Order picking is the process of selecting items from storage to fulfil a customer order. Somebody — or something — travels to a location, takes the right product in the right quantity, and brings it to be packed.
It sounds like the simplest activity in a warehouse and it is consistently the most expensive. Picking typically accounts for somewhere around half of all warehouse operating cost, and the great majority of that is not the picking itself but the walking between picks. Which is why almost every improvement in this area is really an improvement in travel.
Where order picking fits, and why it costs so much
The flow through a warehouse runs receiving, putaway, storage, picking, packing, dispatch. Picking is the point where stored inventory becomes a specific customer’s order, and it is the first stage that is driven by demand rather than by supply.
Three things make it the dominant cost.
It is the most labour-intensive step. Receiving and putaway handle goods in bulk; picking handles them one order at a time, and orders arrive in whatever combination customers choose.
Travel dominates the time. Studies of manual picking consistently put walking at around half the picker’s day, and it adds no value whatever. A picker who walks less picks more without working harder.
It is where accuracy reaches the customer. A mis-pick is not caught by the system — it ships, arrives wrong, and comes back as a return, a replacement and a customer deciding whether to order again. Errors at receiving cost a minute; errors at picking cost an order.
That combination is why picking method, slotting and layout get more attention than any other warehouse decision, and why the warehouse KPIs that matter most sit here.
The main order picking methods
| Method | How it works | Best for |
|---|---|---|
| Single order (discrete) | One picker takes one order from start to finish | Low volumes, large or few-line orders, businesses starting out |
| Batch | One picker collects the same item for several orders in one trip, sorted afterwards | Many small orders sharing common products |
| Multi-batch (cluster) | A picker carries several order containers at once and sorts into them whilst picking | Small multi-line orders; very common in ecommerce |
| Zone | The warehouse is divided into areas; each picker works only their own zone | Large buildings, high volumes, distinct product types |
| Zone pick-and-pass | The order tote travels between zones, each picker adding their part | Multi-line orders spanning a large building |
| Wave | Picking is released in scheduled groups, timed to carrier cut-offs or dispatch slots | Operations with fixed collection times and sortation |
| Pick-to-order | Items are picked directly into the shipping carton | Reduces a handling step where packaging allows |
| Piece / each | Individual units picked from a shelf or bin | Ecommerce, spare parts, anything sold singly |
| Case | Full cartons picked from pallets | Wholesale, retail replenishment, distribution |
| Pallet (full unit) | Whole pallets moved without breaking down | Bulk distribution and cross-docking |
Single order picking is the default everybody starts with. It is simple, the order is never split so nothing can be lost between stages, and it is the least efficient by a wide margin because the picker walks the whole route for every order.
Batch picking attacks that directly. If forty orders all contain the same product, one trip collects forty units. The saving is large and it introduces a sortation step afterwards, where the picked quantities are split back into individual orders — which is where errors enter if it is done badly.
Cluster or multi-batch picking removes the separate sortation. The picker pushes a trolley holding several totes, one per order, and drops each unit straight into the right tote as they go. It combines the travel saving of batching with the accuracy of picking to the order, and it is the reason trolleys with six or twelve tote positions are ubiquitous in ecommerce warehouses.
Zone picking suits large buildings. Each picker knows their area intimately and never leaves it, which cuts travel sharply and raises familiarity. The cost is coordination: an order spanning four zones has to be assembled from four contributions.
Pick-and-pass handles that by moving the tote rather than the picker. It works well and it creates a chain — if one zone is slow, everything behind it waits.
Wave picking is about timing rather than routing. Work is released in scheduled batches aligned to carrier collections, so everything for the 4pm courier is picked, packed and staged before it arrives. It is what prevents correctly picked orders missing their van.
Most real operations combine several. A common shape is zone picking across the building, batch picking within each zone, and waves timed to carrier cut-offs.
Case picking vs piece picking
The unit you pick changes almost everything about how the area is designed.
Piece picking, also called each picking or broken-case picking, means taking individual units. It is what ecommerce and spare parts operations do, and it is the most labour-intensive form because every unit is handled separately. It wants shelving, bins, short travel distances and a pick face at comfortable working height.
Case picking means taking full cartons, usually from pallets. Common in wholesale, retail replenishment and food distribution. Fewer picks for the same volume of goods, heavier handling, and it wants pallet racking with a ground-level pick face and reserve stock above it.
Pallet picking moves whole unit loads and is really materials handling rather than picking.
Many warehouses do all three, and the mistake is running them from the same area. Piece picking from a pallet rack aisle wastes a great deal of space and walking; case picking from shelving is slow and awkward. Separating the pick faces by unit type is one of the more reliable layout improvements available.
Manual, semi-automated and automated picking
Manual picking is a person with a paper list or a handheld scanner walking to locations. It is flexible, cheap to start, and scales by adding people — which is also its limit, because cost rises in a straight line with volume.
Semi-automated picking keeps the person and removes some of the walking or the errors:
- Pick-to-light — lights at each location show what to take and how many, confirmed with a button press. Fast and accurate in dense pick faces.
- Voice picking — a headset directs the picker, who confirms by speaking. Keeps both hands free and both eyes on the work, which suits case picking well.
- Conveyors — totes travel between zones instead of pickers walking them.
- Autonomous mobile robots — carts that meet pickers, follow them, and take completed totes away, removing the walk back to dispatch.
Automated picking removes the travel entirely or removes the person:
- Goods-to-person systems — an AS/RS or a shuttle system brings the stock to a fixed picking station. Travel drops to near zero and throughput rises sharply.
- Robotic piece picking — arms with vision systems selecting individual items. Real, improving quickly, and still limited by the variety of shapes it can handle reliably.
- Fully automated case and pallet handling — common in high-volume food and beverage distribution.
How to think about the choice. Automation buys throughput and consistency and costs capital and flexibility. It suits high, stable volumes and a product range that does not change shape often. Manual picking suits variety, volatility and businesses whose volumes are still moving. Most operations get more from slotting, batching and layout — which cost very little — than from equipment, and the ones that automate before doing those things tend to automate an inefficient process.
How to choose a picking method
Work through these in order:
Order profile first. How many lines does a typical order contain, and how many units per line? Single-line orders favour batch picking heavily. Multi-line orders favour cluster or pick-and-pass.
Order volume. Under a hundred orders a day, single order picking is usually fine and the coordination overhead of anything else is not repaid. Above a few hundred, batching or clustering pays immediately.
Product overlap. If many orders contain the same products, batching saves a lot. If every order is different, it saves little.
Building size. Long walks favour zoning. A small unit does not need it and will be slowed by the handoffs.
Dispatch pattern. Fixed carrier cut-offs favour wave picking. Rolling dispatch throughout the day does not need it.
Product characteristics. Size, weight, temperature requirements and whether items are picked as pieces, cases or pallets.
Peak ratio. A business whose December is five times its June needs a method that scales with temporary staff, which favours simple, well-directed methods over ones requiring deep building knowledge.
Then test the choice on paper before committing. Take a real day of orders and count the picks and the approximate travel under each method. The answer is usually clear and frequently not the one people assumed.
How to improve picking efficiency and accuracy
Slot by velocity. Put your fastest-moving lines closest to dispatch and at comfortable working height. This is free, it can be done in a weekend, and it typically removes a substantial share of walking immediately. Re-slot at least annually, because order profiles drift and layouts do not.
Label every location clearly and logically. Aisle, bay, level, position, in a scheme a new starter can follow. Everything else depends on this.
Batch or cluster wherever the order profile allows. The single largest method-level saving available to most operations.
Sequence the pick path. Pick lists should run in the order the picker walks, in one direction, never doubling back. A list sorted by product code instead of by location is a common and expensive default.
Scan the location and the product. Scanning the product confirms what was taken; scanning the location confirms it came from the right place, which catches the most common error — picking a similar item from the adjacent bay.
Keep the pick face replenished. An empty pick face with full pallets above it stalls the pick and is invisible in most reporting. Trigger replenishment on a level, not on someone noticing.
Separate the unit types. Piece picking, case picking and pallet picking from their own areas, designed for the job.
Group products that sell together. If two items appear in the same order repeatedly, storing them near each other cuts travel with no system change.
Measure the pair, not the single figure. Pick rate alone drives speed at the cost of accuracy. Track picking accuracy alongside it, always, and review both as a trend rather than a daily number.
Review errors by cause, not by count. Wrong item, wrong quantity, wrong location, damaged. Each points at something different — similar products stored adjacently, poor labelling, unclear quantities on the list.
Train and then check. Most picking errors come from a small number of causes and a small number of locations. Finding those is more productive than general exhortation to be careful.
Where to go deeper
- Warehouse processes — where picking sits in the full flow.
- Warehouse layout design — the arrangement that determines travel.
- Warehouse KPIs — pick rate, accuracy and cost per pick.
- What is a WMS — the software that directs and sequences picking.
- What is order fulfilment — the process picking belongs to.
- Barcode scanning for inventory — the accuracy mechanism.
The one-line takeaway
Order picking is selecting items from storage to fulfil orders, and it is the most expensive activity in a warehouse — mostly because of walking rather than picking. The improvements that pay are slotting by velocity, batching or clustering where the order profile allows, sequencing the pick path, and scanning both the location and the product.
Frequently asked questions
What is order picking? The process of selecting items from storage locations to fulfil a customer order. It is the stage where stored inventory becomes a specific order, and it typically accounts for around half of warehouse operating cost.
What are the main order picking methods? Single order (discrete), batch, multi-batch or cluster, zone, zone pick-and-pass, wave, and pick-to-order. Alongside those, picking is classified by unit as piece, case or pallet picking. Most operations combine several — commonly zone picking across the building with batching inside each zone and waves timed to carrier cut-offs.
What is the difference between batch picking and zone picking? Batch picking means one picker collects the same product for several orders in one trip, which is then sorted into individual orders. Zone picking means the warehouse is divided into areas and each picker works only their own zone, with multi-zone orders assembled from several contributions. Batching reduces travel per unit; zoning reduces travel per picker.
What is cluster picking? Picking several orders at once into separate containers on one trolley, dropping each unit straight into the correct order’s tote. It combines the travel saving of batch picking with the accuracy of picking to the order, and it removes the separate sortation step batching requires.
Manual or automated picking — which is better? Automation buys throughput and consistency at the cost of capital and flexibility, and suits high, stable volumes with a settled product range. Manual picking suits variety and changing volumes. Most operations get more from slotting, batching and pick-path sequencing — which cost almost nothing — than from equipment.
How do you improve picking accuracy? Scan both the location and the product at each pick, label locations clearly and logically, avoid storing visually similar products next to each other, sequence pick lists by location, keep the pick face replenished, and track accuracy alongside speed rather than measuring rate alone.
Why is order picking so expensive? Because it is the most labour-intensive step, because travel between locations consumes around half the picker’s time and adds no value, and because errors here reach the customer directly as a return and a replacement rather than being caught internally.
Warehouse Wise is a free, independent library of guides on inventory management, warehousing and fulfilment, written for the people actually doing the work — new warehouse and inventory staff, small and growing retail and ecommerce businesses, and anyone who holds stock and wants to understand it properly. Every guide explains the term, shows a worked example with real numbers where it helps, and stops. We are independent: we do not sell software or equipment, we take no commission on referrals, and our guides are not ranked by who pays us.