People searching for warehouse location management usually need more than a shelf-numbering convention. They need to decide between fixed and random storage, direct inbound loads to valid putaway locations, select inventory under FIFO or FEFO, and reduce stock-count discrepancies. Location management is not merely an address register. It converts item, handling-unit, lot, quality, capacity and process requirements into an executable answer to two questions: where should this load be put, and where should this demand be picked? This guide is written for factories and distribution operations in Thailand and follows the decision from storage strategy through the location master, movement traceability, an RFP, a proposed 90-day PoC and acceptance testing.
1. Executive answer: location management aligns address, inventory and work
A dependable operation keeps three representations aligned: the physical position of the goods, the inventory position recorded by the system and the movement instruction issued to the operator. Shelf labels alone do not prevent searches when movements go unrecorded. A WMS does not repair an ambiguous location master or unclear putaway and allocation policy; it only applies the configured ambiguity faster.
Design in this order:
| Sequence | Decision | Deliverable |
|---|---|---|
| 1 | Physical and process areas | Warehouse zone map |
| 2 | Address granularity and codes | Location master |
| 3 | Item, pack, quality and capacity constraints | Compatibility matrix |
| 4 | Inbound putaway policy | Ranked placement rules |
| 5 | Outbound pick policy | Allocation and picking rules |
| 6 | Movement, count and exception records | Transactions and audit trail |
| 7 | Behaviour on real data | Acceptance-test evidence |
Fixed, random and hybrid are not maturity levels. The right choice depends on SKU behaviour, shipment frequency, handling units, quality controls and shop-floor capability. A practical factory warehouse often fixes high runners and regulated items, uses random storage for variable general stock, and reserves dedicated zones for inspection, hold, returns and staging.
2. What warehouse location management means: from address book to execution control
A location is a storage address composed of a facility, zone, aisle, rack, bay, level and position, or a subset appropriate to the site. Granularity must be small enough for an operator to reach the correct position and for a discrepancy to be isolated, but not so detailed that every transaction becomes burdensome.
The system should answer operational questions, not just display an address:
| Process moment | Question the system must answer |
|---|---|
| Receiving | Where may this load be stored? |
| Replenishment | From which reserve slot to which pick face? |
| Shipping | Which lot, in what order, from which slot? |
| Quality | How are available, inspection, hold and rejected stock separated? |
| Movement | Who moved what, from where to where, and when? |
| Counting | Which boundary is frozen, counted and approved? |
Microsoft describes a location directive in Dynamics 365 as a rule that identifies pick and put locations for inventory movement, covering inbound putaway, outbound pick and stage, production material movement and replenishment. That is a product-specific implementation, but the separation between the address master and the work-selection rule is useful regardless of platform.
3. Fixed, random and hybrid storage: selection conditions
Fixed storage assigns a normal home to an SKU. It is easy to teach and visual deviations are easier to notice, but the reserved space remains unavailable when the item is out of stock. Random storage selects a compatible available slot for each receipt. It shares capacity better, but accurate scanning and real-time inventory records are prerequisites.
| Method | Suitable conditions | Main value | Main caution |
|---|---|---|---|
| Fixed | Stable SKUs, high runners, dedicated or regulated items | Easy to learn and mark | Reserved space can sit empty |
| Random | Volatile assortment and volume, shared racking | Flexible capacity sharing | One unrecorded move creates a search |
| Hybrid | Factory stores, multiple packs, quality segregation | Optimisation by stock characteristic | Rule boundaries and priorities must be explicit |
In a hybrid design, classify by SKU × handling unit × purpose rather than SKU alone. The same component can have random case storage, a fixed open-case pick face, a dedicated inspection area and a line-side supply location.
Before choosing, collect SKU count, daily inbound and outbound lines, days of stock, peak factor, pallet/case/each mix, environmental and dangerous-goods constraints, lot and expiry requirements, forklift and walking routes, and network reliability. Test compatibility and travel with real transactions rather than assuming random storage automatically improves density.
4. Designing the location master: attributes matter more than clever codes
Location codes should be short, unique and difficult to misread on labels and mobile screens. A hierarchical code such as WH1-A03-B12-L02-P01 is intuitive, but codes that embed changeable layout meaning are expensive to maintain. Consider separating a stable identifier from human-readable descriptions and hierarchy fields.
At minimum, maintain these attributes:
| Attribute group | Examples | Design purpose |
|---|---|---|
| Identity | Warehouse, zone, aisle, rack, level, position | Unique address and hierarchy |
| Purpose | Receive, reserve, pick, stage, inspect, quarantine, return | Prevent process mixing |
| Physical limits | Width, depth, height, maximum weight, floor load | Storage compatibility |
| Handling limits | Truck type, approach direction, manual handling | Safe work instruction |
| Item restrictions | Temperature, dangerous goods, family, commingling | Compliance and compatibility |
| Operations | Active/blocked, fixed SKU, priority, count class | Daily control and change management |
| Identification | Barcode, QR, and GS1 identifiers where needed | Accurate scanning |
The GS1 GLN Data Model Solution Standard provides a model for identifying parties and locations. It does not mean every internal bin must receive a GLN. Define which locations must be shared with customers, suppliers or logistics partners and how those identifiers map to internal WMS addresses.
As product-specific comparison examples, Oracle Warehouse Management 26C documents “Active Locations” with settings such as area, aisle, bay, level and barcode; dimensions; pick and putaway sequence; capacity; and restrictions on mixing multiple SKUs or lots. SAP EWM storage bins likewise use attributes such as storage type, section, maximum weight and total capacity. These are product data models, not a common standard. In an RFP, compare how each candidate represents and validates your requirements rather than requiring matching field names.
5. Capacity, handling unit and commingling rules
Random storage frequently fails when capacity is treated as quantity alone. A slot may appear to have one pallet position available but still be invalid because of height, weight, orientation, sprinkler clearance or truck access. Choose measurable capacity dimensions such as volume, mass, pallet positions or cases.
| Constraint | Master or transaction data | Example decision |
|---|---|---|
| Dimensions | Item/pack L×W×H, internal slot dimensions | Height and depth fit |
| Weight | Unit weight, load weight, rated capacity | Within residual load limit |
| Handling unit | Pallet, case, each, license plate | Permitted pack matches |
| Commingling | SKU, lot, expiry, owner, quality | No prohibited combination |
| Facility | Chilled, explosion-safe, locked, charging zone | Required capability exists |
Unit conversion is a control point. If one case contains 12 pieces and one pallet contains 40 cases, while supplier packs vary, preserve effective dates and actual handling-unit identity. Include broken cases, damage, oversized loads and multiple packs in acceptance tests, not only the ideal full-pallet scenario.
6. Putaway rules in an inbound and outbound management system
Putaway is the decision and work that transfers received stock from a dock or receiving buffer into storage. A robust rule does not simply choose the emptiest position. It evaluates safety, quality, handling-unit fit, capacity, demand velocity, replenishment impact and travel in an explicit order.
Separate mandatory filters from preferences:
| Rule class | Example | Treatment |
|---|---|---|
| Mandatory | Quality status, temperature, dangerous goods, load rating | Exclude candidate if any condition fails |
| Strong preference | Fixed home, consolidate SKU, consolidate lot | Rank within valid candidates |
| Efficiency | Distance from receipt, future replenishment travel | Score or sequence |
| Fallback | Overflow or temporary buffer | Alert, owner and expiry time |
Microsoft’s location-directive documentation gives an example that attempts to consolidate with existing stock before seeking an empty location with no incoming work. It also describes using disposition codes to direct received goods to inspection or quarantine. The essential design point is the no-candidate path: do not silently return an arbitrary position. Stop the work, route it to an exception queue or require supervisor approval.
7. Pick design: sequence FIFO, FEFO, distance and pack constraints
Picking selects which inventory fulfils production or customer demand. Shortest distance alone can strand old stock; rigid FIFO can create excess travel and broken-pack work. Apply shipment eligibility and customer conditions first, then stock rotation, pack rules and travel efficiency.
| Priority layer | Typical rule | Exception |
|---|---|---|
| 1. Eligibility | Available, quality passed, owner matches | Hold and quarantine excluded |
| 2. Order condition | Specified lot, origin, minimum remaining life | Contract takes priority |
| 3. Rotation | FIFO or FEFO | Expiry-managed stock generally uses FEFO |
| 4. Pack | Full pallet, case, each | Split or packing-quantity restrictions |
| 5. Efficiency | Zone, route, consolidation | Wave and cut-off considered |
Microsoft describes location-aging FIFO as allocation from the location with the oldest warehouse entry ageing date and documents FEFO batch reservation for batch-enabled items. These are product examples. Your process must separately define whether the controlling date is receipt, quality release, production, manufacture or expiry. Our FIFO inventory management guide for Thailand covers rotation in depth; this article concentrates on how it connects to location selection.
8. Quality quarantine, hold and returns
Real-time quantity is not usable inventory when quality status is wrong. Awaiting inspection, quality hold, nonconforming, return-pending, scrap-pending and customer-owned stock should not share the same logical availability as released stock.
Use both physical and logical segregation where risk requires it. A logical status alone may not prevent an operator from taking the load, while a fence alone may not prevent the ERP or WMS from allocating it. A quality event should update inventory status and recommend a controlled move; the location balance changes only after the scan is completed.
| Status | Suggested location | Allocation | Release authority |
|---|---|---|---|
| Awaiting inspection | QA-IN | Blocked | Quality operator |
| Released | Reserve/pick | Allowed | Approved workflow |
| Hold | QA-HOLD | Blocked | Quality owner |
| Nonconforming | NCR/LOCK | Blocked | Named approver |
| Return pending | RETURN | Blocked | Quality + logistics |
If emergency temporary placement is allowed, make the temporary location, owner, resolution deadline and overdue alert mandatory. Otherwise “temporary” floor stock becomes a permanent break between the system and the physical warehouse.
9. Movement traceability: transactions behind real-time inventory
Inventory visibility begins before the dashboard. Every physical move needs a record of source, destination, item, quantity, unit, lot/serial, handling-unit ID, operator, device, timestamp and reason. If back-dated entry is allowed, distinguish physical-event time from recording time.
GS1 EPCIS 2.0.1 provides a standard model for sharing visibility events—what happened, when, where, why and under which business context. Not every internal bin move needs to be published through EPCIS, but its event dimensions are useful for checking whether an internal movement record is complete.
| Event | Required evidence | Audit question |
|---|---|---|
| Receive | Source, handling unit, quantity, buffer | Match to PO/ASN? |
| Putaway | Source, target, suggested and actual location | Why was it overridden? |
| Move | Source, target, reason | Was the move authorised? |
| Pick | Demand, lot, quantity, source | Was allocation policy followed? |
| Count | Book, physical, variance, recount | Who approved the correction? |
Scanning controls sequence as well as data entry. A source-location, load and destination-location check is a sound default. Test offline queues, duplicate transmission, reversal and operator identity on shared devices.
10. Inventory visualisation: separate balance, capability and exception
An inventory dashboard should distinguish available, allocated, inspection, hold, in-transit, near-expiry and missing stock. A location view should include expected inbound and outbound work, not just current balance. A slot that is empty now may already be committed to an incoming pallet.
Separate executive, supervisor and operator views:
| Layer | Example indicators | Decision supported |
|---|---|---|
| Executive | Inventory value, write-off, delivery adherence, storage cost | Investment and capacity |
| Supervisor | Occupancy, putaway override, replenishment shortage, count variance | Daily rebalancing |
| Operator | Next work, exception, block, rescan | Immediate execution |
Do not make 100% occupancy the objective. It removes room for peak receipts, controlled movement, quarantine and safety clearances. Define targets by zone using operational headroom and peak conditions; do not present one universal percentage as an industry standard.
11. Cycle counting and variance: control movement before counting
Count variance originates not only in counting technique but also in movement during the count, unit conversion, unfinished work, temporary placement, damage, incorrect lot and production postings. For cycle counting, define the location boundary, movement control, blind-count option, recount, variance approval and reason codes.
| Step | Control | Purpose |
|---|---|---|
| Freeze/boundary | Control work entering the location | Prevent double movement |
| Count | Option to hide book quantity | Reduce confirmation bias |
| Recount | Threshold and independent operator | Separate entry error |
| Approve | Value/quantity authority | Prevent unauthorised adjustment |
| Correct | Update stock, lot and location together | Restore integrity |
| Analyse | Cause code and corrective action | Prevent recurrence |
Our guide to reducing inventory count time provides the detailed counting process. Here, the acceptance condition is that movement boundaries and variance approval can be controlled at location level.
12. Devices and labels: a correct master fails if the code cannot be read
Place location labels for the actual scan distance, lighting, contamination and reflection. Show human-readable text beside the machine-readable symbol and avoid layouts that make adjacent slots easy to confuse. The GS1 Logistic Label Guideline is a primary reference for logistics-unit labels and barcodes, but a storage-location label is not the same object. Define the item being identified, key and scan step separately.
Device requirements include gloves, Thai and English display, offline behaviour, camera or scan engine, drop resistance, battery and Wi-Fi roaming. Guide the operator through source → load → destination rather than presenting unrestricted item, quantity and location fields. A rule violation should require a controlled reason.
Monitor scan failure, manual entry and reprint rates. Do not allow warehouse familiarity to hide deteriorating labels.
The GS1 Serial Shipping Container Code (SSCC) identifies a logistics unit such as a pallet or case. It is not a shelf or bin location identifier. Scan the SSCC on the load and the WMS location code on the storage position as separate identifiers, then link them in the movement transaction.
13. Rule priority and exceptions: explainability before optimisation
Putaway and pick rules conflict as they accumulate. A single opaque formula combining dangerous goods, quality, fixed home, capacity, consolidation, expiry and distance is hard to support. Divide logic into mandatory filter, priority, score and fallback.
| Stage | Example | Evidence to retain |
|---|---|---|
| Filter | Quality, temperature, load rating | Exclusion reason |
| Priority | Fixed home, same-SKU consolidation | Applied rule ID |
| Score | Distance, occupancy, replenishment impact | Candidate set and score |
| Fallback | Overflow or supervisor selection | Approver and expiry |
Microsoft’s location-directive model includes conditions, ordered actions, failure behaviour and acceptance tests under different starting conditions. Regardless of product, require representative regression tests before a rule change is released.
14. Dynamic placement: how to read Microsoft 2026 release wave 1
Microsoft’s 2026 release wave 1 plan describes dynamic item placement with preferred storage locations and target quantities by item, balancing during inbound flow, automated replenishment and spatial locations intended to improve routing. The page shows public preview on 5 June 2026 and general availability planned for October 2026.
The same page explicitly warns that some planned functionality has not been released, delivery timing may change and projected functionality may not be delivered. As of this article date, 9 September 2026, treat it as release-plan and preview information—not as a generally available capability or a guaranteed outcome. Recheck tenant, geography, language, licence, preview conditions and GA status at the point of decision.
Dynamic placement does not eliminate master data. It makes the separation more important: immutable safety, quality and physical constraints versus target quantities and preferences that may be optimised. As the algorithm changes placement, retain the recommendation rationale, manual override, history and regression evidence.
15. RFP requirements: compare scenarios and evidence, not checkboxes
Asking “random location supported?” or “real-time inventory available?” produces an unhelpful yes/no comparison. Use your own item, handling-unit, quality and exception scenarios and state the evidence each bidder must show.
| Requirement | Example RFP question | Evidence |
|---|---|---|
| Master | How are dimensions, weight, commingling and facility restrictions represented? | Configuration and data dictionary |
| Putaway | What happens when no compatible candidate exists? | Demonstration and log |
| Pick | Can FIFO/FEFO, pack and customer rules be sequenced? | Rule configuration and tests |
| Quality | How is held stock blocked physically and logically? | State transition |
| Movement | How are offline, duplicate and reversal handled? | Failure test |
| Count | How are work conflict, recount and variance approval managed? | UAT scenario |
| Operations | Who approves and rolls back a rule change? | Audit log |
Our Thailand WMS implementation guide covers the broader programme. This RFP section is intentionally limited to location placement and execution control so it does not duplicate a general WMS selection article.
16. A proposed 90-day PoC: test the system and the physical operation together
Ninety days is a TOMAS TECH planning example, not an external standard or a performance promise. More important than the duration is fixing the pilot zone, SKU set, handling units, shifts and exception cases and completing the path from baseline to acceptance.
| Example period | Main work | Exit condition |
|---|---|---|
| Days 1–15 | Baseline, zone map, SKU/pack/quality analysis | Scope and baseline approved |
| Days 16–30 | Master, labels and device verification | Sample area scans reliably |
| Days 31–50 | Putaway, pick, move and quarantine rules | Scenario unit tests pass |
| Days 51–70 | Physical migration, training, limited operation | Variance and exceptions recorded |
| Days 71–85 | Peak, failure, count and regression tests | Critical failures closed |
| Days 86–90 | Results, residual risks and rollout decision | Go, conditional go or stop |
Do not predeclare an improvement percentage or ROI. “Reduce search time by 30%” may be a test hypothesis, but it is meaningless without the same baseline definition, sample, volume and training conditions. Measure search time, putaway overrides, mispicks, count variance, incomplete moves and travel on the same basis before and during the PoC.
17. Acceptance tests: boundary values and failure behaviour matter
Testing one ideal receipt and shipment is insufficient. Include full slots, residual load limits, prohibited commingling, expiry inversion, quality hold, network interruption, duplicate scan, partial pack, incorrect label and concurrent work.
| Test group | Scenario | Expected behaviour |
|---|---|---|
| Putaway | No compatible location | Stop and notify; do not invent a slot |
| Capacity | Quantity fits but weight exceeds limit | Exclude with reason |
| Quality | Demand against a held lot | Reject allocation and pick |
| FIFO/FEFO | Old inventory is farther away | Follow approved priority |
| Movement | Destination scanned without source | Cannot complete |
| Offline | Same load processed by two devices | No duplicate posting on reconnect |
| Count | Replenishment arrives during count | Controlled hold or conflict workflow |
| Audit | Administrator changes a rule | Record before/after, reason and approver |
Acceptance thresholds are operation-specific. Never let an average score offset a critical violation such as allocating held stock or exceeding rack load. Record the owner, due date and retest condition for every open exception.
FAQ: Does random storage always improve space utilisation?
No. It increases the opportunity to share capacity, but dimensions, weight, commingling rules, unrecorded moves, replenishment work and travel can increase total effort. Compare fixed, random and hybrid on your SKU and peak-volume data.
FAQ: Is RFID mandatory for real-time inventory management?
No. Barcode-controlled moves can provide timely records if source, load and destination are captured at execution and offline and duplicate behaviour is controlled. Evaluate RFID for read zone, tag cost, metal/liquid environment and false reads.
FAQ: How is an inbound/outbound system different from a WMS?
Names are inconsistent. Compare requirements: location-level inventory, putaway and pick work, replenishment, quality status, counting, operational logs and audit controls—not the product label.
FAQ: Should a warehouse use FIFO or FEFO?
Expiry-controlled stock may require FEFO, while non-expiring stock may use FIFO to prevent ageing. Customer conditions, remaining shelf life, lot and quality eligibility must be resolved before the rotation preference.
FAQ: How should location-management ROI be estimated?
Measure current search, travel, replenishment, counting, mis-shipment, write-off and downtime. Treat improvement percentages as TOMAS TECH calculation scenarios, not industry benchmarks or guarantees, and validate them in the PoC.
18. Summary: turn location management from shelf codes into an acceptable process
Warehouse location management is more than selecting fixed, random or hybrid storage. Give the location master purpose, capacity, pack and quality attributes; sequence mandatory and preferred putaway and pick logic; and record every physical movement. Connect FIFO/FEFO, quarantine and cycle counting to the same address, status and event model so inventory visibility agrees with shop-floor execution.
For implementation, prefer scenario-based RFP evidence, a constrained 90-day PoC and boundary-focused acceptance tests over a long feature checklist. Treat emerging capabilities such as Microsoft’s dynamic placement according to their status at the point of decision, and keep non-negotiable safety and quality constraints separate from optimisable targets.
TOMAS TECH can help factories and warehouses in Thailand assess current operations, design the location master, compare fixed/random/hybrid storage, integrate WMS and ERP, and define a PoC and acceptance tests. You can contact us while the question is still “How should we redesign our Excel register and shelf labels?” or “What should the WMS RFP require?”
References
- Microsoft Learn, Work with location directives: https://learn.microsoft.com/en-us/dynamics365/supply-chain/warehousing/create-location-directive
- Microsoft Learn, Enhance warehouse efficiency with dynamic item placement (2026 release wave 1): https://learn.microsoft.com/en-us/dynamics365/release-plan/2026wave1/enterprise-resource-planning/dynamics365-supply-chain-management/enhance-warehouse-efficiency-dynamic-item-placement
- Oracle, Warehouse Management 26C, Active Locations: https://docs.oracle.com/en/cloud/saas/warehouse-management/26c/owmim/active-locations.html
- SAP Help Portal, warehouse location/product master reference: https://help.sap.com/docs/SAP_SUPPLY_CHAIN_MANAGEMENT/57574d15fa1d414792d74047b66c3e41/55c8cb53ad377114e10000000a174cb4.html
- GS1, GLN Data Model Solution Standard: https://www.gs1.org/docs/barcodes/GS1_GLN_DataModelSolution_Standard.pdf
- GS1, EPCIS and CBV 2.0.1: https://ref.gs1.org/standards/epcis/2.0.1/
- GS1, GS1 Logistic Label Guideline 1.3: https://www.gs1.org/standards/gs1-logistic-label-guideline/1-3
This guide reflects public information checked on 9 September 2026. Reconfirm product functions, preview/GA dates, licensing and legal requirements at the point of decision. Ninety days, improvement rates, ROI and thresholds are TOMAS TECH planning or calculation examples, not external standards or guaranteed results.
