That distinction matters in a customer bid. A sourcing team can use should-cost to test a bought-out component, but the bid still needs the supplier’s committed price, lead time, validity and terms. Treat the model as evidence for a decision. Do not let it quietly become a purchase commitment or a customer promise.
The software market spans CAD-driven manufacturing simulation, enterprise product-costing systems, process-specific estimators, market-data products and newer AI tools. They solve different problems. The right choice depends on what you buy, how early you need the answer and who will maintain the cost logic.
Define the decision first
Teams use “should-cost” to describe several calculations. A procurement model for a machined casting asks a different question from a target-cost exercise for a new product platform. Put the decision, subject and date on the first page of every calculation.
| Decision | Question | Evidence that matters |
|---|---|---|
| Supplier negotiation | What would an efficient supplier spend to make and deliver this requirement? | Drawing, process route, material, cycle time, batch size, regional rates, yield, overhead and reasonable profit |
| Source selection | Which qualified offer has the best economic basis? | Should-cost, supplier breakdowns, capacity, lead time, quality, freight, duty and commercial terms |
| Make or buy | Should our plant make the part or should we purchase it? | Avoidable internal cost, capacity constraint, investment, supplier price, logistics and risk |
| Design to cost | Which features or requirements create cost before the design freezes? | Geometry, tolerance, material, process choice, tooling, test and volume |
| Customer bid | What should the purchased scope in this offered configuration cost us? | Customer revision, chosen design, supplier evidence, escalation, delivery and unresolved risk |
The aPriori should-cost guide draws a useful line between bottom-up should-cost and a top-down estimate based on prior orders. Siemens describes a similar bottom-up use of Teamcenter Product Cost Management for purchased parts, design alternatives and supplier negotiation. Both approaches can inform a customer estimate, but neither equals a supplier commitment.
Separate should-cost from the numbers around it
| Number | What it means | Common misuse |
|---|---|---|
| Should-cost | A modeled cost and price under defined manufacturing and commercial assumptions | Presenting an efficient benchmark as the price any supplier can achieve now |
| Will-cost | The expected purchase price under the supplier, volume, timing and market conditions you expect to use | Treating it as independent when it came from the same quote under review |
| Target cost | The cost the product must reach for the business case or customer price to work | Calling a commercial requirement a manufacturing fact |
| Supplier quote | A supplier’s offer for a defined scope, quantity, date and terms | Comparing the headline unit price without checking exclusions or amortization |
| Standard cost | The approved accounting value used by ERP for inventory and variance | Using an old standard as a current sourcing benchmark |
| Total cost of ownership | Purchase price plus logistics, inventory, quality, support, risk and other lifecycle effects | Adding vague risk percentages without an event, exposure or owner |
This vocabulary prevents avoidable fights. A buyer may have a sound should-cost of $320, a target of $300 and a supplier quote of $347. Those numbers can all be correct. The work is to explain their bases and decide what can change.
Build a model another cost engineer can reproduce
A should-cost model should read like a manufacturing plan with prices attached. Each line needs a quantity, unit, rate, source, effective date and calculation. Each override needs an owner and a reason. If a second cost engineer cannot reproduce the result, the model is only an opinion with decimal places.
| Cost block | Model inputs | Checks |
|---|---|---|
| Material | Grade, stock form, gross mass, finished mass, yield, buy rate and scrap credit | Match the drawing and buying unit; avoid counting scrap credit twice |
| Conversion | Operation sequence, machine class, cycle time, setup, labor attendance, energy and consumables | Confirm the process can meet tolerance, finish and volume |
| Tooling | Tool design, manufacture, maintenance, life and amortization quantity | Separate one-time cash from per-unit recovery |
| Quality | Inspection, test, qualification, documentation, scrap and rework | Use the actual customer and regulatory requirements |
| Factory burden | Machine, labor, facility and indirect cost basis | Know which costs the hourly rates already contain |
| Commercial load | Supplier SG&A, profit, warranty and financing | State whether margin is a markup on cost or a percent of selling price |
| Landed cost | Packaging, freight, duty, insurance and inventory | Use the named lane, Incoterm and delivery pattern |
GAO’s Cost Estimating and Assessment Guide gives the same practical rule at a larger scale: document the source, content, time, units, accuracy and reliability of cost data. A software product should make that discipline easy and expose the inputs behind every score.
Follow one supplier quote down to the process
Assume an OEM buys a machined ductile-iron pump housing at 1,200 units per year. The current supplier quotes $347 per housing plus $18,000 of new tooling. The drawing calls for machined mounting faces, two precision bores, a pressure test and a protective coating. The buyer wants a price position before renewing the agreement.
The first model uses a 36 kg casting blank, three machining operations and a 600-piece production lot. It does not start from the supplier price. It starts from a route that can make the drawing.
| Modeled element | Basis | Cost per housing |
|---|---|---|
| Casting | Material, melt loss, molding, cleaning and supplier conversion | $138.00 |
| Machining | Three operations, 41 minutes total, machine and attended labor rates | $76.00 |
| Coating and cleaning | External coating plus handling | $18.00 |
| Inspection and pressure test | Dimensional inspection, test stand time and records | $14.00 |
| Packaging | Returnable dunnage allocation and preservation | $6.00 |
| Factory overhead | Costs not already carried in the process rates | $23.00 |
| Manufacturing cost | Sum of modeled production cost | $275.00 |
| SG&A and profit | 12% of selling price | $37.50 |
| Freight | Named plant-to-plant lane | $9.00 |
| Modeled landed price | Before separate tooling | $321.50 |
The initial gap is $25.50 per unit, or 7.3% of the quoted price. That does not prove the quote is high. The supplier may run smaller lots, use a different inspection plan, carry dedicated capacity or include warranty exposure that the model missed. The gap tells the team where to investigate.
Reconcile the gap before you negotiate
| Question to test | Evidence to request | Possible effect |
|---|---|---|
| Why does the supplier machine in four setups? | Routing, fixture concept and process capability | The fourth setup may be necessary for datum control |
| Which volume drives the quote? | Lot size, release pattern and annual commitment | Smaller releases can raise setup and material cost |
| What does the test requirement include? | Test time, equipment, records and rejection handling | A customer-specific test record may add real labor |
| How is tooling recovered? | Tool scope, ownership, life and amortization schedule | The unit price may already contain part of the tooling |
| Which commercial risks did the supplier price? | Warranty, payment, liability, currency and capacity terms | A cost gap may sit outside the factory route |
Suppose the supplier shows that each monthly release is 200 units, the customer requires an extra pressure-test record and the quote includes premium inbound freight. Those facts add $8, $12 and $6. The reconciled should-cost becomes $347.50. Procurement learned that the supplier price is sound and found the actions that could lower it: larger releases, a simpler record or a different freight plan.
That is a successful should-cost review. The value lies in the explanation and the options, even when the final answer supports the supplier.
Use the right modeling method for the evidence you have
| Method | Best input | Strong use | Main weakness |
|---|---|---|---|
| CAD-driven simulation | 3D model, material, volume and location | Machined, fabricated, cast, molded or forged parts with recognizable geometry | A clean geometry result can still miss customer-specific quality, logistics or commercial scope |
| Bottom-up process model | Route, cycle-time drivers, machines and rates | Parts where a cost engineer understands how the process works | Model quality depends on route and reference-data maintenance |
| Parametric model | A small set of product or process characteristics | Early concepts and repeated families | Relationships can fail outside the data range that created them |
| BOM and assembly roll-up | Structured BOM, purchased content, routings and labor | Complex assemblies and product families | Weak part-level assumptions can disappear inside the roll-up |
| Historical analogue | Comparable quotes, orders and actuals | Fast checks on stable, repeated parts | The nearest part number may not share the cost drivers |
| Supplier cost breakdown | Supplier route, rates and commercial loads | Joint review and negotiation | The structure may reflect the supplier’s accounting rather than economic causality |
| Market benchmark | Material, labor, machine and regional data | Location and escalation checks | A market rate does not define the supplier’s actual factory |
A mature team uses more than one method. A CAD simulation can create the independent baseline. The supplier breakdown explains the current offer. Historical prices show the commercial range. Actual performance tells the team which model deserves more trust.
Choose software by the work it owns
The products below overlap, but their strongest jobs differ. Judge each one on the work your team needs to repeat, not on the length of its feature list.
| Product | Strongest fit | How it builds cost | What to prove |
|---|---|---|---|
| aPriori | Automated cost and manufacturability analysis from 3D CAD across multiple processes and regions | Recognizes geometry and simulates production in configurable digital factories | Coverage for your parts, routing accuracy, factory calibration, unsupported features and model-maintenance effort |
| Teamcenter Product Cost Management | Enterprise cost engineering tied to products, BOMs, tools, sourcing and lifecycle decisions | Bottom-up process and assembly calculations with reference data and Teamcenter context | PLM and ERP integration, master-data ownership, calculation governance and adoption outside specialist teams |
| FACTON EPC | Standardized enterprise product costing, quotation costing, should-cost and design-to-cost | Company calculation schemes, BOM structures, master data and collaborative workflows | Configuration time for your methods, variant handling, change control and reporting |
| DFMA Should Costing | Transparent part-level process models plus design simplification | Manufacturing-process mechanics, material and machine libraries, CAD or manual feature input | Process coverage, local rates, treatment of tolerances and the handoff from part result to sourcing case |
| costdata calculation | Should-cost with current international market and location data | Process calculations combined with material, labor, machine and country benchmarks | Data coverage, update cadence, regional specificity, TCO logic and export into your sourcing process |
| Costimator | Machining and fabrication estimates built around shop processes | Prebuilt cost models and cycle-time calculators, with or without CAD | Match to your machine and process set, estimator overrides, ERP exchange and use on purchased parts |
| PartSpace | Cost engineering and pricing intelligence across drawings, CAD, history and supplier data | AI analysis and benchmarks across technical and commercial records | Explainability, coverage on your commodities, data isolation, benchmark provenance and reviewer control |
| Bourne | Customer-specific cost cases that cross RFQs, ERP, engineering, specialist cost tools and supplier quotes | Assembles source records, calls the right calculation, tracks missing evidence and routes decisions | Connector depth, exception handling, source traceability and transfer into pricing, proposal and order review |
The vendors describe these positions in their current product material: aPriori explains its Cost Model Workbench and digital factories; Siemens documents bottom-up purchased-part costing and design alternatives; FACTON covers should-cost, design-to-cost and quotation calculation; DFMA Should Costing lists its process models and CAD/manual inputs; costdata combines calculation with market data; Costimator focuses on manufacturing process models; and PartSpace combines technical and purchasing evidence.
Test process coverage at the level that changes cost
A vendor may say it covers machining, casting or fabrication. That label is too broad for a buying decision. Ask which routes, materials, feature types and secondary operations the model understands. Then test the cost drivers that matter in your portfolio.
| Commodity | Cost drivers the proof should exercise |
|---|---|
| Machining | Stock choice, feature recognition, removal rate, tool changes, setup, fixturing, tolerance, finish and unattended time |
| Sheet metal | Nesting, yield, cut length, pierces, bend count, tooling, welding, finish and batch handling |
| Castings and forgings | Gross-to-net mass, alloy, process, tooling, cavity or mold life, melt loss, yield, heat treatment and finish machining |
| Injection molding | Part and runner mass, cavitation, cycle time, press size, tool construction, tool life and resin |
| Electrical assemblies | BOM, placement or assembly labor, test, harness content, scrap, compliance and supplier-specific components |
| Industrial equipment | BOM roll-up, bought-out equipment, fabrication, assembly, engineering, test, documentation, freight and commissioning scope |
A product that models a turned shaft well may add little value to a configured compressor skid dominated by bought-out motors, controls, certification, engineering and test. Match the engine to the spend and the decision.
Treat reference data as a product you must maintain
Software demos often focus on the calculation. Production results depend just as much on the reference data: material prices, labor, machines, overhead, exchange rates, logistics and supplier economics. Ask which data the vendor provides, which data your team must own and how a calculation records the version it used.
| Data question | Why it matters |
|---|---|
| What geography and effective date does the rate represent? | A regional average cannot silently stand in for a named supplier plant |
| Does the machine rate include labor, energy and overhead? | Different rate structures can double-count or omit the same cost |
| Can we override a rate without losing the original? | Negotiation needs the benchmark, the supplier fact and the approved decision |
| Who reviews vendor data updates? | A new library version can move thousands of models at once |
| Can we reproduce an old calculation after rates change? | Customer bids and supplier agreements need an auditable historical basis |
| Can we export our models and data? | The cost method is company knowledge, not disposable software setup |
aPriori lets customers configure digital factories and edit model logic through its workbench. Siemens supplies reference data for labor, material, machines and processes. costdata makes current international market data part of its offer. Each product assigns different work to the customer; test the one your cost-engineering team can sustain.
Model supplier profit and commercial scope openly
A should-cost used for purchasing needs a view of the supplier’s business. Manufacturing cost alone is not a fair price. Suppliers need to cover selling, administration, working capital, warranty, investment and profit. The model should show those assumptions as separate lines instead of burying them in a plant overhead rate.
| Commercial item | Question to state |
|---|---|
| Profit | Is the rate a markup on cost or a margin on selling price? |
| Capacity | Does the buyer require reserved equipment, labor or inventory? |
| Payment | How much working capital follows from the payment schedule? |
| Warranty | Which failure exposure and field support sit with the supplier? |
| Currency and escalation | Who carries movement between quote, order and delivery? |
| Liability and compliance | Which insurance, certification, audit or reporting cost applies? |
The U.S. Defense Department’s commercial pricing study defines commercial should-cost as the internal view of design, manufacturing and delivery cost plus reasonable profit. The same study found commercial companies using should-cost for offer evaluation, make-or-buy decisions and post-award review. That broader view is more useful than treating every variance as excess supplier margin.
Make uncertainty visible
A model built from an early drawing and benchmark rates should not display the same confidence as a model calibrated with the supplier route and recent actuals. Put a range around the result and identify which assumptions move it.
| Scenario | Casting | Machining | Other cost | Landed price |
|---|---|---|---|---|
| Low | $131 | $69 | $70 | $306 |
| Base | $138 | $76 | $70 | $322 |
| High | $149 | $88 | $74 | $350 |
In the pump-housing example, the $347 quote sits near the high end of the credible range. Sensitivity shows that batch size, the fourth setup and inspection scope drive most of the spread. Those become the review agenda. A generic “model confidence: 87%” does not.
GAO distinguishes sensitivity analysis from risk and uncertainty analysis. The first changes one assumption to show its effect. The second describes the combined range of possible outcomes. Good software should support both and preserve the assumptions behind each scenario.
Use the model with suppliers, not against them
Send questions that a manufacturing engineer can answer. “Your quote is 12% too high” invites a positional fight. “We modeled two machining setups; your breakdown shows four. Which features force the extra fixtures?” can uncover a drawing issue, a supplier constraint or a better route.
The Defense Federal Acquisition Regulation’s should-cost review rules offer a useful operating principle even outside defense: involve people with relevant process and cost expertise, engage the supplier early, use information already available and define measurable savings actions. The review should create a joint fact base, not a surprise demand.
| Bad challenge | Useful question |
|---|---|
| “The model says you should be cheaper.” | “Which requirement or factory condition explains the $26 gap?” |
| “Your labor rate is too high.” | “Does this rate include machine burden, supervision or benefits that our benchmark holds elsewhere?” |
| “Cut your margin to 8%.” | “Which volume or commitment would let us remove capacity and inventory cost?” |
| “Another region is cheaper.” | “What changes when we include freight, duty, lead time, minimum order and buffer stock?” |
Connect should-cost to the customer bid
Industrial OEMs often run should-cost inside a live customer response. The buying team may not have time to finish a sourcing event before sales must price the package. The system therefore needs to distinguish the benchmark, budgetary supplier input, firm quote and approved bid assumption.
| Bid event | Required should-cost action |
|---|---|
| Customer changes quantity | Recalculate batch, tooling amortization, supplier minimums and logistics |
| Drawing revision arrives | Identify affected geometry, material, process, test and bought-out content |
| Supplier quote expires | Reopen the exposed line or apply an approved, traceable escalation basis |
| Sourcing selects another supplier | Replace the provisional price and record lead-time, terms and scope differences |
| Price approval starts | Show the chosen cost, alternatives, uncertainty, owners and unresolved risk |
| Customer order arrives | Compare the PO and final scope with the cost basis before release |
This is where a specialist cost engine and Bourne do different work. The engine should calculate the part or assembly using its manufacturing logic. Bourne can read the customer request, send the correct revision to that engine, collect supplier evidence, expose missing inputs and carry the approved cost into price, proposal and order review.
Run a pilot that can disprove the software
A polished demo proves very little. Select 12 to 20 items that represent the cost problems your team actually faces. Include stable repeat parts, a new design, an awkward outlier and at least one model the incumbent method gets wrong. Hold back recent supplier and actual data until the first models are complete.
| Pilot test | Pass condition |
|---|---|
| Input readiness | The team can build a first model from the records it normally has at that decision point |
| Route validity | A manufacturing expert agrees that the modeled process can make the requirement |
| Driver response | Material, volume, tolerance, region and process changes move cost for explainable reasons |
| Quote reconciliation | The tool helps explain material gaps against supplier breakdowns and prices |
| Repeatability | A second cost engineer can reproduce the result from the stored model and data version |
| Throughput | The team can cover enough annual spend or bid volume to justify model creation and upkeep |
| Closed loop | Supplier facts and actual outcomes improve the model instead of living in separate files |
Measure the median absolute variance, but do not stop there. Count how much of each gap the team explains, how long a model takes, how often a specialist must intervene and whether the result changes a sourcing, design or bid decision. Accuracy on easy parts can hide failure on the parts that matter.
When Bourne is the best fit
Choose aPriori, DFMA, Costimator, Teamcenter, FACTON or costdata when the main problem is the manufacturing calculation itself. They contain process models, reference data or enterprise costing methods that Bourne should not pretend to replace.
Bourne is best when the should-cost must become part of a customer-specific decision that crosses sales, engineering, sourcing and finance. It can assemble the current RFQ and drawing, call the specialist model, compare its result with supplier offers, route the real gaps to the right people and preserve the approved assumption through pricing and order review.
That distinction also lets the OEM use different engines for different commodities. A machined part can go to a CAD-driven model. A configured assembly can use an enterprise costing system. A bought-out package can use supplier RFQs and benchmarks. Bourne gives the bid one source record and makes the unresolved work visible. See the product costing workflow for the application.
Arda Bulut
Arda Bulut is the co-founder and CTO of Bourne and HockeyStack. He leads engineering at Bourne, building the platform people use to create AI products, agents and automations.