Summary
Any slotting plan goes out of date over time and in peak season it is the most visible. These are mistakes we have seen 120+ warehouses make before and during peak seasons.
A slot plan built for a normal week stops working in warehouse peak season. Once more pickers are working the same aisles, waiting can cost more than walking. Black Friday, Cyber Monday, holiday shopping and back-to-school all do the same thing to a warehouse. Order volumes go up, and the site has to run at a higher capacity to keep up. More orders, more pickers, more movement through the same aisles.
What a slot plan optimizes for
A slot plan balances multiple objectives at once. These five matter most, and three of them need different weights during peak, though the method itself does not change.
Walk distance drops down the list because waiting time gets more expensive than walking time. Concentrating demand and picks is exactly what shortens the walk, and it is also what creates the traffic.
Replenishment climbs for the opposite reason. A plan that minimizes picker walking can double the replenishment trips, which moves labor cost around without removing it.
We went through the mechanics of that trade in more depth in our article on what a slot plan is optimizing for.
The new weights have to be in place before the season starts. A slot plan is a physical thing. Moving stock takes shifts, and shifts are what you run out of first.
Signs your slot plan is the problem
The clearest sign is picks per hour falling while headcount goes up. If output per person drops every time you add pickers, the problem is where the items sit rather than how many people you have on the floor.
Three other signals are worth checking.
- Pickers queueing at the same pick faces during the morning wave.
- Replenishment teams working around the picking teams when they should be working ahead of them.
- A slot plan whose last real change is older than your current order profile.
The 6 slotting mistakes we see before peak season
1. More pickers does not mean higher productivity
Congestion gets much worse in peak. Warehouses answer volume by adding pickers, and pickers get in each other's way. Order volume goes up in a straight line. The time lost waiting does not.
.png)
It happens in three ways, and none of them show up on a slotting report.
- In the aisle. Someone is stopped at a pick face and you cannot get past.
- At the pick face. Two pickers need the same location at the same moment.
- Across trips. Some carry far more picks than others.
Open access research on picking route blockages describes the first one plainly. Aisle blocking happens when a forklift is parked in the aisle, when carts or pallets with products on them are sitting in the area, or when someone else is picking in the area (Mathematics, 2024).
Wide aisles help less than the argument assumes. Pratik Parikh and Russell Meller modelled pick-face blocking in aisles wide enough for pickers to pass each other. A wide aisle beats a narrow one only when each picker takes a single SKU at a pick face. Once pickers take more than one SKU at the same face, blocking rises monotonically with the number of SKUs picked (IIE Transactions, 2009, 41(3), 232 to 246).
The lever is where the items sit, not how many people you put in the aisles. Spread the picks evenly and the extra pickers have somewhere to go.
Volume pushes productivity in two directions at once. More volume puts more pickers in the same aisles, so blocking and waiting time rise. It also puts more lines in every batch and more picks in every aisle a picker already visits, so the walk per pick gets shorter. Which force wins depends on your layout and your order profile.
Both forces are non-linear and they act on each other, which is why a spreadsheet cannot answer it and a simulation can.
2. Slotting was done once
Someone set the slot plan when the site opened, or a consultant did it once and handed over a file. Then it was left alone. Since then items go wherever there is space, and nobody moves them back.
The fix is not a full reslot. Audit your most frequently picked items and move them closer to outbound, ranked by impact. Five or ten moves at the end of a shift is enough to start.
Moving them is the easy part. Working out which SKUs are costing you is the hard part. Our own modelling puts a ten-move list at roughly 3 to 4% more picks per hour.
3. Slotting lives in Excel
Warehouses that keep the slot plan in Excel see a tidy file and a messy floor. The sheet holds SKUs and location codes. It holds no aisles, no travel distances and no traffic.
You cannot see a queue in a list. Work the plan against a layout map. Distance, congestion and replenishment effort only exist once the warehouse has a shape.
4. You only slot by ABC
ABC ranks items by how often they move and stops there. It puts every A-item in the same block near the dispatch end. Shorter walks, and one traffic jam.
More pickers in fixed aisles means blocking drives performance, which is the exact situation peak creates. Go past velocity classes. Spread top movers across more aisles, weight congestion alongside walk distance, and put slow or empty locations between the fast faces so the stops spread out along the aisle.
5. Your WMS "does it"
Most WMS slotting modules apply rules one constraint at a time. They do not optimize against competing constraints. Many also cannot handle a SKU without a fixed home, one that moves location as stock comes in.
Check what your slotting module actually weighs. If it cannot see aisles and travel distance, it cannot balance them. The plan has to be worked out somewhere that can, then go back into the WMS.
An example of what it looks like when the WMS is left to do it: a medical technology distribution center ran a well known WMS that handled the picking optimization, both the clustering and the pick sequence. Routes came out of static business rules and a general Z-snake, a serpentine path that covers the aisles in a set order. Those rules never looked at the layout. The trips took far longer than they needed to.
Warehouse optimization software (WOS) sitting on top of the same WMS took over how picks were clustered and sequenced before a trolley ever reached a picker. The operators kept working exactly as before, and picking productivity went up 27%.
6. Ignoring SKU frequency and pairing
Two items that ship in the same order sit at opposite ends of the building. Every order touching both pays for the walk, and in peak that pair shows up in far more orders.
Pull the pairs out of your order history and put frequency and pairing into the plan. Then re-check the groups for peak. Affinity that cuts distance can also stack pickers into one aisle, so the grouping that worked in October is not the grouping you want in November.
Why slotting is hard
Every fix above trades something away. Spreading top movers across more aisles cuts congestion and adds walk distance. Duplicating an extreme mover buys parallel picking and costs you replenishment trips.
Which trade is worth making depends on your layout and your peak order profile, and neither of those matches the industry average.
This is why a slotting checklist alone does not get you there. A checklist tells you which levers exist. It cannot tell you where your own break-even sits.
What changes in the plan when congestion gets more weight
Raising the congestion weight is not abandoning velocity. It is accepting a slightly longer average trip in exchange for far less standing still. Five moves do most of the work.
- Spread the top movers over more aisles. Concentrating every fast mover into the same block of aisles is what creates the traffic.
- Balance forecast picks per aisle, not just distance. Use the peak forecast rather than the annual one. A C-velocity item in June can be your number three line in November.
- Duplicate the extreme movers. For the handful of SKUs that appear in a very large share of orders, a second pick face in another zone buys parallel picking and costs you replenishment trips.
- Re-check the affinity groups. Affinity slotting cuts distance per order, and it also puts pickers working similar orders in the same place at the same time.
- Leave gaps between the most-visited pick faces. Slow or empty locations between fast movers spread the stops along the aisle, so two pickers are less likely to slow each other down.
What to fix first, in order of impact
There are four levers on picking productivity and they are not equal. Slotting is the biggest one and the slowest to change, which is why it belongs in the quarter before peak rather than in November, when moving stock costs you shifts you no longer have.
The reason slotting sits at the top is that travel dominates the work. Travel is the most expensive activity in order picking, at around 55% of picking time (Axioms, 2023).
- Slotting. The basics, and difficult and slow to change once peak has started.
- Order clustering and batching. Which orders travel together directly controls how many pickers are in the same aisle at the same moment.
- Wave release. Holding a wave back ten minutes to level out volumes is congestion control, not just scheduling.
- Pick path. Sequencing the walk inside a batch is real but small. If an optimization vendor leads with pick path, they are selling you the smallest lever.
Slotting itself breaks into four actions, and only one of them is a project. Re-slotting is the capped end-of-shift move list. A full re-slot assigns every SKU its best location from scratch and has to be paired with ongoing re-slotting, or it starts decaying the day it goes live. In replenishment means every inbound SKU gets the best open location instead of the first free one. SKU housekeeping merges and splits locations to free up space and stop lanes running dry. Three of the four happen inside normal operations, which is why they are the ones to reach for when the season is already close.
Testing the plan is cheaper than moving the stock twice. Pulse Slotting runs on a warehouse digital twin. A slot plan gets rebuilt against your own layout and order history, then run at 150, 200 and 250 percent of a normal week before anything moves on the floor.
Whether a full digital twin or a lighter simulation fits depends on how complex your warehouse actually is.
Slotting on its own cuts travel time by 15 to 25% across customer deployments.
Questions?
Picker blocking is time lost when one picker has to wait for another. It comes in two forms: in-aisle blocking, where a picker cannot pass a colleague stopped at a pick face, and pick-face blocking, where two pickers need the same location at the same time.
Pick-face blocking happens in wide aisles too.
Because pickers interfere with each other, and that interference does not scale linearly with volume.
Volume rises in a straight line. Blocking does not.
No. Wide aisles remove in-aisle blocking, where one picker cannot pass another, but not pick-face blocking, where two pickers need the same location at the same moment.
Yes, in what it optimizes for rather than in method. Most slot plans weight picking walk distance highest, which is correct at normal volumes.
In peak, waiting time costs more than walking time, so the congestion objective should carry more weight and picks per aisle should be balanced on the peak forecast rather than trailing velocity.
Slotting first, because it is slow to change once peak has started and needs doing in the quarter before peak. Batching second, because which orders travel together decides how many pickers share an aisle. Wave release third.
Pick path last, because sequencing the walk inside a batch has the smallest impact of the four.