Latest / The Buyer's Side with Fexingo: Procurement, Sourcing, and B2B Purchasing Decisions / How One Buyer Used Should-Cost Modeling to Cut Costs 17 Percent
Transcript
- Lucas: So here's a sentence you hear in procurement circles: 'We need to know what something should cost, not just what the supplier says it costs.' That's the essence of should-cost modeling. And we're going to talk about one buyer who used it to cut spend by 17 percent. Luna: That sounds like a big number for a single negotiation. What was the product? Lucas: It was a powertrain component — think of a transmission housing — for a mid-size automotive parts manufacturer. The buyer, a procurement manager named Rachel, had been paying $18.50 per unit for this casting from a single-source supplier. She suspected there was slack in the price, but she didn't have the engineering data to prove it. Luna: Right, so she needed to build a model from the ground up — material, labor, machine time, overhead. Lucas: Exactly. And that's the core of should-cost. You're not asking the supplier to open their books. You're modeling what a theoretically efficient manufacturer would need to charge to produce that part at a reasonable margin. Rachel worked with her engineering team to estimate the weight of the casting — about 12 pounds of aluminum. Then she looked at current aluminum prices, which were around $1.10 per pound at the time. That gave her a raw material cost of $13.20 per unit, before any processing. Luna: But the part isn't just a block of metal. There's machining, finishing, quality checks. Lucas: Right. She added standard industry estimates for die-casting cycle time — about 90 seconds per part — and labor rates for the region. She factored in tooling depreciation, energy costs for the furnaces, and a 10 percent scrap rate. When she added up all the direct costs, she got to about $12.80 per unit. Then she added 15 percent for SG&A and 8 percent profit, which gave her a should-cost of $15.35. Luna: So $15.35 versus the current price of $18.50. That's a gap of over three dollars — 17 percent. Lucas: That gap gave her the confidence to go to the supplier. She didn't reveal her model at first. She just said, 'We've done some internal analysis, and we believe there's room to move on price.' The supplier came back with a best and final of $16.80. But Rachel was patient. She asked for a breakdown of their cost structure. They resisted. So she eventually shared her should-cost model — not as a demand, but as a collaborative tool. Luna: That's a delicate moment. Some suppliers see a should-cost model as an accusation of price gouging. Lucas: And that's the risk. But Rachel framed it as, 'Help me understand where my model is wrong.' And to their credit, the supplier pointed out that her scrap rate was too low — they actually ran at 14 percent. And they had higher energy costs because their facility was in a region with expensive electricity. So she adjusted her model to $16.00. They settled at $15.35, exactly her original target. The supplier agreed because they saw it as a fair, data-driven number. Luna: That's the ideal outcome. But I've seen should-cost backfire. I worked with a buyer who tried this on a specialty chemical supplier. He built a model using commodity chemical prices, but the supplier's process involved a proprietary catalyst that cost ten times more. The supplier got offended, and the relationship soured. Lucas: That's a great cautionary tale. Should-cost works best when the product is standardized and the input costs are transparent. For highly engineered or proprietary items, you need a different approach. Which brings me to a rule of thumb: should-cost is most effective for mature products with multiple potential suppliers. Rachel's casting was basically a commodity-grade aluminum part. The technology wasn't cutting-edge. Luna: Let's talk about the data inputs. Building a should-cost model sounds like it requires deep engineering expertise. Is that something a typical procurement team can do? Lucas: It depends. For mechanical parts, you can often use standard cost-estimating software like aPriori or Siemens Teamcenter. These tools have libraries of manufacturing processes — machining, injection molding, stamping — with cycle times and cost rates built in. Rachel used a combination of aPriori and her own Excel spreadsheet. The key is to involve your engineers early. If procurement builds the model in isolation, it's easy to miss critical specs. Luna: And the time investment? Rachel's project sounds like it took weeks. Lucas: About three weeks from start to negotiation. But she was only doing one part. For a portfolio of hundreds of parts, you'd prioritize by spend. The 80-20 rule applies: focus on the components that account for the biggest cost. Rachel's company had about $2 million annual spend on that casting alone, so the 17 percent savings translated to $340,000 per year. That justified the effort. Luna: Speaking of effort, if today's deep dive on should-cost modeling gave you a practical takeaway you can use, that's exactly why this show exists. Lucas and I love doing this, and the way we keep it ad-free is through listener support. Lucas: Yeah. If these conversations have moved your work forward in some small way, buy me a coffee dot com slash fexingo. It's that simple. No tiers, no perks — just a way to say 'this matters.' And it genuinely helps us keep going. Luna: All right, back to should-cost. One thing I want to dig into is the negotiation tactic. Rachel shared her model — is that always the right move? Lucas: Not always. If you're dealing with a supplier that has a strong competitive position — say, a sole source with proprietary technology — sharing your model can look naive. But in a competitive market, transparency can build trust. Rachel's supplier didn't have many alternatives either; they were the only qualified source for that casting. But they also knew that if Rachel's model was accurate, a competitor could eventually enter at that price. Luna: So should-cost is a tool, not a weapon. Use it to open a conversation, not to dictate terms. Lucas: Exactly. And the best practitioners use should-cost in combination with other data. For example, they'll benchmark the should-cost against market intelligence from sources like IHS Markit or Procurement Leaders. If your model says $15 and the market index says $16, you might be missing something. Rachel's model was within 5 percent of the index, which gave her confidence. Luna: What about the opposite scenario — what if the supplier's price is already below your should-cost? Do you go back and tell them they're undercharging? Lucas: That happens more often than you'd think. And the ethical answer is: you don't. If a supplier quotes below your should-cost, it could mean they have a more efficient process, or they're buying raw materials at a discount you didn't know about. Your model might be wrong. You don't punish them for being efficient. But you do keep your model updated. If their cost advantage is real, it's a sign that your other suppliers might be overpriced. Luna: Let's talk about the long-term impact. After Rachel's negotiation, did the relationship stay healthy? Lucas: According to the case study I read — this was published in a procurement journal — yes. The supplier actually improved their own processes to hit the $15.35 target. They invested in a more efficient furnace that reduced their energy cost by 8 percent. A year later, they were making a healthy margin at that price. And Rachel awarded them additional volume for other parts. The should-cost model became the basis for a transparent cost structure going forward. Luna: That's the ideal — cost transparency as a foundation for partnership. But I've heard critics say should-cost is just a fancy way to squeeze suppliers. How do you respond to that? Lucas: I'd say it can be used that way. But when it's used as a one-way hammer, it damages the supply base. The best procurement organizations use should-cost as a diagnostic. If a supplier's price is above should-cost, the question is why. Maybe they have a legitimate cost disadvantage — older equipment, higher wages. In that case, you work together to close the gap. Rachel's approach was collaborative. She didn't demand $15.35 on day one. She listened, adjusted, and found a win-win. Luna: So for listeners who want to try should-cost modeling, what's the first step? Lucas: Pick one high-spend, non-critical part — something you buy in volume, where you have a rough idea of the materials and process. Pull together your engineering team for a two-hour workshop. Map the manufacturing steps. Estimate cycle times and material usage. Use online resources like Industrial Quick Search or ThomasNet for rough cost benchmarks. Your first model won't be perfect, but it will give you a starting point. Then go to your supplier and say, 'We want to understand your cost structure better so we can work together on competitiveness.' That opens the door. Luna: I like that. It's less about the number and more about the conversation. One thing I want to circle back on: Rachel's 17 percent savings — that's impressive, but it's not sustainable if every negotiation is a one-time event. How do you institutionalize should-cost? Lucas: Great question. A lot of companies build should-cost models into their annual supplier reviews. They update the input data — commodity prices, labor rates, exchange rates — and then revisit the target price. Some even share a 'cost model dashboard' with strategic suppliers. That way, when aluminum prices go up 5 percent, both sides agree on the impact. It takes the guesswork out of price adjustments. Luna: And I'm guessing it also helps with the 'we're in this together' dynamic. When a supplier knows you're not just trying to squeeze them, they're more likely to share cost-saving ideas proactively. Lucas: Exactly. One company I read about had a supplier suggest a design change — switching from a machined part to a near net shape casting — that reduced the should-cost by 12 percent. The supplier got a share of the savings. That never would have happened if the buyer had just demanded a lower price without understanding the cost drivers. Luna: All right, so should-cost is a powerful tool, but it requires data, collaboration, and a willingness to be wrong. Rachel's 17 percent cut is a great case study, but the real lesson is the process — not the number. Lucas: Right. The number is a byproduct. The process builds a more intelligent, transparent supply chain. And that's something worth investing in.