Service Level vs. Cost: The Trade-Off Teams Keep Getting Wrong
Every supply chain lives on a single tension: how much do you spend to make sure the customer gets what they ordered, on time? Push service level toward perfection and inventory, expediting, and capacity costs balloon. Squeeze cost too hard and you stock out, lose sales, and burn goodwill. The teams that struggle are rarely bad at math; they just make the trade-off by accident instead of on purpose.
That accident gets expensive in 2022. Lead times are long and erratic, materials are scarce, and carrying costs are rising with inflation. The old rules of thumb, set when freight was cheap and reliable, quietly stopped being true. Here are the mistakes I see teams make as they try to rebalance, and how to avoid them.
Treating every product the same
The single biggest error is applying one service-level target across the whole catalogue. A 98% target on a slow-moving, high-margin specialty item and on a cheap, fast-moving commodity makes no sense; the cost of holding stock and the cost of a stockout differ wildly between them. Segment your items first, by demand volume, variability, margin, and how much a stockout actually hurts, then set service levels per segment. A simple ABC or service-criticality split does more good than any clever formula applied uniformly.
Confusing service level with fill rate. Cycle service level (the chance an order cycle ends without a stockout) and fill rate (the fraction of demand met from stock) are different numbers. Pick the one that matches the promise you actually make to customers, and measure it consistently.
Sizing safety stock off stale assumptions. Safety stock depends on demand variability AND lead-time variability. When lead times became long and unpredictable in 2021–2022, the variability term exploded; teams that kept their old safety-stock settings were quietly under-protected.
Ignoring the real cost of a stockout. A stockout on a part that halts a construction site costs far more than the lost margin on the part. If you price the stockout only at lost gross margin, you will systematically under-serve your most critical items.
Optimizing one number while breaking another
Cost and service are not the only two variables, and treating them as if they are leads to local wins that hurt the whole. Cutting inventory to hit a working-capital target can wreck on-time delivery; chasing a service number can bury cash in slow stock that later gets written off. Look at the system: service level, inventory investment, expediting spend, and obsolescence risk together. A balanced scorecard beats a single optimized metric every time.
Segment items and set a service target appropriate to each segment.
Recalculate safety stock with current, not historical, lead-time variability.
Cost a stockout at its true downstream impact, not just lost margin.
Review the full picture, cost, service, expediting, and obsolescence, not one number alone.
Make the choice on purpose
The goal is not to maximize service or minimize cost; it is to choose the balance that fits your strategy and to know what that choice costs. Write down the target service level for each segment, the inventory it implies, and the assumptions behind it. When conditions shift, and in this market they shift often, you can revisit a documented decision. You cannot revisit one that was never actually made.
The teams that come through volatility well are not the ones with the highest service or the leanest inventory. They are the ones who decided, deliberately and in writing, what level of risk they were willing to carry, and revisited it when the ground moved.
If you are rebalancing service against cost under today's supply pressures, XNM's procurement, sourcing & contract management can help you set targets and supplier terms that fit your real risk.