%
Of direct spend recovered on average, across the families we run
Cut complexity 1 in 4
References found interchangeable within the buyer's own tolerance
Move faster 1 week
From scattered transactions and drawings to a matched, costed portfolio
Worked examples, all modelled
Screws & bolts · Automotive
The same part, five times, at twelve prices.
Five plants, five article codifications, and a bolt that was created independently in each of them. Eight suppliers ended up quoting it, at twelve prices, and no report in the group could see that because every report groups by part number. Parsio matches on diameter, length, grade, thread and coating, ignores the code, and the duplicates surface on their own. The negotiation that follows is about volume, not about price.
Fasteners
5 → 1part numbers behind one part
12prices for the same thing
Spec, not SKUas the unit of comparison
Illustrative. Modelled on portfolio structures mapped with commodity buyers and cost engineers. Not customer results.
Method built with teams in
Automotive & commercial vehiclesIndustrial manufacturingMachinery & machine toolsElectrical equipment
Named references available under NDA.
FAQ
Questions, answered.
We have a cost engineering team and a should-cost tool already. What do you add?
The scenario, not the number. If you already compute should-cost, you know what a part costs on the route it takes today. What is missing is what happens when the route changes, and on this family, what happens when two hundred references become forty. Parsio runs that across the whole portfolio rather than one part at a time.
Our fasteners cost a few cents each. Is there really anything there?
Not in the piece, and we would not pitch you on the piece. It is in the count. Thousands of references, each quoted, coded, stocked and negotiated separately, most of them duplicates of each other. The money is in collapsing that, and in the tooling that then amortises over more volume.
Each of our plants has its own article codes. Can you still match parts across them?
That is the job. We match on what the part is rather than on what it is called: diameter, length, grade, thread, head, coating, read off the drawing. Codes and descriptions are treated as unreliable, because in every group we have seen, they are.
Do we have to replace SAP, Ariba or our sourcing tool?
No. Parsio is the layer before you decide an event is worth running. It exports the costed pack, and the list of references that collapse into each other, into whatever you already use. The ontology, the rules and the cost model are yours.
We tried a statistical cost model once and engineering rejected it. Why is this different?
They were probably right to. A regression that says a part should cost less invites the answer "that is statistics, not the real cost", and the conversation ends there. What we build is a route: wire, stations, thread rolling, heat treatment, coating, each with an assumption you can open and change. It exists to start a conversation with engineering, not to win one.
Bring one fastener family.
Screws, studs, pins: a rough spend, the transactions from as many sites as you can get, and whatever drawings exist. We'll tell you within a week whether it's worth a first phase, and we'll say so if it isn't.