As the Head of Business and Engineering at Acquaint Softtech, an AI development services partner, I hear the same question from warehouse operators: our pickers are busy all shift, so why are orders still leaving late?
The answer usually sits in slotting, which decides where each product lives and therefore how far pickers walk per order line. This guide covers the algorithms behind it, from ABC velocity to affinity and re-slotting rules.
The approach draws on Acquaint Softtech’s work as an Official Laravel Partner, with 1,300+ projects, 70+ in-house engineers, and clients across the USA, UK, Europe, and Australia.
What warehouse slot optimization is
Warehouse slot optimization is assigning each SKU (stock keeping unit, one distinct stored item) to the location that minimizes picking effort within physical limits. It turns order history into a placement plan.
Why does picking speed depend on slotting?
Travel is often cited as around half of total picking time, notably in Tompkins’ Facilities Planning. Each fast mover stored far from dispatch adds walking to thousands of order lines.
The slotting methods compared
Four logics cover most slotting engines, and production systems usually combine them into one score per SKU.
| Method | Rule it applies | Best for |
| ABC velocity | Fastest movers closest to dispatch | Stable, high-volume catalogs |
| Cube-per-order index | Least space per pick goes closest | Mixed item sizes |
| Affinity | Items ordered together sit together | Multi-line e-commerce orders |
| Constraint rules | Heavy items low, hazards segregated | Safety and compliance |
Velocity and ABC slotting
ABC slotting ranks SKUs by pick frequency and puts the fastest movers in the most accessible slots. It works because a small share of SKUs usually drives most picks.
How do you implement ABC slotting?
Count order lines per SKU over a recent window, sort descending, and cut by cumulative share of picks. A common split puts the top 80 percent of picks in class A, the next 15 percent in B, and the rest in C. Class A goes into the golden zone, the waist-to-shoulder height where picks are fastest.
The cube-per-order index
The cube-per-order index (COI) is the storage space an item needs divided by how often it is picked. The lowest COI goes closest to dispatch, delivering the most picks per unit of prime space.
Why does COI beat pure velocity?
Under COI, a shoebox-sized item picked 200 times a week outranks a pallet-sized item picked 250 times.
Affinity slotting for items ordered together
Affinity slotting stores SKUs that often share an order side by side, so one trip collects several lines. It pays off most in e-commerce.
How do you find product affinity?
Build a co-occurrence matrix, a table counting how often each SKU pair shares an order, and keep pairs above a threshold. Clustering then turns pairs into zones, and teams that hire Python developers get pandas and scikit-learn working on real order data quickly.
Solving the slot assignment
Once every SKU has a score, slotting becomes an assignment problem: match SKUs to slots so total picks multiplied by distance is as low as possible.
Greedy first or optimizer first?
Start greedy: assign the best-scored SKU to the best free slot and repeat, fast enough to run nightly. A solver such as Google OR-Tools then refines it within weight and hazard constraints. Measure distance along real aisles, since pickers cannot walk through racks.
Deciding when to re-slot
Re-slot only when travel saved exceeds the labor cost of moving stock. Moving every SKU every week burns more hours than it saves.
How do you score a move?
Score each move as weekly travel saved minus one-time move labor, then run only the top moves in a capped nightly batch. Operators that hire data engineers keep that velocity data clean instead of feeding the planner stale spreadsheets.
What it costs to build
Cost depends on whether you need static rules or a learning engine that plans its own moves.
| Scope | What you get | Typical range (USD) |
| Rules-based module | ABC, COI, constraints, slot reports | 12,000 to 30,000 |
| AI slotting engine | Affinity, forecast velocity, re-slot planner | 30,000 to 70,000 |
Where are the savings?
Start with ABC rules, then add affinity and forecasting. This staged build runs at up to forty percent savings versus Western agencies, at roughly twenty-five to forty-nine US dollars per hour.
FAQs
What is slotting in warehouse management?
Slotting is deciding which location each SKU occupies. Good slotting puts fast and frequently co-ordered items where pickers reach them fastest.
How does slotting reduce pick time?
It shortens walking. High-velocity SKUs near dispatch and in the golden zone mean most lines come from the easiest slots.
What algorithm is used for warehouse slotting?
Most engines combine ABC velocity, the cube-per-order index, and affinity analysis, then solve placement with a greedy heuristic or integer programming.
How often should a warehouse re-slot?
When expected travel savings exceed move labor. Small nightly or weekly batches beat a full re-slot.
Can machine learning improve slotting?
Yes. Machine learning forecasts future velocity, so seasonal items reach prime slots before demand peaks, not after.
