Report: Why Spare Parts Inventories and Customer Escalations Keep Rising Together
Spare parts inventories at manufacturers are sitting at record levels, and customer escalations are climbing right alongside them, according to research published in June by Unilog. Conventional inventory logic says those two lines should move in opposite directions, and the research argues that their rising together points to a problem with the shape of service networks rather than the amount companies spend on them.
The stakes are bigger than the planning function that usually owns the problem. Manufacturers tie up roughly a tenth of annual sales in spare parts, after-sales satisfaction consistently runs 10 to 15% below customer expectations, and in many networks half of all service calls are delayed because the needed part isn’t where it should be, the research finds. And the aftermarket is no longer a sideline: BCG has estimated that service represents a third or more of total income for industrial leaders, while a McKinsey survey of 250 senior aftermarket procurement executives found 70% reported no service improvement from their providers over a full decade, with parts availability among the top complaints.
The paradox persists, the research argues, because the standard responses all make it worse. Cutting inventory to satisfy finance shifts the failure into customer downtime and next quarter’s escalation report. Adding inventory feels safer but isn’t, since a meaningful share of parts on the shelf goes obsolete each year, meaning volume poured into a badly shaped network buys depreciation rather than availability. Tightening the SLA without redesigning what sits behind it converts a contract clause into a liability paid out in service credits and exposed renewals. The common thread, per the Unilog analysis, is that networks are managed to a single aggregate availability number, which planners defend by overstocking cheap, fast-moving parts while the expensive, low-volume, mission-critical parts that decide whether a customer is up or down quietly run thin. The result is a network that looks optimized on paper and fails on the customer’s floor. The research captures it in one image: a ten-dollar fan can take down a quarter-million-dollar server rack, yet most networks stock the spare power supplies instead.
The structural insight is that a service network behaves like a chain rather than a hierarchy of warehouse tiers, with outcomes decided at the handoffs between links — global pool, regional hub, forward stocking location, technician’s van, consignment stock — each of which exists to do a different job. The practical tool that follows is segmentation on two axes, part criticality against demand predictability, to determine which parts belong on which link, with a named owner for each quadrant so the design doesn’t drift back into a single planner’s spreadsheet within a year.
The sharpest critique in the research is aimed at measurement. A network can hit every monthly average and still lose the renewal, because customers remember the morning it failed rather than the months it worked.
“The average is comfort. The worst day is consequence.” The worst incidents cluster predictably in slow-moving, high-criticality parts, which forecasting tools rate as weak demand signals and finance flags as easy cuts, which is how a routine budget review becomes the cause of next quarter’s escalation,” said Eyal Yossef, VP of supply chain solutions at Unilog. “The recommended fix starts with visibility: adding the worst 5% and worst 1% of the parts-wait clock to standing service reviews, figures most service organizations cannot produce today.”
Three anonymized case studies in the research show how networks fail while their dashboards stay green — a data center service provider whose 32 forward stocking locations filled tickets locally only one time in three, leaving an expedite bill larger than the inventory savings; a medtech manufacturer that duplicated expensive insurance-part kits at every regional hub when a single central set with a pre-arranged expedite lane would have covered the same risk; and a European OEM whose well-run Amsterdam hub could honor its same-day promise in the Benelux and almost nowhere else. Three verticals, three failure modes, one lesson: operational excellence cannot rescue a poorly designed network.
For operations leaders wondering where to start, the research argues against the full redesign and for two deliberately small moves next quarter: build the segmentation matrix for only the top three SLAs and top hundred SKUs, forcing operations, commercial, parts engineering, and finance to decide ownership in one room, and put the worst-day numbers into the monthly service review. The redesign takes longer than 90 days. The visibility that makes it possible doesn’t.
