# Class-Based Storage

> Site directory and reading guidance: [llms.txt](https://wiki.skusavvy.com/llms.txt).

Canonical URL: https://wiki.skusavvy.com/article/class-based-storage
Published: 2026-11-11T10:09:18.050Z
Last updated: 2026-11-11T10:30:08.270Z
Category: Warehousing

> This article is AI-assisted reference material. Verify important or time-sensitive claims against the cited sources; AI review does not establish independent factual verification.

## Overview

**Class-based storage** is a warehouse location policy that divides stock-keeping units (SKUs) into distinct classes and assigns each class to a designated zone or set of storage locations. A common approach places high-velocity items in prime, highly accessible positions (such as near input/output points or at ergonomic pick heights) and slower-moving items farther away. Within an assigned class zone, individual items may occupy any suitable open slot under a shared-storage rule, such as randomized or closest-open-location assignment, rather than a strictly fixed slot for each SKU.

The policy occupies a middle ground between fully dedicated storage (where every SKU has a fixed, reserved location) and purely randomized storage (where incoming units may be assigned to any open location in the facility). Class-based storage is intended to reduce travel time relative to random storage while using less reserved space than dedicated storage, a comparison developed in the [order-picking research literature](https://doi.org/10.1016/j.ejor.2006.07.009) and in analytical models of [storage assignment](https://doi.org/10.1287/mnsc.22.6.629). In practice, the policy depends on stable [warehouse bin locations](https://www.skusavvy.com/blog/how-to-set-up-bin-locations-for-warehouse-efficiency) and accurate location-level inventory records. Textbook treatments of the same three policies—dedicated, randomized, and class-based—appear in [Warehouse & Distribution Science](https://www.warehouse-science.com/).

## Terminology and Origins

The term *class-based storage* comes from operations research and warehouse engineering. It combines **class**, a group of items that share a defined attribute, with physical **storage**. Implementations often use two, three, or more zones. Three zones labeled A, B, and C are common because they borrow the familiar tripartite structure of ABC analysis, but the number of classes is a design choice, not a fixed rule.

ABC inventory classification was popularized in mid-20th-century production management by adapting Vilfredo Pareto’s observation about unequal distributions. In purchasing and inventory control, the usual ranking metric is value, such as annual usage value. In storage assignment, the ranking more often uses handling activity: order lines, pick frequency, cube-per-order, or another measure of retrieval effort. Those are related ideas, not the same policy. Research on [warehouse storage policies](https://doi.org/10.1287/mnsc.22.6.629) treated class-based assignment as a distinct archetype beside dedicated (full turnover-based) and randomized assignment.

## Historical Development

Storage-position rules developed alongside analytical models of material handling and, later, computerized warehouse control:

- **1963:** James L. Heskett published the [Cube-Per-Order Index (COI)](https://doi.org/10.1080/05695556308978149), the ratio of an item’s storage-space requirement to its order frequency. The COI rule places items that use little space per order handled closer to input/output points. It is a foundational slotting index, not itself a complete class-based zoning procedure.
- **1976:** Warren H. Hausman, Leroy B. Schwarz, and Stephen C. Graves published [“Optimal Storage Assignment in Automatic Warehousing Systems”](https://doi.org/10.1287/mnsc.22.6.629). For a square-in-time automated storage and retrieval system (AS/RS), they compared randomized storage, full turnover-based dedicated storage, and class-based storage. Their models showed that a small number of classes, including a three-class policy, can capture much of the expected travel-time improvement of full turnover-based assignment while requiring less space than reserving a location for every item.
- **1980s–1990s:** As materials-requirements planning, and later systems sold as enterprise resource planning, spread, storage rules were applied beyond the single-aisle AS/RS models of the 1970s to multi-aisle and conventional rack warehouses. Results from those models do not transfer unchanged to every manual layout.
- **2007:** René de Koster, Tho Le-Duc, and Kees Jan Roodbergen published a [review of order-picking design and control](https://doi.org/10.1016/j.ejor.2006.07.009). It placed storage assignment beside batching, zoning, and routing methods such as S-shape, midpoint, and largest-gap heuristics, and noted that storage policy and routing interact.
- **2010s onward:** Warehouse-management systems made it practical to recompute velocity classes from order history and to direct putaway into eligible zones. [Warehouse putaway workflows](https://www.skusavvy.com/docs/guides/warehouse/putaway) and [warehouse mapping software](https://www.skusavvy.com/warehouse-mapping-software) can support that execution; how often classes should be rebuilt remains an operating decision, because relocation labor can offset travel savings.

A later survey of warehouse design research, [Gu, Goetschalckx, and McGinnis (2010)](https://doi.org/10.1016/j.ejor.2009.07.031), reviews storage-assignment models together with layout, batching, and performance evaluation, and is a useful check on claims that were first proved only for AS/RS.

## Core Concepts

### Classification Metrics and Zone Assignment

A class-based system needs an explicit ranking metric and explicit zone boundaries. Common metrics include:

- **Pick frequency:** Count of order-line hits, or picks, over a stated history window.
- **Cube-movement or COI:** Frequency adjusted by the space the item occupies, so bulky fast movers do not automatically claim the entire prime pick face.
- **Handling effort:** Adjustments for equipment, weight, or retrieval difficulty, used when raw hit counts would mis-rank the work.

SKUs are sorted on the chosen metric and split at cutoff points into classes. Locations are partitioned by travel accessibility from receiving, packing, or other input/output points. Class A is assigned to the most accessible zone, later classes to progressively less accessible zones. [Layout choices](https://www.skusavvy.com/blog/optimize-warehouse-layout)—aisle orientation, cross-aisles, and forward-pick capacity—determine whether those zones can actually absorb the assigned volume.

The shares below are planning heuristics in the Pareto tradition, not physical laws. Real cutoffs should be set from the facility’s own order and cube data.

| Class | Often-cited SKU share | Often-cited activity share | Typical location characteristics |
| :--- | :--- | :--- | :--- |
| **Class A** | about 10%–20% | about 60%–80% | Near pack or I/O points; floor or waist height; short travel |
| **Class B** | about 20%–30% | about 15%–25% | Intermediate aisles and rack levels |
| **Class C** | about 50%–70% | about 5%–15% | Farther aisles, upper levels, or reserve storage |

### Class Pools versus Dedicated Slotting

Dedicated slotting reserves a location for an SKU even when that SKU is out of stock. Class-based storage treats each zone as a shared pool: any SKU in the class may use any eligible open location in the zone. During receiving, [bin putaway criteria](https://www.skusavvy.com/docs/guides/warehouse/bin-putaway-criteria) restrict the search to that pool. Sharing locations reduces the amount of space held empty for absent items, at the cost of less predictable item positions inside the zone.

Many warehouses also separate a forward pick face from bulk reserve. [Stock replenishment](https://www.skusavvy.com/docs/guides/warehouse/stock-replenishment) then moves units from reserve into the forward class zone. Class-based rules can apply to either tier, or to both, and the forward face is often sized for a limited number of days of demand.

Full turnover-based dedicated storage is the limiting case in which each item is effectively its own class, ranked by turnover and given a fixed region. Randomized storage is the opposite limit: one class covering the whole warehouse.

## Types and Configurations

- **Velocity-based (fast-mover / slow-mover):** Ranks goods by pick activity. Class A may be placed nearest the main cross-aisle, in the “golden zone” between about waist and shoulder height, or both.
- **Space-adjusted (COI-based):** Divides required storage space by turnover so that small, frequently picked items outrank large items with the same hit count.
- **Constraint-first, then class:** Temperature, hazardous materials, security, weight, or lot/expiry rules define legal areas first. Velocity classes are then formed only inside each compliant area.
- **Forward-reserve hybrid:** A class-based or dedicated forward pick area, such as carton flow or low-level shelving, is supplied from a high-bay reserve that may itself be class-based or closer to random pallet storage.

## Travel-Time Comparison

In the square-in-time, single-command AS/RS model of [Hausman, Schwarz, and Graves (1976)](https://doi.org/10.1287/mnsc.22.6.629), expected crane travel falls as storage becomes more turnover-based. A few classes capture a large part of that reduction relative to random storage, while using less space than full dedicated assignment. The exact percentage depends on the demand curve, the number of classes, zone boundaries, and the I/O location. It should not be quoted as a universal 70–80 percent saving for every warehouse.

In manual order picking, class-based zoning can shorten the distance walked on popular routes, but the gain depends on the routing method, order batch size, and whether fast movers are clustered so tightly that pickers block one another. [Batch picking](https://www.skusavvy.com/batch-picking-software) can multiply picks per trip inside a dense zone; it can also increase congestion if too many pickers share the same Class A bays. The [2007 design-and-control review](https://doi.org/10.1016/j.ejor.2006.07.009) treats storage, batching, and routing as a joint design problem rather than as independent levers.

## Benefits and Operational Trade-Offs

### Benefits

- **Shorter retrieval travel, under the right demand skew:** When a small share of SKUs generates most picks, concentrating those SKUs near I/O points reduces average travel compared with storing them at random.
- **Better space use than dedicated storage:** Open locations in a class zone can be used by other SKUs in that class, so space is not permanently reserved for every out-of-stock item.
- **Ergonomic placement:** High-frequency items can be kept in a reachable band, reducing bending and ladder use, provided weight and beam ratings still govern the final level.

### Limitations and Trade-Offs

- **Aisle congestion:** Putting most activity into a few bays can create picker interference. Waiting can erase the travel-time gain that the storage model predicted under the assumption of a single crane or an uncongested aisle.
- **Re-slotting labor:** Seasonal demand and promotions change class membership. Moving stock with [bin-to-bin transfers](https://www.skusavvy.com/docs/guides/warehouse/bin-to-bin-inventory-transfer) is justified only when the expected picking saving exceeds the move cost.
- **Replenishment pressure:** Smaller Class A pick faces shorten travel but increase the number of replenishment trips. Forward quantity and [replenishment triggers](https://www.skusavvy.com/docs/guides/warehouse/stock-replenishment) have to be set together with the class boundaries.
- **Physical and regulatory constraints:** Beam capacity, clearance, and secure storage are not optional ranking factors. [OSHA 29 CFR 1910.176](https://www.osha.gov/laws-regs/regulations/standardnumber/1910/1910.176) covers general handling, secure storage, and aisle clearance. Flammable liquids and other regulated materials are governed by separate standards, including [29 CFR 1910.106](https://www.osha.gov/laws-regs/regulations/standardnumber/1910/1910.106), and cannot be slotted into a “golden zone” merely because they are fast movers.
- **Correlated demand and inaccurate history:** Models that rank items independently can understate travel when items that are ordered together are split across distant classes. Short or promotion-distorted history can also mis-classify SKUs.

## Implementation

1. **Profile demand and space together.** Use a stated history window, exclude one-off promotional spikes or tag them separately, and rank by the metric the operation actually wants to optimize (hits, cube, or COI). Record baseline travel, replenishment trips, and location utilization before changing zones.
2. **Fix the location system first.** Class pools fail if bins are unlabeled or inventory is not scanned to a location. Barcoded locations and [inventory scanning](https://www.skusavvy.com/blog/inventory-scanning) are prerequisites, not optional extras.
3. **Size each zone to its class.** Class A locations need enough forward quantity for the replenishment cycle. A zone that is too small forces constant overflow; a zone that is too large gives back the travel reduction.
4. **Spread activity to limit congestion.** Place Class A stock along the pick path or across several aisles rather than in one bay, and put heavy items on appropriate beams even if their velocity would otherwise suggest waist height.
5. **Enforce putaway in software.** Direct receivers only to locations that are both open and legal for the SKU’s class and constraints, using predefined [putaway criteria](https://www.skusavvy.com/docs/guides/warehouse/bin-putaway-criteria).
6. **Reclassify on a schedule, not continuously by default.** [Cycle counting](https://www.skusavvy.com/docs/guides/warehouse/cycle-counting) protects location accuracy. Velocity reviews should move an SKU only when the class change is persistent enough to repay the transfer.

## See Also

Dedicated storage, randomized storage, cube-per-order index, ABC analysis, forward-reserve storage, and order-picking routing are the adjacent topics most often confused with class-based storage. Class-based storage is a location policy. It does not, by itself, set reorder points, batch sizes, or pick-path algorithms.


## Source links

- [Design and Control of Warehouse Order Picking: A Literature Review](<https://doi.org/10.1016/j.ejor.2006.07.009>)
- [Optimal Storage Assignment in Automatic Warehousing Systems](<https://doi.org/10.1287/mnsc.22.6.629>)
- [How to Set Up Bin Locations for Warehouse Efficiency](<https://www.skusavvy.com/blog/how-to-set-up-bin-locations-for-warehouse-efficiency>)
- [Warehouse Putaway Guide](<https://www.skusavvy.com/docs/guides/warehouse/putaway>)
- [Optimize Warehouse Layout](<https://www.skusavvy.com/blog/optimize-warehouse-layout>)
- [Stock Replenishment Guide](<https://www.skusavvy.com/docs/guides/warehouse/stock-replenishment>)
- [Bin Put-Away Criteria](<https://www.skusavvy.com/docs/guides/warehouse/bin-putaway-criteria>)
- [OSHA 29 CFR 1910.176: Handling Materials—General](<https://www.osha.gov/laws-regs/regulations/standardnumber/1910/1910.176>)
- [Inventory Scanning](<https://www.skusavvy.com/blog/inventory-scanning>)
- [Cycle Counting Guide](<https://www.skusavvy.com/docs/guides/warehouse/cycle-counting>)
- [Cube-Per-Order Index—A Key to Warehouse Stock Location](<https://doi.org/10.1080/05695556308978149>)
- [Warehouse Mapping Software](<https://www.skusavvy.com/warehouse-mapping-software>)
- [Batch Picking Software](<https://www.skusavvy.com/batch-picking-software>)
- [Bin-to-Bin Inventory Transfers](<https://www.skusavvy.com/docs/guides/warehouse/bin-to-bin-inventory-transfer>)
- [Warehouse & Distribution Science](<https://www.warehouse-science.com/>)
- [Research on Warehouse Design and Performance Evaluation: A Comprehensive Review](<https://doi.org/10.1016/j.ejor.2009.07.031>)
- [OSHA 29 CFR 1910.106: Flammable Liquids](<https://www.osha.gov/laws-regs/regulations/standardnumber/1910/1910.106>)