What "Should-Cost" Actually Means in Foundry Pricing

Should-cost analysis originated in defence procurement - the US Department of Defense used it systematically from the 1960s to estimate what a contracted component should cost a supplier to produce, as a basis for negotiating prices rather than accepting quotes at face value. The concept has since spread across industrial procurement, but the implementation quality varies enormously.

In the context of sand casting, should-cost has a specific meaning. It is a bottom-up estimate of what a competently run foundry, operating with efficient process parameters, would need to charge to cover its costs and earn a reasonable margin on a specific part - at current market prices for inputs.

What should-cost is not

Should-cost is not the cheapest price in the market. It is not an industry average. It is not what a competitor quotes. Each of these is a different concept, and using them interchangeably leads to poor decisions.

The cheapest price in the market may reflect a foundry that is underinvesting in quality, running old equipment at high utilisation, or - in some cases - deliberately pricing below cost to fill capacity with the intention of increasing prices later. Using the lowest market quote as a should-cost benchmark leads procurement teams to demand prices that are unsustainable and to attract suppliers who will cut corners to deliver them.

An industry average is a lagging indicator - it reflects what buyers have been paying, not what an efficient producer should cost today. In a period of rapid input price movement (as casting markets have experienced since 2020), average pricing data from last year may be significantly wrong in either direction.

How a foundry should-cost model works

A proper should-cost model for sand castings builds up from first principles, component by component. The input parameters are:

1. Part geometry: casting weight, internal volume (for core estimation), complexity classification, and the moulding line type required.
2. Material specification: alloy grade, which determines scrap mix, alloy additions, and melt treatment requirements.
3. Production parameters: annual volume (which determines whether automatic or manual moulding is appropriate), cavities per mould, and finishing requirements.
4. Market data: current scrap prices, energy tariffs, and labour costs in the target market - updated to reflect current conditions, not annual averages.
5. Process assumptions: yield, scrap rate, cycle time - based on what an efficient foundry running this part type should achieve, not what a specific supplier claims.

The model then calculates a cost per kilogram and per piece, broken down by component: material cost, moulding cost, melt cost, overhead, finishing, and margin. Each component is separately visible and separately defensible.

The efficiency assumption is critical

The most important - and most contested - input in any should-cost model is the efficiency assumption. What yield, scrap rate, and cycle time should you assume?

CastCalc uses what we call the "efficient foundry" benchmark: the process parameters achievable by a competently run, modern foundry operating the appropriate equipment for the part type. Not the best foundry in the world. Not the average foundry. A competent, efficient operator.

This matters because it determines the fair range of the should-cost output. If you assume world-class efficiency, you calculate a should-cost that only the best foundries can achieve - and using it as a negotiating target is unreasonable. If you assume average efficiency, you are benchmarking against the mediocre, which tells you nothing about whether there is room for improvement.

The efficient-foundry assumption gives you a reference point that a good supplier can achieve and sustain. The gap between your current supplier's price and the efficient-foundry should-cost is the negotiating space - it represents either inefficiency that can be addressed, or margin that exceeds a reasonable level.

How to use should-cost in practice

There are four practical applications of a should-cost benchmark in casting procurement:

1. RFQ evaluation: When quotes arrive, compare each to the should-cost benchmark. Quotes significantly above warrant a cost breakdown request. Quotes significantly below warrant a quality risk assessment - a price 20% below should-cost usually means an unsustainable concession.
2. Price increase challenges: When a supplier requests an increase, recalculate the should-cost at new input prices. The justified increase is the difference between old and new should-cost. If the claimed increase exceeds this, the excess needs to be explained.
3. Design review: Run the should-cost model on a proposed design before tooling is committed. Then run it again with modifications - eliminating a core, adjusting wall thickness, changing the parting line. Cost impact becomes visible before it is locked in.
4. New supplier evaluation: Should-cost gives you a reference price to present to new suppliers rather than asking for a blind quote. "Our model suggests approximately X in your market - can you confirm whether you can be competitive?" This signals commercial sophistication and saves time on both sides.

The limits of should-cost

Should-cost is a reference point, not a target price. There are legitimate reasons why a supplier's actual cost may exceed the efficient-foundry benchmark - and a procurement team that uses should-cost as a ceiling rather than a reference will damage supplier relationships and eventually find itself without good partners.

Legitimate cost drivers above the benchmark include: higher quality assurance costs for safety-critical applications, tooling amortisation on low-volume parts, geographic premiums for proximity and flexibility, and overhead structures at smaller foundries without the volume to achieve large-foundry efficiency.

The right use of should-cost is to open a conversation, not to close one. "Our benchmark suggests a cost in the range of X - your quote is Y. Can you help us understand what drives the gap?" This question, asked in good faith, usually produces a productive discussion. Used as a blunt instrument - "we will not pay more than X" - it produces resentment and eventually worse outcomes.

Why most should-cost models fail

The most common failure mode in should-cost analysis is stale data. A model built on last year's scrap prices, last cycle's energy costs, and outdated labour benchmarks produces a should-cost figure that is confidently wrong. In casting markets specifically, where scrap and energy move meaningfully month to month, a should-cost model that is not updated regularly is worse than no model - because it gives a false sense of certainty.

The second failure mode is insufficient process knowledge. A should-cost model that uses the right formula but wrong process parameters - assuming 80% yield when the part geometry realistically achieves 60%, or assuming automatic moulding when the part requires manual - will produce an output that is systematically biased. This is why the model must be built by someone who understands foundry processes, not just accounting.

CastCalc addresses both: monthly market data updates in 22 markets, and process parameters built from real foundry experience rather than academic models. The combination gives a benchmark that is current and realistic - the two properties that matter most for practical use.