How to Build a Resin Should-Cost Model That Holds Up in a Negotiation

Your supplier sends over next year's number. Sixty-eight cents a pound, starting in January, on the grade you run every single day. Sure, you can check it against an index or last year's contract, or ask your plant manager what you paid the year before that. But without a resin should-cost model, you can't say what that pound actually costs to make and move.

That cost is the floor under the quote. A should-cost model is your own estimate of what something ought to cost, built up from the costs underneath it instead of worked backward from the quote. Procurement teams already run this kind of should-cost analysis on machined parts and molded components, where it sometimes goes by the name clean-sheet costing. Yet almost nobody uses it on resin, which is pretty strange when resin usually makes up 50 to 80% of COGS for a plastics processor.

Key Takeaways

• A resin should-cost model estimates what a delivered pound ought to cost by building up the feedstock, freight, adders and producer margin underneath it, rather than starting from the quote.
• Standard should-cost frameworks assume a machined part, so they treat raw material as a pass-through line instead of the whole purchase.
• Producer margin and freight are the two layers a buyer can realistically push back on, and a published resin price doesn't break out either one.
• The model gives you a defensible range and a specific question rather than a number to the tenth of a cent. You rebuild it monthly, not annually.
• If your last three price conversations all started from the supplier's number, the problem probably wasn't how hard you pushed.

Why Standard Should-Cost Frameworks Break on Resin

Most guides to should-cost modeling are written for a manufactured part, like a machined bracket. They add up material, labor, overhead, and a supplier profit, with almost all the detail going into cycle time, tooling, and yield.

With resin, the material line is the whole purchase, and a pound of pellets has no cycle time or tooling to cost out. So you build the model around what actually goes into a pound of resin, with a live resin price benchmark doing the job a cycle-time database does for a machined part. Makes sense, right?

What Actually Goes Into a Pound of Delivered Resin

Once you build it around the material, a resin cost breakdown has five layers. They don't look much like the cost buckets in a part model.

• Feedstock: Ethane and naphtha get cracked into ethylene and propylene, which producers then polymerize into the pellets you buy. Feedstock costs move the most of any layer. They also come with public reference points, since polymer-grade propylene spot pricing gets published and so does ethane.
• Conversion: This is what it costs a producer to turn monomer into polymer. You'll never see that number, so any model claiming precision here is guessing. You can still read whether a producer has room to absorb an increase, and operating rates are the clearest public signal.
• Logistics and packaging: Your landed cost depends on how the resin ships, whether that's rail car, bulk truck or gaylord, along with how far you are from the Gulf Coast and whatever demurrage clock came with the contract. Published prices also carry an assumption buyers tend to skip past. Plastics News quotes prime resin, unfilled and natural color, FOB supplier, while your dock isn't FOB supplier and your grade probably isn't natural color.
• Grade and additive adders: Color, UV package, antiblock, slip and small-lot penalties. You could negotiate each one separately, and almost nobody bothers.
• Producer margin: The swing factor, and the only layer where the number is a decision rather than a cost.

That last one is the point of the whole exercise. You can track feedstock and measure freight and adders down to the cent, even though conversion stays out of sight. Whatever's left once you've stacked the defensible layers is margin. Whether you call it producer margin or supplier margin, resin suppliers would much rather you called it "the market."

Building Your Resin Should-Cost Model

So you know the layers. Filling them in starts with data you already have, and usually more of it than you'd think: twelve to twenty-four months of invoices by grade, plant and supplier, with freight terms and volumes attached. Most of it's already in the detail you pull together when you work a resin quote properly. If it's spread across three spreadsheets and a purchasing system, pulling it into one resin cost model is the actual first task.

Then anchor the model to feedstock instead of price. Take your two or three highest-volume grades and map each one back to its monomer. From there, compare how far the feedstock moved over the last few quarters with how far your price moved. If polymer-grade propylene came off nine cents and your PP copolymer came off three, you've found something to ask about using nothing but public data.

After that, layer freight and adders on at your actual delivery points and packaging. It's pretty unglamorous arithmetic, but it's usually where you find the first savings, because nobody goes back and re-prices a freight assumption from four years ago.

Then check the whole stack against transaction data. A cost build-up tells you what a pound should cost, but it won't tell you what comparable processors are paying this month for the same grade at the same volume. You need both numbers before you sit down with your supplier.

Think there's margin hiding in your portfolio company's resin procurement?

Start your free trial at resinsmart.ai and see where the opportunity sits. 

What a Resin Should-Cost Model Can and Can't Do

Once you've built one, the easiest mistake to make is chasing precision, which is why plenty of these models never get finished. Even on a machined part, procurement software company Precoro says a good model lands within about 10 to 15% of the actual price. And Boothroyd Dewhurst, who developed the DFMA method, says a model was never meant to predict a supplier's exact price, only to show which assumption explains the difference.

For resin, you can aim even looser than that. Instead of landing on 61.4 cents per pound, you just want to walk in able to say the stack supports something in the low sixties. The quote came in at sixty-eight, so you'd like to understand the six cents. Suppliers handle that a lot better than pressure, because a supplier can't do much with "that's too high" and can do plenty with a feedstock number.

There are limits, though. A should-cost model won't do much when supply is genuinely short, when you're single-sourced on a specialty grade, or when a producer has decided margin recovery is this year's plan regardless of cost. Boothroyd Dewhurst's guidance flags those same conditions.

The model does work well on commodity grades in a balanced market, and the same math is useful outside purchasing too. Resin procurement due diligence on a plastics target asks the same question, and a portfolio company without that discipline usually carries a few cents a pound nobody priced into the deal.

How ResinSmart Supports a Resin Should-Cost Model

A should-cost model is only as good as the market number you check it against, which is the one piece buyers can't build themselves. ResinSmart benchmarks fair-market resin pricing from 3B+ lbs of actual buyer transactions across polyethylene, polypropylene, polystyrene, PVC, PET and engineered resins. Our benchmarks update as transactions come through rather than on a survey cycle. Once you put them next to feedstock and operating-rate data, your build-up becomes something you can take to a supplier, not just an internal estimate.

Start a free trial and see where your delivered resin cost sits against what comparable buyers are paying.

Resin Should-Cost Model FAQs

These are the questions buyers usually raise when they start building one.

What is a resin should-cost model?

A resin should-cost model is a buyer-side estimate of what a delivered pound of resin ought to cost, built from the feedstock, conversion, freight, grade adders and producer margin underneath it rather than from the supplier's quote. Unlike a part-level should-cost model, it treats raw material as the entire purchase rather than a pass-through line.

Does should-cost modeling actually work for a commodity like resin?

Should-cost modeling works for resin, though not in its standard form. Generic frameworks assume a manufactured part and weight everything toward cycle time, tooling and yield, none of which apply to pellets. A resin model is built instead around feedstock, freight, grade adders and producer margin.

What data do you need to build a resin should-cost model?

A resin should-cost model needs four inputs: 12 to 24 months of purchase history by grade, plant and supplier, public feedstock pricing for the relevant monomers, your actual freight terms and delivery points, and a transaction-based market benchmark. The first three you already own. The fourth is the one most buyers have to source.

How is a resin should-cost model different from a resin price index?

A resin should-cost model estimates what a pound ought to cost, while a resin price index reports what the market is understood to have paid. The model is assembled from resin cost drivers you can trace, so it explains why a price should move. An index gives the level without the reasoning, which is why buyers run both.

How accurate does a resin should-cost model need to be?

A resin should-cost model needs to be accurate enough to identify which cost layer explains a price difference, not accurate to the tenth of a cent. Precoro's guidance puts achievable variance on part-level models at roughly plus or minus 10 to 15%. For resin the useful output is a defensible range and a specific question about the difference.

How often should a resin should-cost model be updated?

A resin should-cost model should be updated monthly. Standard guidance recommends annual updates with quarterly refreshes for volatile categories, which is far too slow here. Since resin contract prices move monthly on producer nominations and settlements, and feedstocks move even faster, a model rebuilt once a year is out of date most of the time.