Summary
Warehouse picking strategies are the rules that decide which orders a picker collects together, which part of the floor each picker covers, when work is released and which route each tour follows. The six common order picking methods are discrete, batch, cluster, zone, wave and case picking, and each one runs with a pick path such as S-shape, return or largest gap. The right mix depends on lines per order, SKU overlap between orders, carrier cut-offs and how much of a picker's day goes to walking.
Order picking is where most warehouse labor goes. A 2023 study in the journal Logistics puts it at up to 55% of the total operating costs of the warehouse, and most of a picker's time on a tour is spent walking between locations. The picking strategy sets how far people walk for every order line, so it is one of the few cost levers a warehouse manager controls directly.
A strategy is also more than a method name. A cluster picking trolley on a poor route can walk further than a discrete picker on a good one, and a wave schedule built for last year's order mix can undo the gains of both. None of these layers needs robots or a new WMS to change, because each one is a rule in software.
What a warehouse picking strategy decides
Every picking operation answers four questions, whether someone wrote the answers down or the WMS defaults answered them.
- Grouping: does a picker collect one order per trip, or the lines of several orders at once?
- Floor split: does every picker cover the whole warehouse, or does each one stay in a zone?
- Release: does work drop to the floor in scheduled waves, or continuously as orders arrive?
- Routing: in what sequence does the picker visit the locations on one tour?
The method names in common use mix these layers. Wave picking is a release rule, zone picking is a floor split and cluster picking is a grouping rule with the sorting done on the trolley. That is why one site can run waves, zones and batches at the same time without any contradiction, and why "wave vs batch" comparisons often confuse more than they help.
Grouping and routing decide most of the walking, and walking is most of the job. A 2025 paper in the journal Operations Research and Decisions notes that travelling "constitutes over half of total order picking time", which makes it "the biggest area for improvement". Searching, picking and confirming share the rest.
Routing is the layer most often left to chance. Many WMS setups sort a pick list by location number and let that sequence stand in for a route. The picker then snakes through every aisle on the list in number order, whatever the shape of the order, and through aisles that hold a single item.
The 6 order picking methods, and when each one fits
Six methods cover almost every manual warehouse. Each method fits a specific order profile, and each one moves some of the work somewhere else: to a sort station, to a hand-off between zones or to the planning desk.
1. Discrete picking (single order picking)
One picker, one order, one trip. The picker walks the order from the first line to the last and drops it at packing before starting the next one. Discrete picking is the simplest method to run and the most expensive in walking, because two orders that need the same aisle send two people down it. It fits long orders that fill a tour on their own, urgent single orders and sites with low daily volume.
2. Batch picking
One picker collects the lines of several orders in one trip, often totalled per SKU, and the orders are separated afterwards at a sort or put wall. Batch picking cuts trips when many small orders share the same SKUs, which is why e-commerce sites with one- and two-line orders rely on it. The cost is a second touch, because every item is handled again at sorting.
How much batching saves depends on how the batches are built. The 2023 Logistics study built batches and routes together for a real warehouse and reports that "the picking time was reduced by 57% compared to single-order strategy". Batches built by order number or arrival time save less, because they can put orders from opposite ends of the building on the same trolley.
3. Cluster picking
Cluster picking is batch picking with the sort done on the trolley. The picker pushes a cart with one tote or box per order and puts each item straight into its order's tote at the pick face, so nothing needs sorting at packing. It fits small and medium orders that share aisles. The limit is the cart itself: the number, size and weight of the totes decide how many orders ride along, and bulky items fill it fast. The cluster vs batch picking comparison goes through the labor and sorting trade-off.
Which orders share a trolley matters more than how many. Group orders whose lines sit in the same part of the floor and each tour covers a short stretch. Group them by order number and each tour crosses the building. Batch, cluster and other forms of multi-order picking all depend on this choice. The diagram shows the same picks before and after order clustering, with each colour one tour.
4. Zone picking and pick and pass
The floor is split into zones and each picker stays in one. Orders either travel from zone to zone in a tote, which is pick and pass, or every zone picks its part at the same time and the parts are merged before packing. Zone picking caps how far one picker walks and keeps pickers in an area they know, which helps with large SKU ranges and with areas that must stay apart, such as cold, hazardous or high-value goods. The cost is the hand-off: one slow zone holds up every order that passes through it, and parallel zones need a merge step.
Zoning and batching deal with congestion in different ways. Zones keep pickers out of each other's aisles, while batching needs fewer trips in the first place, so fewer people are in the aisles at once. The article on zone picking vs batch picking covers when each fits and why many sites combine them.
5. Wave picking
Wave picking releases orders to the floor in scheduled blocks, often timed to carrier cut-offs or shift changes. A wave is a release rule, not a way of walking: inside a wave, people still pick discrete, in batches or in zones. Waves make it easy to line up labor with truck departures. The price is waiting time, since an order that arrives just after a wave closes sits until the next one. Wave picking in warehousing explains how to size and improve waves.
6. Case and pallet picking
Case picking moves full cases rather than single units, usually onto a pallet, and pallet picking moves whole pallets. Case picking fits B2B, wholesale and store replenishment, where customers order by the case. The pick sequence matters for a second reason here: heavy, strong cases go on the pallet first so the load stays stable, so the shortest route is not always the right one. Many sites run case picking and each picking as two separate flows.
Two more terms you will meet
Each picking, also called piece picking, means picking single units rather than full cases, and the first five methods above are usually each picking. Pick to zero means a pick that empties a location. Many WMS setups then ask the picker to confirm that the location is empty, which doubles as a quick stock check during normal picking.
Pick path strategies: how the picker walks the tour
The picking method decides which lines share a tour. The pick path decides the order in which the picker visits them, and any method can run with any pick path. The classic routing policies are fixed rules for walking the aisles, applied the same way to every pick list.
Simple routing policies leave distance on the floor. In the 2025 Operations Research and Decisions paper, S-shape routes came out on average 25% longer and return routes 33% longer than the routes from the authors' own algorithm. The paper also compares the midpoint policy, which came closer but still fell short.
The location sequence in a WMS is usually an S-shape in practice. A pick list sorted by location number sends the picker down aisle 1, 2 and 3 in order, often walking a full aisle for one item. Real floors add one-way aisles, cross-aisles, blocked areas and drop-off points that no textbook policy accounts for. The diagram shows the same pick list walked as a full snake and as a calculated route that uses the cross-aisle.
How to choose a picking strategy for your order profile
Order profile decides the method more than warehouse size does. Four measurements, all available from a few weeks of order history and one shift of observation, place you in the decision table below.
Measure four things first
- Travel share. Follow a handful of pick tours and split the time into walking, searching, picking and confirming. If walking is under a quarter of the tour, look at replenishment, search time or confirmation steps before changing the method.
- Lines per order. Mostly one- and two-line orders point to batch or cluster picking. Orders long enough to fill a trolley on their own suit discrete or zone picking.
- SKU overlap. Count how often orders released in the same hour share a SKU or an aisle. High overlap is what makes batching pay.
- Cut-offs and staffing. Hard carrier cut-offs favour waves, or at least release rules tied to departure times. Peaks with many new people on the floor favour methods with clear instructions on a screen over methods that depend on knowing the floor.
The KPIs and formulas for the first measurement are in how to measure and improve warehouse picking performance. Measure before choosing, because a method change on a site where walking is a small share of the tour moves very little.
Most sites end up with a mix, for example case picking for store orders, cluster picking for web orders and waves timed to the two daily trucks. The table gives the starting point for each flow. After that, the mix has to keep pace with an order profile that keeps changing.
Why a picking strategy that worked last year stops working
The order profile moves and the rules stay put. A new client, a promotion or a shift from store orders to web orders changes lines per order and SKU overlap within weeks. A batch size, zone boundary or wave schedule set two years ago keeps running, because it sits in a WMS setting that nobody has a reason to open.
Batching and routing depend on each other. The best batch is the one with the shortest route, and the shortest route depends on which orders are in the batch. Deciding them one after the other, batches by a fixed rule and routes by location sequence, settles for less than deciding them together. The 2023 Logistics study made the same point by modelling batching, batch sequencing and routing as one problem.
Constraints pile up. Trolley capacity, item weight, cut-off times, order priorities and one-way aisles all limit which orders can ride together. A fixed rule can respect one or two of them. Weighing all of them for every batch, every hour, is a calculation that a WMS rule engine is not built for.
Most WMSs do not know how far apart two locations are. They manage stock and transactions well, but without distances they fall back on location order for routes and on order number or arrival time for batches. That gap is what warehouse picking optimization fills: real distances, real constraints and a fresh calculation for every batch.
Running batching and routing on top of your WMS
You can change the strategy without changing the WMS or buying robots. Optioryx Pulse sits on top of the existing WMS. It reads the open orders, decides which orders share a trolley and calculates the route for each pick list on a digital map of the floor, with one-way aisles, cross-aisles and trolley limits included. The pick tasks go back to the WMS, so pickers keep their scanners and their way of working.
Pulse decides batching and routing together, so each batch is built around the route it will produce. It finds the optimal route within one second for 95% of routes. Across Optioryx customer deployments, clustering, routing and slotting together cut walking distance by up to 50%, with up to 30% increase in pick rate. The animation shows one tour on a demo layout, walked as the as-is snake path and as the calculated route.
A MedTech distribution center measured the effect on its own floor. The site picks about 30,000 order lines a day by hand, with 60 operators. With Pulse clustering orders onto each trolley on top of the existing WMS, the first measured result was a 15% increase in picking productivity, worth up to 9 FTE. With route optimization and double-stacked trolleys added, productivity rose to 27% above the level before any optimization software. The customer video tells the story.
Bleckmann, a 3PL for fashion and lifestyle brands, runs Pulse route optimization alongside its existing WMS. A route is generated for each pick list as soon as the picker is ready to walk, and Bleckmann reports an average of 15 to 20% less walking than with non-optimized routes.
Start with your own numbers. The order picking ROI calculator turns four inputs (yearly cost per picker, orders per tour, picking transport and number of pickers) into a yearly saving and a payback time. Pulse can also run on a few weeks of your order history and your layout before any WMS connection, so you see the routes and the walking against your current strategy first. Running it live then needs the WMS connection, by file exchange or API.
Pulse fits manual picking operations from around 10 pickers. For a one-page summary of Pulse picking optimization, download the AI picking overview.
Questions?
The six common order picking methods are discrete (single order) picking, batch picking, cluster picking, zone picking, wave picking and case or pallet picking. Each one runs with a pick path, such as S-shape, return, largest gap or a calculated shortest route. Most warehouses combine two or three methods, for example case picking for store orders and cluster picking for web orders.
It depends on the order profile. Batch and cluster picking save the most walking when many small orders share SKUs and aisles, while discrete or zone picking suits long orders that fill a tour on their own. How batches and routes are built matters as much as the method: a 2023 study in the journal Logistics planned batching and routing together for a real warehouse and reports 57% less picking time than single-order picking.
A picking method decides which order lines a picker collects on one tour, such as one order, a batch or one zone. A pick path strategy decides the sequence in which the picker visits those locations, such as S-shape, return or largest gap routing. Any method can run with any pick path, and both change the walking distance.
Pick to zero means a pick that empties a storage location. Many WMS setups then ask the picker to confirm that the location is empty, which works as a quick stock check during normal picking. It is a confirmation rule that works with any of the six picking methods.
Yes. Batching, routing and wave release are decisions that an optimization layer can make on top of the WMS, from the order and location data the WMS already holds. Optioryx Pulse sends the optimized pick tasks back to the WMS, so pickers keep their scanners, and customers see up to 30% increase in pick rate. You can test it on your own order history before connecting it.