Warehouse KPIs and Metrics Every Ecommerce Business Should Track

Why Warehouse Metrics Matter for Ecommerce

Running a warehouse without metrics is like running a business without financial statements. You know generally whether things feel good or bad, but you cannot identify specific problems, measure the impact of changes, or make data-driven decisions about where to invest time and money. The most common failure mode for growing ecommerce warehouses is not a sudden catastrophe but a gradual decline in performance that nobody notices until customers start leaving negative reviews about wrong items and slow shipping. By the time the reviews appear, the underlying problems have been compounding for weeks or months.

Metrics create accountability and visibility. When every picker knows that their individual pick accuracy and units per hour are being tracked, performance improves even without incentive programs, simply because people perform better when they know performance is visible. When the warehouse manager reviews a daily dashboard showing yesterday's on-time shipping rate dropped to 91%, they can investigate immediately (was it a staffing shortage? a late inbound shipment? a WMS issue?) rather than discovering the problem through a spike in customer complaints three days later. The investment in tracking warehouse KPIs is primarily time, perhaps 15 to 30 minutes daily to review dashboards and investigate anomalies, plus the WMS software that captures the underlying data automatically.

Order Accuracy Rate

Order accuracy rate measures the percentage of orders shipped with the correct items in the correct quantities. It is calculated by dividing the number of correctly fulfilled orders by the total number of orders shipped over a given period, then multiplying by 100. An order accuracy rate of 99.5% means that 5 out of every 1,000 orders contain an error.

For ecommerce, the target is 99.5% or higher. World-class operations achieve 99.9%+. The industry average for operations without barcode scan verification is 97% to 99%, meaning 1 to 3 errors per 100 orders. This might sound acceptable until you calculate the cost. Each mispick generates a return (average return shipping cost $5 to $8 for a prepaid label), labor to receive and process the return ($3 to $5), re-picking and re-shipping the correct order ($4 to $6 in labor plus the new shipping label), and the packing materials consumed twice ($1 to $3). Total cost per error: $13 to $22. For a 500-order-per-day operation at 98% accuracy, that is 10 errors per day, costing $130 to $220 daily or $3,900 to $6,600 per month in pure waste.

The hidden cost of inaccuracy is even larger: negative reviews mentioning wrong items, customer service time handling complaints, and the permanent loss of customers who had a bad experience. Research from Convey (now project44) found that 84% of consumers will not return to a brand after a single poor delivery experience. That makes each mispick not just a $15 to $20 direct cost but a potential lifetime customer value loss of hundreds or thousands of dollars.

To improve order accuracy, implement scan verification at every handoff point: scanning at pick (verify the right item from the right location), scanning at pack (verify every item is present and correct before sealing), and scanning at ship (verify the right label is on the right package). Each scanning checkpoint catches errors that slipped through previous steps. Operations that scan at all three points routinely achieve 99.9%+ accuracy regardless of order complexity or picker experience level.

On-Time Shipping Rate

On-time shipping rate measures the percentage of orders that are picked, packed, and handed to a carrier within your promised fulfillment window. For most ecommerce businesses, the standard is same-day shipping for orders placed before a cutoff time (typically 2:00 PM or 3:00 PM local time). The metric is calculated as: orders shipped on time divided by total orders eligible for same-day shipping, multiplied by 100.

The target is 98% or higher during normal periods and 95%+ during peak periods (holiday season, major promotions). Consistently falling below 95% means your warehouse capacity does not match your order volume, and you need to address the bottleneck, whether it is picking speed, pack station throughput, staffing levels, or inbound receiving delays that prevent inventory from reaching pick locations in time.

Track this metric daily and pay attention to the pattern. If on-time rate drops every Monday (because weekend orders pile up), you may need Sunday or early Monday staffing to work through the backlog. If it drops during the last week of every month (when promotional campaigns drive volume spikes), you need better demand forecasting and temporary staffing plans. If it drops on days when large inbound shipments arrive, your receiving process is pulling labor away from outbound fulfillment, and you need to stagger receiving schedules to avoid conflicts.

Amazon Seller fulfilled metrics impose a Late Shipment Rate threshold of under 4% for marketplace sellers. Exceeding this rate triggers account health warnings and can lead to suspension. If you sell on Amazon Seller Central and fulfill orders yourself rather than using FBA, tracking on-time shipping rate is not just a best practice but a requirement for maintaining selling privileges.

Cost Per Order

Cost per order (CPO) is your total warehouse operating cost divided by the number of orders shipped in the same period. It captures everything: rent, labor (wages, benefits, payroll taxes), equipment depreciation, software subscriptions, packing materials, utilities, insurance, and maintenance. CPO is the master metric that tells you whether your warehouse is efficient overall, and tracking it monthly reveals trends that individual metrics might miss.

Industry benchmarks for ecommerce CPO depend heavily on order complexity, average items per order, and automation level. Single-item orders with basic packaging in a well-run manual warehouse typically cost $2.50 to $3.50 per order. Multi-item orders (3 to 5 items average) in a manual warehouse run $4.00 to $6.00. Operations with significant automation (AMRs or conveyor systems) processing 1,000+ orders per day can achieve $2.00 to $3.00 per order even for multi-item orders. For context, most 3PLs charge ecommerce brands $3.00 to $5.00 per order for pick and pack (not including the shipping label), so keeping your in-house CPO below 3PL pricing is a key benchmark if you are running your own fulfillment.

To calculate CPO accurately, capture all warehouse costs including those that are easy to forget: the warehouse manager's salary allocated to fulfillment operations, WMS software subscription, scanner and printer depreciation (spread the purchase cost over 3 to 5 year useful life), packing tape, void fill, labels, insurance for warehouse operations, and maintenance on equipment. Divide the total monthly cost by the total orders shipped that month. Common mistakes include omitting rent (if you own the building, use the opportunity cost of market rent), omitting management overhead, and omitting packing materials, all of which understate the true cost and create a misleading comparison against 3PL pricing.

The biggest levers for reducing CPO are labor efficiency (through better picking methods and automation), which typically represents 50% to 65% of total warehouse cost, followed by rent optimization (getting more throughput from your existing space through better layout and dense storage), and packing material optimization (right-sizing boxes to reduce both material cost and dimensional weight shipping charges).

Units Per Labor Hour

Units per labor hour (UPLH) measures how many individual items your warehouse team processes per hour of paid labor time. It is the most direct measure of warehouse productivity, and it captures the combined effect of pick speed, travel time, pack speed, process efficiency, and downtime. Calculate it by dividing the total units shipped in a period by the total warehouse labor hours worked in that same period.

Benchmarks vary significantly by operation type and automation level. A manual, unstructured ecommerce warehouse doing discrete picking typically achieves 20 to 40 UPLH. Implementing batch picking with organized bin locations pushes this to 50 to 80 UPLH. Adding barcode scanning and a WMS with pick path optimization reaches 70 to 120 UPLH. Operations with AMRs or conveyor-based sortation achieve 100 to 200 UPLH. Fully automated goods-to-person systems can reach 200 to 400+ UPLH because the human operator never walks and picks at a sustained pace dictated by the system's bin presentation rate.

Track UPLH by individual worker as well as warehouse-wide. Individual UPLH data reveals training needs (a new picker at 25 UPLH compared to veterans at 60 UPLH needs coaching on efficient pick paths and product location familiarity), identifies top performers who can serve as trainers or leads, and flags potential issues (a previously strong picker whose UPLH drops suddenly may have an ergonomic issue, a personal problem, or may be assigned to a poorly slotted zone that needs re-organization).

When comparing UPLH across periods, control for order mix. A day with 90% single-item orders will show higher UPLH than a day with 60% multi-item orders requiring multiple picks per order, even if your team worked with identical efficiency. Normalize by comparing UPLH across similar order profiles, or use units per labor hour for each process step (pick, pack, receive) separately so that changes in order mix do not distort the overall number.

Inventory Accuracy

Inventory accuracy measures how well your physical on-hand quantities match what your system says you have. Calculate it by dividing the number of SKUs where the physical count matches the system count by the total number of SKUs counted, multiplied by 100. You can also calculate it by dollar value (total value of accurately tracked inventory divided by total inventory value) which gives a more business-relevant view since accuracy on your high-value A items matters more than accuracy on low-value C items.

The target is 99% or higher by SKU count and 99.5%+ by value. The industry average for ecommerce warehouses without systematic controls is 85% to 95%, meaning 5% to 15% of SKU counts are wrong at any given time. That level of inaccuracy causes two expensive problems: overselling (listing items as available when physical stock has been depleted, leading to canceled orders and angry customers) and phantom stock (the system shows zero when physical inventory exists, causing missed sales on products you actually have in the building).

Inventory accuracy is maintained through a combination of scan-verified warehouse processes (every receive, putaway, pick, and adjustment is scanned) and regular cycle counting to catch and correct the discrepancies that inevitably creep in despite scanning. The connection between warehouse management and inventory management is tightest here: your inventory data can only be as accurate as your warehouse processes allow. If products are received without scanning, put away in the wrong location, or picked without verification, no amount of software sophistication will keep your counts accurate.

Additional Metrics Worth Tracking

Receiving dock-to-stock time measures how long it takes from when a shipment arrives at the receiving dock until all items are counted, inspected, labeled, and placed in their designated storage locations. Target: under 24 hours for standard shipments, under 4 hours for urgent replenishment. Long dock-to-stock times mean incoming inventory sits in receiving instead of being available for order fulfillment, which can cause stockouts on your website even though the product is physically in the building. This is a common hidden cause of "out of stock" problems that frustrate both customers and inventory planners.

Return processing time measures how long it takes from when a returned package arrives until the item is inspected, restocked (or disposed), the customer is refunded, and the inventory system is updated. Target: under 48 hours. Returns that sit unprocessed for days or weeks create two problems: customers waiting for refunds become frustrated and contact support (increasing customer service costs), and returned items that could be resold sit in a returns pile instead of being available for new orders.

Space utilization rate measures the percentage of your available storage positions that are currently occupied. Calculate it as occupied positions divided by total positions. Track it weekly. Below 60% means you are paying for space you are not using. Above 85% means you are approaching capacity and need to plan for expansion. At 90%+, efficiency starts declining because pickers encounter congestion, putaway locations become scarce, and overflow product ends up in temporary locations that your WMS does not track, degrading inventory accuracy.

Perfect order rate is the percentage of orders that are delivered to the customer on time, with the correct items, undamaged, and with the correct documentation (invoice, packing slip). It is the ultimate customer-facing quality metric because it combines warehouse accuracy (right items), warehouse speed (shipped on time), packing quality (undamaged in transit), and administrative accuracy (correct paperwork). A 99% order accuracy rate and 98% on-time shipping rate does not mean 97% perfect order rate, because some orders that shipped on time may have been inaccurate, and some accurate orders may have shipped late. Perfect order rate is typically 2 to 5 percentage points lower than your worst individual metric.