Warehouse Order Picking Methods: Productivity & Strategy

55 Percent

Fifty-five percent. That’s the share of total warehouse labor costs that picking accounts for. Not the whole building. Not receiving, putaway, packing, and shipping combined. Just picking. One process, more than half your labor spend.

When I tell clients that method selection is the single highest-leverage operational decision they can make inside a warehouse, I can prove it with that number. Labor represents 45 to 57% of total warehouse operating costs. Picking accounts for 55% of that labor. Which means picking alone drives roughly 25 to 30 cents of every dollar your operation spends.

The operation that chooses the right picking method for its order profile, order volume, and SKU density has a structural cost advantage over one that defaults to what they’ve always done. This module gives you the seven methods, the benchmark data for each, and the decision logic to know which one is right for a given situation.

The Picking Productivity Benchmark Table

Before diving into each method, here’s the full comparison you’ll reference repeatedly in real projects. Every number in this table is sourced:

Method 1: Discrete Picking

One picker, one order, one trip. The simplest method there is. Picker takes a paper ticket or RF gun, walks the full pick path for that single order, collects every item, delivers to pack. No sorting, no consolidation, no WMS sophistication required.

The benefit: 100% order integrity. Zero risk of mixing product between orders on the floor. Accountability is clear — one picker, one order, one person responsible if something goes wrong.

The cost: Every picker walks the entire facility for every single order. If you’re running 500 orders a day, you have 500 separate trips through the warehouse. In a facility with a 600-foot pick path, that’s 300,000 feet of travel for that day’s orders — before you account for any variation or returns.

Productivity:

• Paper-based discrete: 60–80 lines/hour
• Carton flow rack with paper: 184 lines/man-hour (Newcastle Systems)
• With RF scanner and WMS-optimized pick path: 100–140 lines/hour typically

When to use discrete:

• Under 100 orders per day
• High-complexity orders where each item needs careful attention
• Operations without the order volume to justify anything more sophisticated
• Special-handling requirements (fragile items, customer-specific instructions, regulated products)

Method 2: Batch Picking

One picker, multiple orders, one trip. The WMS groups orders that share common SKUs and pick locations, then sends one picker to service all of them on a single pass through the warehouse.

Here’s the efficiency logic: instead of four separate pickers each walking to Location A-03-12 and pulling one unit, one picker walks to A-03-12 once and pulls four units — one for each order in the batch. Same physical item, same pick location, one-quarter of the travel.

The trade-off: those four items go into a shared tote or container, which means a sort step at the pack station is required to separate them back into individual orders. With proper WMS labeling (each unit scanned to a specific order position), this sort is fast and accurate. The net result is still strongly positive.

Productivity improvement: 30–50% increase in lines per hour over discrete picking (ShipHero analysis). One e-commerce operation documented fulfilling 200 orders in 2 hours with one picker using batch picking — the same volume previously required four staff and four to five hours.

Batch size: This matters more than most operations realize.

• Too small (fewer than 4 orders): travel reduction benefit is minimal
• Too large (more than 12–15 orders): sort errors multiply, carts become unwieldy, picker loses track of order assignments
• Typical optimal batch size: 4 to 12 orders per picker trip, depending on order density and SKU overlap

When to use batch:

• High volume of small orders (1–3 items per order)
• Orders sharing common SKUs — WMS identifies co-occurrence
• Operations with reliable WMS and experienced pickers who can execute the sort accurately

Batch picking is the first productivity upgrade available to any operation running discrete. No new equipment. No major WMS upgrade in most cases. Just smarter grouping of orders before you release them to the floor. If you’re running discrete today at any volume above 100 orders per day, batch picking is worth evaluating this week.

Method 3: Wave Picking

Wave picking is frequently confused with batch picking. They’re different, and they work together.

Wave picking is a scheduling mechanism, not a picking method. A wave is a set of orders released to the floor at a specific time, coordinated with carrier cutoffs, packing capacity, and labor availability.

Here’s how wave planning works in practice:

The wave coordinates the whole downstream flow — pickers, packers, and the shipping dock are all synchronized to the same schedule. Without wave planning, you get pickers ahead of packers, packers waiting for picks, and the shipping dock either waiting or getting hit with a wall of freight at the last minute.

Wave and batch are almost always combined in practice: the wave controls when orders hit the floor, batch controls how many orders each picker handles per trip within that wave. A morning wave of 200 orders might release to 20 pickers in batches of 10.

Method 4: Zone Picking

The warehouse is divided into physical zones. Each picker is permanently assigned to one zone. When an order spans multiple zones, each zone picker handles their portion.

Zone picking addresses a specific problem: in a large warehouse with a broad SKU catalog, no single picker can know the entire building. Zone pickers develop expertise in their zone — faster location identification, fewer errors, quicker recognition when a replenishment is needed.

Two fulfillment models:

Zone-and-Pass (Pick-and-Pass)

The order container — tote, carton, pallet — physically travels zone-to-zone. Zone 1 picker adds their items, passes the tote to Zone 2 on a conveyor or flow lane, Zone 2 adds items, passes to Zone 3.

• Advantage: No consolidation step — the order is complete when it exits the last zone.
• Disadvantage: A slow zone creates a bottleneck for every zone downstream. If Zone 2 gets backed up, Zone 3 sits idle waiting. Zone balance is critical.

Zone-and-Consolidate

Each zone picks into separate totes independently. A consolidation station merges all zone totes for an order before packing.

• Advantage: Zones run at their own pace without dependency on upstream/downstream zones.
• Disadvantage: Requires consolidation infrastructure — conveyors, sorter, merge station — and adds a consolidation step to the workflow.

When to use zone picking:

• Large warehouses (200,000+ sq ft) with physically distinct product sections
• Operations with a high SKU count where picker specialization creates real efficiency
• Environments with consistent multi-line orders that span multiple product categories
• When zone-specific expertise (knowledge of products, locations, quality issues) has measurable value

Method 5: Cluster Picking

One picker, one cart with 4–12 totes, one trip through the warehouse — servicing multiple orders simultaneously.

The WMS assigns each tote to one order and groups orders that share pick locations. The picker travels the pick path once, and at each location pulls units for all orders on the cart that need that item. The WMS tells them which tote each unit goes into.

The efficiency gain is structural: a picker visiting Location A-03-12 on a 10-tote cluster cart services 10 potential orders from that single stop. On a traditional discrete pick with 10 separate trips, that same location gets visited 10 times.

Put-to-light on the cart — LED indicators on each tote compartment showing the quantity to deposit — eliminates the cognitive load of reading and processing instructions while picking. The picker looks at the light, deposits, moves on. This is where cluster picking accuracy gets closest to pick-to-light accuracy.

Productivity:

• 2–3× more productive than single-order discrete (JASCI Software)
• 20–30% fewer labor hours per unit of order volume versus discrete (ShipBob data)

When to use cluster picking:

• E-commerce operations with small order profiles: 1–5 items per order
• Order volume of 500 to 5,000 orders per day
• SKU profiles with high co-occurrence — many orders share a subset of popular SKUs
• Operations where pick-to-light infrastructure isn’t justified but batch accuracy needs to be high

Method 6: Pick-to-Light and Put-to-Light

Pick-to-light puts LED modules at each pick location. When the picker reaches a location, the light illuminates, shows quantity. Picker grabs product, confirms with a button press. No scanner. No screen to read. Hands-free, eyes-on-product.

The productivity and accuracy numbers from Newcastle Systems are the ones the industry cites:

Accuracy: 99.9%+ consistently. The system tells you exactly what to pick and requires you to confirm the pick. No mis-reads. No transposition errors. No handwriting interpretation issues.

The trade-off: Pick-to-light infrastructure is fixed to the shelving. High capital cost. It’s not practical to wire an entire warehouse — the ROI only works in dedicated high-velocity forward pick zones: the 10–20% of your SKU catalog that generates 60–80% of your picks, installed in a defined flow rack section near packing.

Put-to-light is the sorting companion used in batch operations. A batch picker brings a multi-order tote to a sort station; as each item is scanned, the light illuminates on the correct order cubby showing how many units to deposit. This eliminates the sort errors that are the primary weakness of batch picking at high volume.

Together, pick-to-light and put-to-light represent the highest accuracy, highest throughput configuration available without full automation — and they’re often the bridge step before an operation commits to goods-to-person capital.

Method 7: Goods-to-Person (GTP) Systems

Traditional picking: picker walks to inventory. In a typical warehouse, that’s 8 to 12 miles per shift of picker travel.

Goods-to-person: automation brings inventory to a stationary pick station. The picker walks less than 1 mile per shift (Invia Robotics data). You’re not eliminating the picker. You’re eliminating 45 minutes of walking per hour.

GTP doesn’t improve pick accuracy — it fundamentally changes what “picking” is. The picker doesn’t travel. They stand at a station, the inventory comes to them, they pick, confirm, and the system moves the storage unit back.

Four GTP System Types

AMR Shelf-to-Person (Locus Robotics, 6 River Systems)

Autonomous mobile robots navigate the warehouse and bring inventory pods/shelves to stationary pick stations.

• Pick rate: ~150 order lines/hour/picker (Dematic comparative analysis)
• Lowest pick rate of the GTP systems, but highest flexibility
• Add or remove robots to scale throughput without structural facility changes
• Best for small-to-medium fulfillment operations or facilities where headroom prevents cube-based systems

AMR Bin-to-Picker (inVia PickerWall)

Robots deliver individual bins to pick stations — more precise than full shelf delivery.

• Average: 350–450 UPH; bursts to 1,000 UPH (Invia Robotics)
• 90% reduction in worker walking
• AMR bin-to-picker: ~400 order lines/hour/picker (Dematic)

AutoStore (Cube-Based Grid)

Items stored in bins stacked in a dense grid. Robots traverse the grid top surface to retrieve bins.

• Up to 4× storage density per square foot versus traditional shelving
• Best-in-class for space-constrained operations and small item profiles
• Pick rates are moderate in GTP terms — bins delivered sequentially
• Used extensively in pharma dispensing, high-value electronics, e-commerce apparel

Dematic Multishuttle

Multiple shuttle vehicles operate within rack aisles to store and retrieve products.

• Pick rate: >600 order lines/hour/station — the highest in the GTP category
• Height utilization up to 20 meters; triple-deep storage configurations
• Highest capital cost; optimal ROI above 20,000 orders per day

The question with goods-to-person is not whether it’s faster — it obviously is. The question is whether your volume justifies the capital, and whether your operation has the technical maintenance capability to sustain it. Automation that’s down is worse than labor you can manage.

Pick Path Optimization: Reducing Travel Within Your Chosen Method

Regardless of which method you’re running, pick path optimization reduces travel within that method. It’s a parallel lever to method selection, not a substitute for it.

Serpentine (S-Pattern)

The picker traverses the warehouse in a continuous S-shape — down one aisle, up the next, alternating end-to-end. No backtracking. Halves travel distance versus the return-to-start method (Cargoz analysis).

This is the default WMS-optimized path for moderate SKU density operations and the most common configuration. Almost every WMS can generate serpentine-optimized pick lists automatically.

Skip-Aisle

Cross aisles are placed approximately two-thirds of the way down each pick aisle. Fast movers are slotted before the cross aisle; slow movers after. If a picker’s order doesn’t require anything in the back third of the aisle, they take the cross aisle and skip that section entirely.

This is most effective when combined with A-item forward slotting — if A-items are all positioned before the cross aisle, the majority of picks complete before the cross aisle, allowing most trips to skip the slower section entirely. Typical result: a meaningful reduction in per-aisle travel time for operations with concentrated A-item demand.

Z-Pattern

The picker goes into an aisle to the deepest required pick, then returns up the same aisle. Less efficient than serpentine for dense orders, but sometimes optimal for sparse orders where a single pick at the end of an aisle makes a full serpentine traverse wasteful.

Modern WMS systems calculate whether serpentine or Z-pattern is shorter for each individual order pick list and route accordingly. Manual selection of a single path pattern for all orders leaves optimization on the table.

Choosing Your Method: A Decision Framework

These are starting points, not prescriptions. The right method for any specific operation depends on your SKU count, order profile distribution, facility layout, and labor market. A 10,000-order-per-day operation with a narrow SKU catalog and large average order size has different optimal methods than a 10,000-order-per-day operation with 50,000 SKUs and 2-line average orders.