Why measure at all
Logistics produces a steady stream of events: orders received, pallets moved, trucks loaded, parcels delivered. Without measures it is hard to tell whether the operation is getting better or worse, or where a problem sits. Key performance indicators, usually shortened to KPIs, turn that stream into a handful of numbers that people can track over time.
The word “key” matters. A business can calculate hundreds of figures, but a small set tied to real goals is more useful than a wall of charts.
Delivery and service measures
These measures describe what the customer experiences.
On-time delivery. The share of orders or shipments that arrive by the promised date or within an agreed window. If 92 of 100 orders arrive as promised, the rate is 92 percent (a hypothetical example). The tricky part is the definition: promised by whom, and measured at which point? A carrier may count a delivery as on time against its own schedule while the customer judges it against the date shown at checkout.
Order fill rate. The share of ordered items, or of orders, that can be supplied from stock when requested. It can be measured by lines, units or complete orders, and those give different answers. An order that is 9 of 10 lines filled looks good by line but counts as a miss if only complete orders count.
Order accuracy. The share of orders shipped with the right items, in the right quantities, to the right address. Errors are usually costly because they trigger returns, re-shipments and unhappy customers.
Damage and loss rate. The share of shipments that arrive damaged or go missing. A low rate is good, but it can depend heavily on how claims are recorded; if customers rarely report minor damage, the figure may look better than reality. Packaging choices and the way freight claims are handled both influence this number.
Time measures
Order cycle time. The time from order placement to delivery. It can be broken into steps, such as processing, picking, packing and transit, which helps locate the slow stage.
Dock-to-stock time. How long goods take to move from arriving at the receiving dock to being available for sale or use. Long delays here mean items exist but cannot be shipped.
Transit time. How long goods spend in transport. Averages can hide important detail, because a route that is usually quick but occasionally very slow can be harder to plan around than one that is steady.
Inventory measures
Inventory turnover. How many times stock is sold and replaced over a period. A common approach divides the cost of goods sold by the average inventory value. As a hypothetical example, if annual cost of goods sold is 600,000 and average inventory is 100,000, the turnover is 6. High turnover can mean efficient use of money, but it can also mean the business is running too lean and risks stockouts.
Days of supply. How long current stock would last at the present rate of use. If 300 units are on hand and 30 are used per day (hypothetical), that is 10 days.
Inventory accuracy. The match between records and what is physically present. Poor accuracy undermines every other number, since planning relies on records. Regular checks such as cycle counting exist for this reason.
Stockout rate. How often an item is requested but unavailable. It is related to inventory management as a whole, and the buffer stock held to prevent it is covered under safety stock.
Cost measures
Cost per unit shipped or per order. Total logistics spending divided by the number of units or orders. It is easy to compare over time, but it can mislead if the mix of products changes. Shipping heavy items one month and light ones the next will move the figure without any real change in efficiency.
Transportation cost as a share of sales. Freight spending divided by revenue. It is simple, but it can rise or fall because of price changes in the product rather than in logistics itself. How these figures are built up is explained further under logistics costs.
Warehouse cost per unit handled. Storage and handling expense divided by units processed. It is sensitive to volume: when volume drops, fixed costs are spread across fewer units and the figure rises.
Productivity and utilization
Capacity utilization. How much of available space or capacity is used, such as the percentage of a trailer’s space filled or the share of warehouse slots occupied. High utilization is usually efficient, but near-full facilities have little room to absorb surges, which matters during peak season.
Lines picked per hour. A measure of warehouse productivity. It rewards speed, so it should be read alongside accuracy; a team can pick quickly and wrongly.
The limits of any single number
A few cautions apply to nearly all logistics metrics.
- Averages hide spread. An average delivery time of three days can include some orders arriving in one day and others in seven.
- Metrics can conflict. Cutting cost per unit may lower service levels, and chasing high service may raise cost. Looking at one without the other gives a distorted picture.
- People respond to targets. When a number is used to judge performance, behavior may shift to improve the number rather than the underlying service. A warehouse rewarded for speed might cut corners on checking.
- Definitions drift. Two companies reporting “on-time delivery” may be measuring quite different things, so comparisons between firms need care.
- Data quality limits everything. A tidy percentage built on incomplete scan data is less reliable than it looks.
Choosing a sensible set
A balanced set usually includes at least one measure each for service, speed, cost and accuracy. Many teams also look at a few measures in the same view, so that improvement in one area does not quietly cause harm in another. Where a business uses a transportation management system or a warehouse management system, many of these numbers can be produced automatically, but someone still needs to check that the definitions match what the business actually cares about.
Treat any metric as a question to investigate rather than a verdict. A change in a number is a prompt to look at what happened underneath.