Supply chain KPIs: SCOR metrics, formulas and the systems behind them

The standard list of supply chain KPIs is ASCM's SCOR model. Version 14.0 groups eight performance attributes into three categories and gives each attribute at least one level-1 metric: perfect order fulfillment, order fulfillment cycle time, supply chain agility, total supply chain management cost, cost of goods sold, cash-to-cash cycle time, return on working capital and others. Run those alongside operational measures such as OTIF, fill rate, inventory turns and forecast accuracy. Pick one or two per attribute, write down the exact formula and data source for each, and review service and inventory together, because improving one usually costs the other.

SCOR's eight attributes and level-1 metrics

ASCM's SCOR Digital Standard introduction (SCOR Version 14.0, CC BY-NC-ND 4.0, © 2025) lists these. Reliability, responsiveness and agility are "customer (resilience) focused." Costs, profit and assets are "internally (economically) focused." Environmental and social are "outward (sustainability) focused."

Category	Attribute	Level-1 metrics
Resilience	Reliability (RL)	Perfect Order Fulfillment (RL.1.1), Perfect Supplier Order (RL.1.2), Perfect Return Order Fulfillment (RL.1.3)
Resilience	Responsiveness (RS)	Order Fulfillment Cycle Time (RS.1.1)
Resilience	Agility (AG)	Supply Chain Agility (AG.1.1)
Economic	Costs (CO)	Total Supply Chain Management Costs (CO.1.1), Cost of Goods Sold (CO.1.2)
Economic	Profit (PR)	EBIT as a Percent of Revenue (PR.1.1), Effective Tax Rate (PR.1.2)
Economic	Assets (AM)	Cash-to-Cash Cycle Time (AM.1.1), Return on Fixed Assets (AM.1.2), Return on Working Capital (AM.1.3)
Sustainability	Environmental (EV)	Materials Used, Energy Consumed, Water Consumed, GHG Emissions, Waste Generation (EV.1.1 to EV.1.5)
Sustainability	Social (SC)	Diversity and Inclusion (SC.1.1), Wage Level (SC.1.2), Training (SC.1.3)

The levels are diagnostic. ASCM's introduction says level-2 metrics "serve as diagnostics for level-1 metrics," and level-3 metrics do the same for level 2. A drop in perfect order fulfillment is explained by its level-2 components, such as Customer Order Perfect Condition (RL.2.4).

The full metric definitions and calculations sit in the SCOR DS itself. The introduction says it is "open access with authenticated ASCM login," and registration on ascm.org is free.

What changed from SCOR 13

If your KPI deck still lists five attributes, it was built on SCOR 13 or earlier. The SCOR 13 introduction (© 2020) had five: reliability, responsiveness, agility, cost and asset management. It listed one reliability metric (RL.1.1) and three agility metrics: Upside Supply Chain Adaptability, Downside Supply Chain Adaptability and Overall Value at Risk.

ASCM released the new SCOR DS on September 19, 2022. Its press release calls it "the most significant update since the inception of SCOR in 1996." Version 14.0:

• adds profit, environmental and social as attributes, and regroups the rest under resilience and economic
• replaces the three agility metrics with one, Supply Chain Agility (AG.1.1)
• adds supplier and return orders to the reliability level-1 metrics
• renames asset management to assets, keeping the AM code
• replaces Plan, Source, Make, Deliver, Return and Enable with Orchestrate, Plan, Order, Source, Transform, Fulfill and Return

The process change matters for KPIs. Deliver is now split into Order and Fulfill, so order-capture measures and shipping measures have separate homes.

The KPI table

The formulas below are common working forms. Each one has variants, which is the reason for the governance section further down.

KPI	Formula	Data source system	What it misses
Perfect order fulfillment (RL.1.1)	Orders delivered complete, on time, undamaged and with correct documents ÷ total orders	ERP orders and invoices, TMS proof of delivery, customer claims	Which component failed. Track the components separately as well.
Order fulfillment cycle time (RS.1.1)	Average days from order receipt to delivery	ERP order timestamps, TMS delivery events	Variability. A good average can hide a long tail of late orders.
Total supply chain management cost (CO.1.1)	Sum of the cost of running plan, order, source, transform, fulfill and return	General ledger, cost center mapping	Service. Cutting cost by holding less stock shows up later as missed orders.
Cash-to-cash cycle time (AM.1.1)	DSO + DIO − DPO	Balance sheet and income statement	How payables were stretched, and what that did to suppliers
Return on working capital (AM.1.3)	Supply chain contribution ÷ (inventory + receivables − payables)	General ledger	Fixed assets, which have their own metric (AM.1.2)
OTIF	Orders (or lines, or cases) delivered inside the agreed window and in full ÷ total	ERP orders, TMS delivery events, customer receiving data	Damage and documentation errors, which perfect order catches
Fill rate	Units (or lines, or orders) shipped from stock on the first attempt ÷ units ordered	ERP or order management, WMS shipments	Lateness. A complete shipment can still arrive late.
Inventory turns	Cost of goods sold ÷ average inventory value	General ledger, ERP inventory valuation	Mix. Fast movers can mask obsolete stock.
Days of supply	On-hand inventory ÷ average daily demand (forecast or actual)	ERP or WMS on-hand, planning system demand	Which demand figure was used. Forecast-based and history-based results differ.
Forecast accuracy (MAPE)	Mean of the absolute value of (actual − forecast) ÷ actual × 100	Planning system forecasts, ERP sales history	Direction of error. See bias.
Forecast bias	Sum of (forecast − actual) ÷ sum of actual	Planning system, ERP sales history	Size of error. Large errors can cancel to zero bias.
Freight cost per unit	Freight spend ÷ units, weight, or orders shipped	TMS freight audit and pay, general ledger	Service level bought with that spend, and expedites booked outside the TMS
Supplier on-time delivery	Receipts arriving inside the agreed window ÷ total receipts	ERP purchase orders, WMS receiving	Quantity and quality of what arrived
Dock-to-stock time	Time from arrival at the dock to inventory available to pick	WMS receipt and put-away timestamps, yard system	Time the truck waited in the yard before the dock, if arrival is not logged

Cash-to-cash from a 10-K: Procter & Gamble, fiscal 2026

Cash-to-cash cycle time is the number of days between paying for inputs and collecting cash from customers. The three inputs come from public filings, so you can calculate it for any listed company.

• DSO = average accounts receivable ÷ net sales × 365
• DIO = average inventory ÷ cost of goods sold × 365
• DPO = average accounts payable ÷ cost of goods sold × 365

Figures are from Procter & Gamble's Form 10-K for the fiscal year ended June 30, 2026, filed August 4, 2026. Amounts are in millions of dollars. Averages use the June 30, 2026 and June 30, 2025 balances.

Input	June 30, 2026	June 30, 2025	Average	Days
Accounts receivable	6,056	6,185	6,120.5	DSO 26 (on net sales of 87,032)
Total inventories	8,170	7,551	7,860.5	DIO 66 (on cost of products sold of 43,362)
Accounts payable	16,306	15,227	15,766.5	DPO 133 (on cost of products sold of 43,362)

Cash-to-cash = 26 + 66 − 133 = −41 days. P&G collects from customers about 41 days before it pays the suppliers whose payables sit on its balance sheet. Inventory turns on the same figures are 43,362 ÷ 7,860.5 = 5.5.

The filing also shows what the number hides:

• Supplier financing. Note 14 says P&G negotiates "extended payment terms with its suppliers" and runs a supply chain finance program with banks. Payment terms under that program "generally range from 60 to 180 days." Confirmed obligations under the program were $6,176 million at year end, about 38% of accounts payable. A long DPO here partly reflects a financing arrangement, not only operating practice.
• What is in payables. The cash flow discussion says payables rose partly from "increased marketing support activity." Dividing all payables by cost of goods sold mixes non-supply spend into DPO.
• What is in COGS. P&G's cost of products sold includes distribution, inbound freight, warehousing and other shipping costs. Another company may book some of those elsewhere. Compare cash-to-cash across companies only after checking the cost lines.

Your internal version should use daily or monthly balances from the ERP, not two year-end points.

Forecast accuracy needs MAPE and bias together

MAPE is the mean absolute percentage error. Hyndman and Athanasopoulos, in Forecasting: Principles and Practice (3rd edition, section 5.8), list its problems. It is "infinite or undefined" when actual demand is zero for any period. It produces extreme values when actuals are close to zero. It puts "a heavier penalty on negative errors than on positive errors."

Zero-demand periods are common for slow-moving SKUs, spare parts and new items. For those, MAPE either breaks or swamps the average. The same section describes scaled errors, proposed by Hyndman and Koehler in 2006, as an alternative for comparing accuracy across series with different units.

Bias measures direction. A forecast that is 20% high one month and 20% low the next has zero bias and poor accuracy. A forecast that is 5% high every month has good accuracy and steady bias, and that bias builds inventory. Report both, at the level where the forecast drives a decision: usually SKU by location, at the lead time of the replenishment order.

OTIF means what your customer's contract says

OTIF has no single standard definition. Retailers set their own terms in supplier agreements and compliance manuals. Check four points in each customer's terms:

• Unit of count. Orders, lines or cases. A 10-line order with one short line is 0% in full by order and 90% by line.
• Window. A must-arrive-by date, an appointment time or a range of days, and whether early arrival counts as a failure.
• Whose record. The customer's receiving data or your shipping data.
• Freight terms. Whether the supplier or the retailer arranges transport, and whether the on-time test differs between the two.

If you sell to several large retailers, you have several OTIF figures. Keep each customer's definition as a separate metric, and keep your own internal OTIF definition for trend reporting.

Pick five to eight, at least one per attribute

A KPI set that covers only cost will push service down. One that covers only service will push inventory up. Use SCOR's attributes as a coverage check:

1. Reliability: perfect order fulfillment, or OTIF if your customers score you on it.
2. Responsiveness: order fulfillment cycle time.
3. Agility: SCOR's AG.1.1, or a scenario measure such as how much demand change you can absorb within a set number of days.
4. Cost: total supply chain management cost as a percent of revenue, with freight cost per unit beneath it.
5. Assets: cash-to-cash cycle time or inventory turns.
6. Planning input: forecast accuracy and bias.

Add environmental or social metrics if the company reports them externally. Put fill rate, supplier on-time delivery and dock-to-stock time in level-2 and level-3 reports, where people use them to explain movement in the level-1 set.

Service and inventory pull against each other

Raising fill rate usually means holding more safety stock, which lowers turns and lengthens cash-to-cash. P&G's fiscal 2026 10-K shows the trade in one sentence: inventories increased, "consuming $641 million of cash, driven primarily by increased safety stock levels and new product initiatives. Days inventory on hand increased by two days."

Common conflicts:

• Fill rate against inventory turns. Fix the service target per customer or segment first, then measure inventory against it.
• Freight cost per unit against order fulfillment cycle time. Full truckloads cost less per unit and take longer to fill.
• DPO against supplier on-time delivery. Longer payment terms improve cash-to-cash and can weaken a supplier's ability to deliver.
• Forecast bias against fill rate. A planner rewarded on fill rate has a reason to forecast high.

Review conflicting pairs in the same meeting, owned by the same executive. Set the trade-off explicitly, for example a line fill rate target for one customer segment paired with a days-of-supply ceiling. A target for one metric without its partner moves the problem.

Where the data comes from

System	Supplies
ERP	Sales orders, purchase orders, invoices, inventory valuation, general ledger cost and balance sheet data
Warehouse management system (WMS)	Receipts, put-away, picks, shipments, on-hand by location, dock-to-stock timestamps
Transportation management system (TMS)	Shipments, carrier events, proof of delivery, freight invoices
Planning system	Forecasts by version and lag, safety stock settings, supply plans
Customer portals and EDI	Customer receiving data and compliance scores for OTIF

Most KPIs need records joined across systems. Perfect order needs the ERP order, the TMS delivery event, the WMS shipment and the customer's claim to agree on one order number. Forecast accuracy needs the forecast version that was frozen at lead time, not the latest one.

Write the definition before you build the dashboard

Each KPI needs a one-page record before it reaches an executive report:

• the formula, including the unit of count (order, line, unit, case)
• the time window and the timestamp that starts and stops the clock
• the system of record for each input
• exclusions, such as customer-requested delays or canceled orders
• the owner, who approves any change
• the date the definition last changed

Store the definitions in the semantic layer of your BI tool, so every report computes the KPI the same way. When a definition changes, restate history or mark the break on the chart. A two-point improvement in OTIF that came from changing the delivery window is not an operational improvement.

Frequently asked questions

What are the SCOR level-1 metrics? In SCOR Version 14.0 they are perfect order fulfillment, perfect supplier order, perfect return order fulfillment, order fulfillment cycle time, supply chain agility, total supply chain management costs, cost of goods sold, EBIT as a percent of revenue, effective tax rate, cash-to-cash cycle time, return on fixed assets, return on working capital, five environmental metrics and three social metrics. The SCOR DS introduction lists them.

What is the difference between OTIF and perfect order? OTIF checks timing and quantity. Perfect order also requires the goods to arrive undamaged and with correct documentation, so it is the stricter measure.

How do you calculate cash-to-cash cycle time? Days sales outstanding plus days inventory outstanding minus days payables outstanding. Procter & Gamble's fiscal 2026 figures give 26 + 66 − 133 = −41 days.

Related Reading

• Supply Chain Management (SCM)
• Inventory Optimization & Replenishment
• Transportation Management Systems (TMS)
• Warehouse Management Systems (WMS)
• Business Intelligence & Analytics
• Supply Chain Technology Strategy
• Supply chain transformation consulting: the transformation process, the firms that sell it, and what to ask
• Measuring What Matters: KPIs for Strategic CIOs