Fulfillment Process: How Does an Order Reach the Customer?

Order fulfillment is the set of physical and digital steps through which an e-commerce order is validated, prepared in the warehouse, handed over to a carrier and updated with status information. When a customer clicks “Place order,” the product has not yet begun moving through the warehouse. The order data must first be received from the sales channel, validated against business rules and matched with available inventory. Picking, control, packing, labeling and carrier handoff then take place inside the warehouse. The journey to the end customer continues separately within the carrier’s network.

Fulfillment therefore cannot be reduced to “pick, pack and ship.” The commercial order, warehouse work order, physical parcel and carrier delivery are related but distinct records. In a well-controlled operation, every step has a defined entry condition, proof of completion, owner and exception path.

Where does the fulfillment process begin and end?

ASCM’s SCOR process model treats Order, Fulfill and Return as separate processes. Fulfill covers the activities that execute a customer order, including delivery planning, picking, packing and shipping. Return is a separate flow in which a product moves back from the customer. This distinction shows that fulfillment extends beyond warehouse picking, but it does not place every delivery and return activity under the control of the same party. ASCM SCOR process model

In practice, three connected timelines exist:

1. Commerce system: The order is created, then validated or held according to payment, risk, cancellation, product and address rules.
2. Warehouse operation: A processable order enters the WMS; inventory is allocated, picked, checked, prepared and handed over to the carrier.
3. Carrier network: The shipment receives its first acceptance scan, then proceeds through transfer, delivery or delivery-exception events.

Confusing these boundaries also weakens customer communication. “Label created” does not mean that the carrier has received the parcel, and “handed over to the carrier” does not mean that it has reached the end customer. The fulfillment provider’s performance and the carrier’s in-network performance should be measured separately and evaluated together when considering the end-to-end outcome.

What are the fulfillment process steps, and how should an end-to-end status map be built?

Status names may differ across an OMS, WMS, marketplace and carrier system. The objective is not to display a large number of statuses, but to ensure that every status represents a real business event supported by evidence. A general status chain might look like this:

Order created → received by system → validated or placed on hold → inventory allocated → released to operations → picking → pick confirmed → packed → label created → ready for dispatch → handed over to carrier → carrier acceptance scan → in transit → delivered or exception

This chain contains four objects that may look similar but are not the same:

• Customer order: Contains the products purchased, quantities, payment information and delivery preference.
• Fulfillment work order: Defines which order lines must be prepared, from which locations and under which operating rules.
• Shipment or parcel: The physical unit handed over to the carrier; one order may be split across several parcels.
• Delivery record: Reports carrier acceptance, transfer, out-for-delivery and final delivery outcomes.

One order may be fulfilled from multiple locations, split into partial shipments or assigned to different delivery methods. Shopify’s current order-management model likewise shows that a single customer order can contain multiple fulfillment orders linked to different locations or delivery methods. This platform-specific model is not mandatory for every system, but it clearly demonstrates why “one order equals one parcel” is not a safe assumption. Shopify order-management apps

Process prerequisites: receiving, product master data and available inventory

Order fulfillment is triggered by a customer order, but it can operate reliably only if products have already entered the system with the correct identity, quantity and inventory status. If that preparation is incomplete, every downstream step will simply accelerate the existing error.

The preparation layer should include at least the following:

• SKU, variant, barcode, dimensions, weight and, where required, serial, lot or expiration data
• Pallet, carton and unit conversions for the product
• The counting unit and tolerances applied at receiving
• Quarantine rules for damaged, missing, excess or unidentified products
• Pallet or shelf location and the product’s physical availability
• Packing requirements, ready-to-ship cartons and special protection rules
• Permitted carrier, service and label information
• The client’s decision matrix for returned products

At receiving, the physical load is compared with the advance document or declared quantity. If a discrepancy exists, its cause and decision owner should be established instead of releasing the goods directly into “available for sale” inventory. Once put-away is complete, the system record and physical location must point to the same product.

On-hand, available and allocated inventory are not the same. A product may be physically present in the warehouse but damaged, quarantined, committed to another order or unavailable to the sales channel. Shopify’s official inventory states distinguish between on-hand, available, committed and unavailable inventory for this reason. Terms may vary by system, but the quantity that can be assigned to a new order must be defined clearly. Shopify inventory states

For inventory policies and counting methods, see our inventory management guide. For physical layout and slotting, see our warehouse layout guide.

How is an order received and validated in the WMS?

An order may arrive through an API, marketplace connection, storefront integration, file transfer or controlled manual entry. Regardless of the channel, the WMS must avoid creating the same order repeatedly and must apply subsequent updates to the correct record.

The first validation layer should answer the following questions:

• Have the external order number and line identifiers already been processed?
• Is the order canceled, on hold or under risk review?
• Does every order line map to a warehouse SKU?
• Are quantity, unit of measure and variant details unambiguous?
• Are the recipient name, address, telephone number and required country or postal-code fields sufficient?
• Are the selected delivery method and carrier rule defined?
• Are there additional instructions for packing, a gift note, serial or lot handling, or a channel-specific label?

Payment approval, fraud checks and order changes are not managed by the same system in every business. There is therefore no universal rule that “an order enters the warehouse automatically as soon as it is created” or that “no order can be processed before payment is received.” The brand, sales channel and fulfillment provider must agree explicitly on which status opens a warehouse work order.

Address normalization can reduce spelling and formatting problems, but it cannot guarantee that the recipient will be present or that the carrier will complete delivery. The original customer-entered data, corrected data, source of the change and any condition requiring approval should remain recorded.

For technical setup, retries, webhooks and status-mapping details, see our e-commerce integration guide.

How should inventory allocation and release to operations be planned?

The next step for a validated order is to reserve suitable inventory and turn the order into warehouse work at the appropriate time. Allocation may consider not only total stock quantity but also location, inventory status, serial or lot rules, expiration date, product condition and commitments already made.

When several fulfillment locations are available, the system may evaluate:

• Whether available inventory can complete the order lines at the same location
• The carrier’s service area and the customer’s delivery preference
• The cost and customer-experience effects of a split shipment
• The location’s workload and product-handling capability
• Storage or documentation requirements specific to the product category

Once inventory is allocated, the order may be released immediately, at scheduled intervals or as part of a group. Wave planning groups selected orders for processing within approximately the same time window, whereas continuous release places an eligible order into the queue without waiting for a wave. No single method is correct for every order profile. Carrier cutoff times, collection plans, order priority, product proximity and workforce capacity should be evaluated together.

Marking one order as “priority” should not cause other orders to age invisibly. Queue age, reason for waiting and planned dispatch time should be visible together. Any manual priority change should record who made it and why.

How do picking and barcode validation work?

The purpose of picking is not merely to send an employee to a shelf. It is to move the correct item and quantity into the dispatch flow against a traceable work order. Order-based, batch, zone, pick-and-pass or automated methods may be used. The choice depends on lines per order, SKU variety, product size, velocity distribution, aisle layout and peak-volume profile.

Order-picking literature shows that performance is shaped by storage policy, batching, zoning, routing and work sequencing together. A method used by another warehouse should therefore not be copied without testing it against the actual order profile. De Koster, Le-Duc and Roodbergen, Van Gils et al.

Control points during picking may include:

1. The operator is directed to the correct location.
2. The location and product barcode are matched against the work order.
3. Quantity—and, where required, serial, lot or expiration data—is verified.
4. The item is assigned to the order tote, cart or consolidation area.
5. A shortage, missing item, damage or barcode mismatch is not closed as a normal completion; it enters an exception flow.

A barcode helps connect physical movement with the system record; it is not a guarantee of zero errors. If master data are incorrect, a barcode is duplicated, a label is wrong, a scan is skipped or an uncontrolled manual override is used, the system may process an incorrect record quickly. GS1’s traceability approach emphasizes that identification, critical tracking events and related master data must be established together. GS1 Global Traceability Standard

Why are consolidation and pre-dispatch quality control separate steps?

When batch or zone picking is used, items for several orders may be collected during the same trip. Picking the correct product from the shelf does not, by itself, prove that it entered the correct customer parcel. The consolidation step regroups the picked items by order.

Pre-dispatch control may cover:

• Matching order line, SKU, variant and quantity
• Compliance with serial, lot or expiration rules where applicable
• Visible product damage or missing components
• Whether a ready-to-ship carton remains suitable for transport
• Notes, documents or channel requirements linked to the order
• Parcel-to-content relationships when an order is split across several packages

The control method should reflect product risk. Fully recounting every order may create unnecessary repetition in some operations, while performing no independent check may allow an incorrect item to move into the next stage. Barcode validation, weight tolerance, photography or a second-person check can be designed according to risk and cost.

When an error is found, the parcel should not merely be corrected. The stage and cause of the error should also be recorded. If a total labeled “incorrect shipment” does not distinguish a location error from an order-sorting error, lasting improvement is not possible.

How should packing rules and shipping documents be managed?

Packing is not the same operation for every product. If a product is already in a sealed, structurally sound sales carton suitable for transport, unnecessary repacking may be avoided. For products that require protection, packaging should be selected according to product dimensions, weight, fragility, companion items, carrier restrictions and the brand’s presentation rules.

Packing instructions should answer at least the following questions:

• Is the existing ready-to-ship carton acceptable?
• Which box or mailer range should be used?
• When are void fill, tape, tamper-evident tape or corner protectors required?
• Are there products that cannot be shipped together?
• Must a dispatch note, invoice, promotional insert or customer note be included?
• What evidence must be retained before the parcel is sealed?
• At what point are actual weight and dimensions updated?

Document and label data must come from the same order version as the physical parcel contents. If an old pick list or old label remains in use after an order changes, two records that each appear correct in the system can be combined incorrectly in the physical flow.

E-commerce warehouses may process a growing share of small, time-sensitive orders, but every brand has a different product and order profile. Research by Boysen, Weidinger and de Koster likewise shows that warehouse-design decisions must be evaluated together with order structure and service expectations. Warehousing in the e-commerce era

How should labels, the dispatch area and carrier handoff be separated?

A carrier label connects the shipment number, service, recipient, address and routing data to the physical parcel. Creating the label proves only that the data were generated. Applying it to the correct parcel, keeping it readable, staging it under the correct carrier and recording the physical handoff are separate controls.

The following errors should be monitored particularly closely at the labeling stage:

• Duplicate or canceled labels for the same order
• A label belonging to another order applied to the parcel
• Address or service data rejected by the carrier
• A barcode that is creased, covered or unreadable
• Several parcels incorrectly associated with a single tracking number
• A label created even though the parcel never reaches the dispatch area

The GS1 Logistic Label uses an SSCC to identify logistics units uniquely and helps connect physical movement with electronic messages. However, an SSCC and a parcel tracking number issued by a carrier are not the same concept, and it should not be assumed that every B2C parcel must use an SSCC. The applicable label standard should be selected according to the actual processes of the trading partners and carrier. GS1 Logistic Label Guideline

Parcels ready for dispatch may be staged by carrier, service, collection run or manifest. Physical handoff should be closed with a loading scan, handover list, manifest, carrier acceptance record or another form of proof defined in the contract. Marking an order as “handed over to the carrier” before that event occurs makes the gap between warehouse and carrier invisible.

How should tracking, proof of delivery and the returns boundary be defined?

After carrier handoff, tracking events usually originate in the carrier’s system. A WMS or sales channel may display those data to the customer, but it does not create the carrier’s scan frequency, transfer route or delivery decision.

Statuses should be distinguished clearly:

• Label created: A shipment identifier exists; physical handoff has not yet been proven.
• Handed over to carrier: The parcel has left warehouse control and a handoff record exists.
• Accepted by carrier: The first acceptance event has been received from the carrier’s network.
• In transit: The parcel is moving through the carrier’s operation.
• Out for delivery: The parcel has entered the final-mile delivery run.
• Delivered: The carrier has reported the delivery outcome.
• Delivery exception: The normal flow was not completed because of an address issue, recipient unavailability, damage, loss or another cause.

The content of a POD, or electronic proof of delivery, varies by carrier and service. A delivery status may include a signature, name, location or photograph; none of these is automatically available for every service. Information shown to the customer should not exceed the evidence supplied by the carrier.

Delivery may close forward fulfillment. A customer’s return request opens a new reverse-logistics flow. Return authorization, transport, warehouse receipt, inspection and the decision to resell, quarantine or take another action should not be combined under one status. For detailed implementation, see our reverse logistics and returns management guide. Carrier selection and the customer delivery experience are covered in our last-mile logistics guide.

How should exceptions be managed without breaking the process?

The quality of a real operation is demonstrated not only by how quickly normal orders move, but also by how visibly and systematically orders outside the normal flow are managed. “The order was delayed” is not a sufficiently precise cause.

Exceptions can be separated by stage and owner:

• Order data: Incomplete address, duplicate record, cancellation, payment or risk hold, or a changed customer request
• Inventory: Insufficient available quantity, damaged or blocked product, or a discrepancy between system and physical inventory
• Picking: Product missing from its location, wrong SKU or quantity, or serial or lot mismatch
• Packing: Undefined rule, unavailable packaging, or weight or dimensional tolerance exceeded
• Dispatch: Label error, carrier-service rejection, incorrect staging area or missed collection
• Transport: No first acceptance scan, delay, damage, loss or unsuccessful delivery

For every exception, record the time first detected, current owner, customer impact, next action and resolution time. If an order is advanced manually, the record should show which control was bypassed and why. Oracle’s official shipping model likewise tracks exceptions related to picking, shipments and automated batch processes separately, with error, warning and information severities. This demonstrates the practical value of distinguishing cause and severity rather than relying on one general error queue. Oracle Shipping Exceptions

The cancellation window must be defined particularly clearly. An order for which inventory has not yet been allocated cannot be canceled in the same way as one that has already been picked, packed or handed over to the carrier. Each stage has different implications for inventory, labels, charges, customer notifications and physical retrieval.

How does a WMS support end-to-end traceability?

A WMS should do more than display the current inventory quantity. It should be able to relate receiving, location, allocation, picking, control, parcel, label and handoff events around the same order and product identity.

A useful event record should answer the following questions:

• What happened? Which SKU, quantity, order or parcel was processed?
• When did it happen? What were the actual event time and the system-recording time?
• Where did it happen? At which warehouse, zone, location or dispatch door?
• Who or what performed it? Which user, device, integration or automated rule?
• Why did it happen? Was it a normal task, exception or manual correction?
• Which record is it linked to? What is the relationship to the order, work order, label, tracking number or document?

The GS1 EPCIS approach records supply-chain events through the dimensions of “what, when, where, why and, where necessary, how.” Memnun Depo is not claimed to use EPCIS; the standard is cited here as an example of the questions a traceable event record should answer. GS1 EPCIS and CBV Implementation Guideline

A large number of records does not, by itself, create visibility. If the same status has different meanings across systems, events can be altered retrospectively or statuses are generated automatically without physical movement, the result may be false confidence. Integrations, the WMS and carrier systems should use a shared status dictionary and a defined hierarchy of source systems.

For more detail on WMS, barcode, automation and integration choices, see our fulfillment technology guide and automated order fulfillment guide.

Which KPIs should be used to measure the fulfillment process?

One “average preparation time” does not explain an entire operation. Every KPI should define its starting event, ending event, denominator, excluded orders, reporting period and responsible party.

An initial KPI set may include:

• Order-ingestion success: Share of eligible source orders received completely and uniquely by the WMS
• Validation time: Time from WMS entry to a decision that the order is processable or on hold
• Hold age: Age and reason distribution of orders that cannot be processed
• Inventory-allocation success: Share of orders or order lines matched with available inventory on the first attempt
• Queue time: Time from release to operations until picking begins
• Picking accuracy: Share of order lines completed with the correct SKU and quantity against a defined denominator
• Right-first-time preparation: Share of shipments that pass control, packing and labeling without rework
• Ready-for-dispatch time: Time from the order becoming processable until physical preparation is complete
• On-time carrier handoff: Share of parcels handed over within the committed handoff window
• Handoff-to-first-scan gap: Time between the warehouse handoff record and the carrier acceptance event
• Exception rate and resolution time: Separated by warehouse, integration, brand or carrier cause
• Warehouse-caused reshipment or return: Events caused by the wrong product, quantity, label or packing

ASCM’s “Perfect Customer Order Fulfillment” metric evaluates not only the correct product and quantity, but also time, place, customer, documentation and product condition. An order is not a “perfect order” simply because it left the warehouse on time. ASCM Perfect Customer Order Fulfillment

Information quality should not be separated from operational KPIs. Research on order fulfillment treats timeliness, accuracy, accessibility and reliability of information as distinct measurement dimensions. Even if the physical parcel is prepared correctly, incorrect or late tracking information can make the process appear incomplete to the customer. Forslund — Measuring information quality in the order fulfilment process

An average can conceal a small number of severely delayed orders. In addition to the median, teams may therefore monitor high percentiles, the oldest open order and the distribution of causes. Universal “correct” speed or error benchmarks taken from other businesses should not be used; targets must reflect the order profile and contract scope.

What are Memnun Depo’s role and responsibility boundaries in the process?

Memnun Depo’s publicly verified fulfillment flow can be described as follows:

• Integration, packing, carrier and operating rules are defined with the client in advance.
• At receiving, goods are unloaded and the declared number of pallets, cartons or individual loose items is checked. Unit-by-unit counting inside sealed cartons whose contents were declared in advance is outside the standard scope. SKU setup, inventory entry and pallet or shelf put-away are then completed.
• The WMS relates receiving, location, inventory, order and dispatch movements within the same record chain.
• Product and shipment matches can be verified through the barcode workflow.
• A processable order has valid order data, available inventory and a defined packing rule.
• Once these orders enter the system, they are picked, prepared and made ready for dispatch within 12 hours.
• Transport-ready boxed products can be picked, labeled and handed over to the carrier without unnecessary repacking.
• When packing is required, it is performed according to the product and client rules and quoted separately by package size and type.
• The carrier label and tracking information are linked to the relevant order record.
• Returns are processed in a separate flow according to the operating and decision rules agreed with the client.

The 12-hour period is not a final-delivery promise. Carrier collection, in-network transport, delivery sequencing, delivery outcomes and POD generation fall outside this period. Memnun Depo must not be presented as managing the carrier’s route, vehicle or final-mile operation.

This article does not create a new performance claim regarding daily capacity, error rate or final-delivery time. The actual SLA should be established contractually after the order profile, integration, inventory, packing and carrier responsibilities are defined.

You can review the current service scope on the Memnun Depo website or contact us to discuss a workflow suited to your operation.

Frequently asked questions about the fulfillment process

Does fulfillment time begin as soon as the customer places the order?

Not always. The order must become processable with valid data, available inventory, required approvals and defined packing and carrier rules. The starting event should be stated clearly in the contract and system statuses.

Does “ready for dispatch within 12 hours” mean delivery to the customer within 12 hours?

No. It means that a processable order is picked and prepared inside the warehouse and made ready for dispatch. Carrier collection and delivery to the end customer are separate processes.

If a shipping label has been created, has the parcel been handed over to the carrier?

No. Label creation, physical carrier handoff and the carrier’s first acceptance scan are separate events. Each should be recorded with its own time and evidence.

Why might one order be split into several parcels?

Products may be stored at different locations, may not be suitable for shipment together, may exceed weight or dimension limits, or may be only partially available. The content and tracking relationship for each parcel should be recorded separately.

Must a product in a ready-to-ship carton be repacked?

No. If the existing carton meets the transport conditions and client rules, the product can be picked, labeled and handed over to the carrier. The scope and price of additional packing are defined separately when it is required.

Does using barcodes eliminate incorrect shipments completely?

No. Barcodes strengthen validation, but they cannot provide a guarantee on their own if master data, labels, scanning discipline or exception controls are incorrect.

Is a return the final step of the fulfillment process?

A return is a separate reverse-logistics process that may begin after forward fulfillment is complete. Authorization, transport, receipt, inspection and product disposition should each be managed with their own statuses.

Sources and further reading

• ASCM. Introduction to SCOR Processes: Order, Fulfill and Return and Fulfill process.
• ASCM. Perfect Customer Order Fulfillment.
• Gu, J., Goetschalckx, M. and McGinnis, L. F. (2007). Research on warehouse operation: A comprehensive review.
• De Koster, R., Le-Duc, T. and Roodbergen, K. J. (2007). Design and control of warehouse order picking: A literature review.
• Van Gils, T. et al. (2018). Designing efficient order picking systems by combining planning problems.
• Boysen, N., Weidinger, F. and de Koster, R. (2019). Warehousing in the e-commerce era: A survey.
• Forslund, H. (2007). Measuring information quality in the order fulfilment process.
• GS1. Global Traceability Standard, Logistic Label Guideline and EPCIS and CBV Implementation Guideline.
• Shopify. Inventory states and Apps in order management.
• Oracle. Overview of Shipping Exceptions.

This article provides general information on operations and process design. Platform APIs, carrier services, product requirements and contractual responsibilities may change, so each implementation should also be validated against current technical documentation and client-specific business rules.