Latest / The Buyer's Side with Fexingo: Procurement, Sourcing, and B2B Purchasing Decisions / How One Manufacturer Reduced Supplier Defects by 80 Percent
Transcript
- Lucas: There's a mid-sized automotive parts manufacturer in Ohio — I'll call them Midwest Precision, not their real name — that had a problem. They were rejecting about 18 percent of incoming shipments from their top five suppliers. Luna: That's huge. Eighteen percent is basically one in five shipments has something wrong. Lucas: Right. And this wasn't a new operation. They'd been buying from these suppliers for years. The defects weren't catastrophic — surface scratches on machined parts, inconsistent hardness in a batch, minor dimensional drift. But the cumulative cost of rework and downtime was eating their margins. Luna: So what did they do? Threaten to switch suppliers? Lucas: That's the obvious move, but no. Their procurement lead told me they realised something: if they switched, the new supplier would probably have similar issues because the root cause wasn't malice or incompetence. It was a mismatch between how they inspected and how the supplier's process varied. Luna: Okay, so they looked inward first. Lucas: Exactly. They started with a Pareto analysis of every defect they'd logged over the prior six months. Three defect types accounted for 73 percent of all rejections — surface finish anomalies, hardness outliers, and a specific tolerance drift on one part family. Luna: So they could focus on the biggest problems first. Lucas: Yes. But here's where it gets interesting. Instead of just sending a complaint report to suppliers, they redesigned their incoming inspection process. They used to do a random sample of ten pieces per lot — which is a pretty standard AQL approach. But that sample size wasn't catching intermittent defects. So they shifted to a stratified sampling plan: they'd take a fixed number of pieces from each production run within a lot rather than blindly grabbing ten from the top of the pallet. Luna: That's smart — because if a defect is concentrated in one part of the production run, you'd miss it with random sampling. Lucas: Right. That change alone doubled their defect detection rate within a month. But they didn't stop there. The real breakthrough came when they built a shared defect dashboard — live, updated every shift — and gave their suppliers direct read-only access. Luna: So the suppliers could see the data as soon as it came in? Lucas: Yes. And not just aggregate numbers — they could drill into each rejected piece: the measurement, the deviation, even a photo from the inspection station. The procurement lead told me that for the first time, suppliers could see exactly what 'out of spec' meant, in their own language. Within two weeks, one supplier had adjusted their grinding process based on the photo evidence. Luna: That's a huge shift in trust. Most companies would never share inspection data that granular — they'd treat it as proprietary. Lucas: Exactly. And Midwest Precision had to overcome internal objections too. The quality team was worried suppliers would use the data to argue about borderline cases. But they decided that the transparency was worth the risk. And it paid off. Luna: What were the numbers at the end of the year? Lucas: Over twelve months, their incoming defect rate dropped from 18 percent to under 4 percent. That's roughly an 80 percent reduction in defects. And they did it without switching a single supplier and without a price increase. Luna: Eighty percent — that's a massive improvement. What was the cost of implementing the dashboard? Lucas: Relatively small. They used a low-code platform — think Airtable or similar — to build the dashboard with a simple API connection to their existing inspection equipment. Total setup cost was under fifteen thousand dollars, plus maybe two hours of training per supplier. Luna: So the ROI was probably enormous. Fifteen thousand dollars versus months of rework — that pays for itself in a week. Lucas: Easily. The procurement lead estimated that the rework costs alone were dropping by about forty thousand dollars per quarter. And that doesn't include the intangible benefits: less production downtime, fewer expedited shipments, less tension in supplier relationships. Luna: It's interesting that the biggest win wasn't a new technology or a new supplier. It was changing how they shared information. Lucas: That's the part I keep coming back to. Most quality improvement programs focus on the supplier side — demanding corrective actions, threatening penalties. But this case suggests that the inspection process itself — and the data-sharing layer — can be the bottleneck. Luna: And it's a bottleneck you can fix without adversarial negotiations. Lucas: Right. There's a behavioral economics angle too. When suppliers could see their own defect data in real time, they started self-correcting before Midwest Precision even flagged it. One supplier noticed a drift in hardness values over three consecutive shifts and called to say they'd caught it and were adjusting the tempering cycle. Luna: That's remarkable. The transparency created a sense of ownership. Lucas: Exactly. And I think that's a lesson that scales beyond automotive parts. Any procurement team buying repeat batches of anything — raw materials, subassemblies, even services — can apply this: start with good data about what's actually failing, then share that data with suppliers in a way they can act on. Luna: You mentioned services — how would you apply a defect dashboard to something like a cleaning contract or an IT outsourcing agreement? Lucas: Good question. The principles are the same: define what a defect is — a missed service level, a delayed ticket resolution, a compliance gap — measure it consistently, and share the trend data with the provider. The key is to make the data granular enough that the provider can identify the root cause. If you just show them a monthly score of 87 percent, they can't improve. But if you show them that response times were above target on three specific days, they can investigate. Luna: So it's about turning a scorecard into a diagnostic tool. Lucas: Exactly. And that's what Midwest Precision did. Their dashboard wasn't just a score — it showed the actual measurements, the photos, the trends. The suppliers could see, for example, that surface finish defects were spiking on Tuesdays, which led them to realise they were changing cutting fluid on Tuesday mornings and not letting the system stabilize. Luna: A classic root cause that would never show up in a monthly report. Lucas: Right. And this is why I think the Pareto analysis step is so important. If you just throw data at suppliers without focusing on the biggest few defects, you overwhelm them. But if you say, 'Hey, 73 percent of our rejections come from these three things, and here's exactly what we're seeing,' you give them a clear starting point. Luna: So for a procurement team listening to this, what's the first step they should take on Monday morning? Lucas: Pull your last six months of rejection data — if you don't have it structured, start tracking it today. Run a simple Pareto analysis: which defect types account for the top 80 percent of rejections? Then pick the top two or three and investigate whether your inspection process is actually catching them consistently. You might find, like Midwest Precision did, that your sampling method is the problem. Luna: And if you don't have inspection data at all? If you're a smaller team that's been relying on occasional audits? Lucas: Start small. Pick one high-volume SKU and one supplier. Set up a basic inspection protocol — even just measuring three critical dimensions on every batch — and share that data weekly. The goal isn't perfection; it's creating a feedback loop. Once that works, scale it to the next SKU. Luna: That feels very doable. Lucas: It is. And it's something we talk about on this show a lot — small, focused changes that have outsize impact. Luna: Speaking of which — if these conversations have moved your work forward in some small way, that's exactly why we keep doing them. We keep this show ad-free and open to everyone, and a couple of dollars a month is genuinely what keeps these going — buy me a coffee dot com slash fexingo, if you've gotten something out of them. Lucas: Yeah, it's as simple as that. Every bit of support lets us spend more time digging into cases like this one. Luna: Alright, back to the data. One thing I want to circle back to: you mentioned that Midwest Precision didn't switch suppliers. Do you think that's always the right call? Lucas: No, not always. If a supplier is consistently out of spec across multiple defect types and shows no willingness to improve, then switching might be the only option. But I think many procurement teams switch too quickly, assuming a new supplier will be better, when the real problem is a broken information flow. Luna: So first, try to fix the information flow. Lucas: Yes. And that's much cheaper and faster than qualifying a new supplier. Midwest Precision's approach cost them fifteen thousand dollars and a few weeks. Qualifying a new automotive supplier can take six months and tens of thousands in audits and samples. Luna: That's a compelling argument for transparency. Lucas: It is. And I think it applies beyond procurement. Any relationship with a recurring transaction — whether it's a landlord, a contractor, a software vendor — benefits from clear, shared data about what's working and what's not. Luna: So the takeaway: stop guessing, start measuring, and share the measurements. Lucas: Exactly. And be specific about what you share. Don't just give them a score — give them the context they need to improve. Luna: Good episode. Next time, I'd love to talk about how procurement teams are handling the shift toward nearshoring — that seems to be a big topic right now. Lucas: It's on the list. We'll dig into a case from a medical device company that moved production from Southeast Asia to Mexico and what they learned about quality and lead time trade-offs. Luna: Looking forward to it.