Supply chain KPIs

Supply chain performance is usually judged on three fronts at once: service to customers, cost, and cash tied up. The trouble is that they pull against each other: more stock improves service but costs cash, and fuller trucks cut cost but can slow delivery. That’s why KPIs must be read together, never alone.

The ones that come up most:

KPI	What it measures	Formula or meaning
OTIF	Orders delivered on time and in full	On-time-in-full orders ÷ total orders
Fill rate	Share of demand met from stock	Units shipped from stock ÷ units ordered
Perfect order	On time, complete, undamaged, correct documents	Product of the component rates
Inventory turns	How fast stock moves	Cost of goods sold ÷ average inventory
Days of inventory (DIO)	How long stock sits	365 ÷ turns, or inventory ÷ daily COGS
Forecast accuracy and bias	Quality of the demand plan	WMAPE or MASE, plus mean error
Logistics cost	Cost efficiency	Logistics cost ÷ revenue, or cost per unit
Cash-to-cash cycle	Working capital tied up	DIO + DSO − DPO
OEE	Production effectiveness	Availability × performance × quality

A company turning inventory 8 times a year holds about 46 days of stock (365 ÷ 8). Push turns to 12 and stock drops to about 30 days, freeing cash, but only if OTIF holds.

Frameworks like the SCOR model group measures into reliability, responsiveness, agility, cost and asset efficiency. A balanced set from each group stops a team from improving one number by damaging another, such as cutting inventory until service collapses.

From KPIs to return on capital

Senior leaders judge the business on return on capital employed (ROCE): operating profit divided by the capital tied up. Supply chain performance moves both parts:

• Profit: cost of goods sold, logistics cost, obsolescence, expediting, and revenue lost to stockouts.
• Capital: inventory, receivables minus payables, and fixed assets such as plants, warehouses and trucks.

Revenue ₹500 crore, operating profit ₹50 crore, capital employed ₹250 crore: ROCE is 20%.

• Logistics savings of ₹2 crore lift profit to ₹52 crore.
• Removing ₹10 crore of excess inventory cuts capital to ₹240 crore, and avoids about ₹1 crore a year of storage and obsolescence cost.
• ROCE becomes 53 / 240 ≈ 22.1%.

A two-point rise in ROCE from supply chain work alone, without any growth in sales.

Define KPIs precisely

The same name can hide different calculations. A KPI definition should state:

• Formula and units: OTIF by order, line or unit?
• Reference point: on time to the customer’s requested date or the promised date? At dispatch or at delivery?
• Scope: which customers, products, sites and channels.
• Data source and timing: where the data comes from, how often it is calculated, and when.
• Owner and target.

Without this, two sites report the same KPI with different meanings, and comparisons mislead.

Leading and lagging indicators

Lagging indicators (OTIF, inventory turns, cost per unit) report results. Leading indicators warn of trouble earlier: forecast bias, projected stock below safety stock, supplier acknowledgement delays, backlog of open orders. A good dashboard pairs each important result with the signals that predict it.

Balance, not a single number

Kaplan and Norton’s balanced scorecard (1992) argued that no single measure can guide an organization; financial results must be read with customer, process and learning measures. For supply chains, the balance is service, cost, cash and, increasingly, resilience and sustainability. SCOR organizes its top-level metrics similarly, across reliability, responsiveness, agility, cost and asset efficiency.

Goodhart’s law

“When a measure becomes a target, it ceases to be a good measure.” A forecast accuracy target invites sandbagging; an inventory target at year end invites delayed receipts; an OTIF target measured against promised dates invites late promises. Designing KPIs means anticipating how people will respond to them and pairing measures so that gaming one shows up in another.

Functional KPIs that fight each other

• Purchase price variance rewards buyers for big discounted buys that raise inventory and obsolescence.
• Plant efficiency rewards long runs that build stock and delay other products.
• Transport cost per kilometre rewards cheap carriers who deliver late.
• Sales revenue at quarter end rewards loading the channel with stock.

Each looks sensible locally. Shared end-to-end measures, such as total landed cost, OTIF to the customer and cash conversion, align the functions.

Benchmarking with care

External benchmarks are useful only when definitions, business models and product mixes are comparable. Inventory turns differ widely between grocery and machinery; OTIF differs by how it is measured. Benchmarks are best used to ask questions (“why are our turns lower than similar companies’?”), not to set targets mechanically.

Incentives

Bonuses tied to a single supply chain KPI tend to distort decisions. Combinations of team-level and end-to-end measures, with a few individual measures that the person can actually influence, work better, as does reviewing how results were achieved, not just whether targets were hit.