Supply Chain Service KPIs Use Case
When Growing Pains Become Customer Complaints
A precision custom metal stamping and machined components manufacturer in Greenfield, Indiana, has been family-owned for nearly six decades. They have add-on plants in Columbus, Ohio, and Marietta, Georgia, completed in 18 months. A third acquisition is under due diligence.
"Since we brought Columbus online, something changed. I'm getting escalations I wasn't getting before — not just from new customers, but from accounts we've had for years. The complaints didn't start until the third plant came in."
That was Greg, the Sales VP. He wasn't focused on a metric. He just knew some process was broken.
What the service metrics said — and left out
Perfect Order Rate | OTIF | Forecast Accuracy & Bias | Customer Reject Rate
The COO pulled what he had:
- OTD (on-time delivery) running around 87%.
- Customer Reject Rate holding around 1.5%.
Reasonable numbers, but incomplete ones. Consolidated OTIF across Greenfield, Columbus, and Marietta didn't exist. When someone pulled the numbers together informally, true OTIF looked closer to 70% — two of the three plants had been crediting on-time delivery for partial shipments. Perfect Order Rate wasn't tracked. Forecast Accuracy and Bias weren't measured, but the team knew it was a problem. The business had lagging indicators — metrics that confirmed a customer was already unhappy — but nothing that identified the cause or provided triggers to take action.
The Columbus plant was expected to add capacity and flexibility, but it also added complexity the service model wasn't designed to handle. Two plants had been manageable through informal coordination. Three plants — each on a different ERP, each with its own scheduling logic and no shared demand signal — created delivery failures before anyone had visibility into them internally.
A customer service team without a complete picture
A Tier 1 datacenter contractor's qualification process made the service gap specific and urgent. Datacenter buildouts operate on compressed construction schedules where a late component delivery doesn't generate a complaint — it triggers contractual penalties that dwarf the value of the parts involved. The customer needed demonstrated OTIF performance, a credible capacity ramp plan, and evidence that delivery consistency could be maintained through a volume increase.
None of that existed in presentable form. OTIF had never been calculated across all three plants. A customer service team that couldn't quote delivery dates confidently on existing products was not in a position to commit to a ramp schedule on a new product family for a customer with no tolerance for uncertainty.
Datacenter customer requirements raised the bar
The customer service function had been consolidated across all three plants after the acquisitions. The expectation was that a single team could manage commitments and exceptions across Greenfield, Columbus, and Marietta — plants that planned differently, ran different systems, and had never aligned on what an accurate delivery promise meant in practice. It wasn't a reasonable expectation, and the team knew it.
When customers called for commit dates or explanations for why deliveries had been pushed out, customer service couldn't provide confident answers. The response became cautious — hedged commitments, delayed explanations, avoided conversations. Customers weren't just receiving late deliveries. They were receiving silence. In a B2B relationship built on operational reliability, silence erodes trust faster than a late shipment with a straight explanation.
Track the right supply chain KPIs
Supply chain excellence is not confined to warehouses and logistics—it radiates outward, transforming every corner of the enterprise. When supply chain operates at its peak, the entire organization elevates.
Get the executive guide to the top supply chain KPIs, including formulas.
Supply chain is not a department—it is a discipline that permeates every function, influences every decision, and determines every outcome. To master supply chain is to master business itself.
An operational gap, not just a metrics gap
The metrics were absent, but the organizational structure was the root cause. No one owned Service performance across all three plants. Greg owned customer relationships. The COO owned plant operations. Neither owned a shared definition of what "on time" and "complete" meant as a cross-plant standard.
Without that definition, plant-level performance couldn't be consolidated, customer service couldn't make credible commitments, and Greg had no way to know whether the pattern of escalations he was seeing was a plant problem, a scheduling problem, or something systemic.
Technology considerations
Each plant managed order status independently, with no visibility across sites. Sales forecasts moved between the sales team and plant schedulers weekly, by spreadsheet. The immediate requirement wasn't a technology platform — it was clear forecast ownership and a defined S&OP cadence that Greenfield, Columbus, and Marietta all operated against.
AI-driven demand sensing and probabilistic forecasting become the right investment once that data infrastructure exists — sharpening responsiveness to shifting demand and bringing forward-looking market data into the planning process rather than relying on trailing history alone. Those capabilities matter more as the company adds larger customers and has to manage component supply across three plants instead of one.
What moved
OTIF was standardized across all three plants, finally meaning the same thing at Greenfield, Columbus, and Marietta and closing the gap that had let partial shipments count as on-time. It improved from 70% to 80% as the definition tightened and the underlying causes became visible for the first time. Perfect Order Rate entered the weekly operating review. Forecast Accuracy and Bias were tracked for the first time, giving the team visibility into a problem they had previously only sensed. The customer service team had a consolidated cross-plant view for the first time and could make and keep delivery commitments with confidence.
Greg's phone was ringing less. That's not a metric. But it's a reliable signal that the right metrics are working.