5 KPIs Every Supply Chain Head Should Track
Five measures that change decisions rather than describe them, why each has to be defined before it is quoted, and how they mislead when read alone.
Most supply chains drown in dashboards but starve for decisions. The fix is not more metrics — it's the right handful, tracked consistently and tied to action. These five KPIs give a supply chain head an honest read on service, cost, and working capital.
1. On-Time In-Full (OTIF)
OTIF is the percentage of orders delivered both on schedule and complete. It is the single best proxy for whether your network is actually serving customers. Best-in-class operations run 95–98%; anything below 90% signals real service issues and, in retail or B2B, potential penalties.
2. Inventory accuracy
If your system says 100 units and the shelf has 94, every downstream plan is wrong. High inventory accuracy — sustained through disciplined cycle counting — reduces stock-outs, lowers carrying cost, and makes inbound receiving and order processing far more reliable.
3. Perfect order rate
The perfect order rate combines on-time, in-full, damage-free, and correctly documented. It is stricter than OTIF and exposes problems in picking, packing, and paperwork that single metrics hide. Treat it as the customer-experience benchmark.
4. Inventory turnover and carrying cost
Turnover shows how quickly stock moves; carrying cost quantifies the working capital tied up in it (storage, insurance, shrinkage, obsolescence). Read together, they reveal whether inventory is productive or quietly eroding margin.
5. Cash-to-cash cycle
This measures the days between paying suppliers and collecting from customers. It connects logistics performance directly to the balance sheet — shorter cycles mean leaner working capital and a more resilient business.
Why both parties report different numbers for the same month
The most common dispute in a logistics relationship is not about performance. It is about definition, and both sides are usually reporting accurately against the definition they hold.
On-time delivery is the standard example. Does the clock start when the order is placed, when it is released to the warehouse, or when it is picked? Is performance measured against the customer's requested date or the date that was confirmed back? Are delays caused by the customer excluded, and who decides whether a given delay qualifies? Each of those choices is defensible, and each produces a materially different number from identical operations.
Inventory accuracy behaves the same way. Measured by location, it asks whether the system knows what is in each bin. Measured by SKU, it asks whether total quantity per item is right, which can be true while every location is wrong. Measured by value, it weights toward expensive stock and can look excellent while a high-runner component is unreliable.
This is why a defined 97 per cent is worth more than an undefined 99, and why the definitions belong in the agreement rather than in the first argument.
- Write down when the clock starts and stops for every time-based measure.
- State whether customer-caused delays are excluded, and who adjudicates.
- For accuracy, specify by location, by SKU or by value - and whether tolerance applies.
Measures that predict, not just describe
Most logistics reporting is retrospective. It tells you accurately what happened last month, which is useful for accountability and close to useless for prevention, because by the time the number moves the cause is weeks old.
A smaller set of measures gives warning. Ageing of anything stalled in a process - receipts not put away, returns not dispositioned, exceptions not closed - predicts next month's accuracy problem, because unresolved work becomes a discrepancy eventually. Rework counted by type points upstream to the decision that made it possible. Supplier readiness against collection windows predicts inbound shortages before they reach the line.
The distinction worth applying to any dashboard is whether a plausible value would change what someone does that day. Measures that would not belong in a monthly pack; measures that would belong on the screen someone checks before a shift.
More Insights
Why Reverse Logistics Is Now a Competitive Advantage
Returns are usually an administrative afterthought. Treating disposition as a timed decision rather than a queue is where the recoverable value sits.
VMI vs Consignment Inventory: Which Model Fits Your Network?
VMI and consignment are often treated as interchangeable. They differ on the point finance cares about: who owns the stock, and until when.
In-Plant Logistics: The Hidden Lever for Production Efficiency
In-plant logistics is measured at the production station, not the warehouse dock - a different discipline even when the same team does both jobs.
Put these ideas to work
Talk to a Vistar logistics expert about applying this to your operations.