Multichannel retailing

Multichannel retailing is selling through more than one channel: an owned online store, marketplaces, physical retail, wholesale, and social commerce. It describes the commercial model rather than the integration between channels, which is the distinction from omnichannel, where the channels share stock, customer records, and fulfillment. The two terms are often used interchangeably in vendor copy and they are not interchangeable, a point Verhoef, Kannan and Inman set out in the Journal of Retailing and one the multichannel and omnichannel comparison page works through in detail. A business can be multichannel with entirely separate systems per channel, and many are, which is where the operational cost sits.

One order layer under every channel

An order management system is the layer that accepts orders from every channel and holds them in one place, deciding where each is fulfilled and tracking its state. Without it, each channel keeps its own order records, and questions that span channels cannot be answered without opening several systems. The capabilities that matter for multichannel specifically are channel-agnostic order intake, allocation rules that account for stock committed to other channels, and a single order status vocabulary, since marketplaces and storefronts define statuses differently and unreconciled definitions produce contradictory customer-facing information. Allocation in particular has settled definitions worth borrowing rather than inventing, in the CSCMP supply chain glossary, and the status vocabulary problem is visible the moment two systems disagree about what fulfilled means, which the Shopify Fulfillment object defines for one of them and a marketplace defines differently for another. What the layer itself is doing is order management.

Worth holding onto while everyone argues about the next big shift in commerce: a new channel is just a new channel. When Google and Shopify dropped the Universal Commerce Protocol, everyone called it a revolution. I see it differently. UCP is just another sales channel, like your owned storefront, like marketplaces, like social and content commerce: another way for customers to find and buy your products. And each new channel adds complexity to the post-purchase process, because every order it brings still has to land in one place, be allocated against the same stock, and be answerable in one status vocabulary. On The Ecommerce Edge I put it the same way: agentic commerce will become another sales channel, like marketplaces or fulfilled by Amazon. New channels keep coming. The brands that win are not the ones chasing every new front door. They are the ones that have already fixed what happens after, which is why omnichannel order management is the part worth investing in rather than the channel.

One process that satisfies the strictest channel

Each channel imposes its own post-purchase rules. Marketplaces frequently mandate messaging templates, dispatch windows, and tracking upload deadlines, with seller-performance consequences for missing them. Owned channels allow branded tracking and free-form communication. The operational requirement is one process that satisfies the strictest channel's obligations while giving the owned channel the branded treatment, driven from one set of order events rather than a per-channel implementation that drifts. What those branded events look like on the owned channel is shipping notifications, and the connection work that feeds them is carrier integration.

I hold every channel to that strictest standard, including the ones you do not run yourself: marketplaces, dropship, B2B, ship from store, click and collect. One brand we work with now tracks a hundred dropshippers in one place. He never had visibility into them before, and now he has absolute clarity on whether each one is delivering to the service the brand promised.

Overselling is two channels working from a figure nobody owned

Inventory synchronization is the defining technical problem of multichannel. One pool of stock is exposed to several channels that each sell independently, so the risk is overselling, and its severity rises with channel count and sales velocity. Mitigations include a single source of truth with short sync intervals, buffer stock per channel, and reservation at the point of order rather than at dispatch. Bundles, pre-orders, and kits are the standard failure cases, because they consume stock in ways simple per-SKU counters model badly. All of it assumes the same physical unit carries the same identifier in every channel, which is what the GS1 GTIN exists to guarantee and what a business running two SKU schemes has already given up.

Here is the thing that surprises people. We have had a couple of retailers tell us they only do a certain number of orders, and once they came onto the platform it was actually double or triple that. My read is that nobody had ever seen every channel in one place at the same time. If a business cannot count its own orders across channels, the stock position underneath those orders is not a number either, and that is what overselling actually is. It is not a synchronization bug so much as the moment two channels find out they were both working from a figure nobody owned. The version of that failure the customer sees is split shipments.

The visible cost lands on the person answering the email

A customer who bought on a marketplace and contacts through the brand's own channel is a routine case that fragmented systems handle badly. Centralized context means an agent can see the order, its channel, its fulfillment state, and prior contact, regardless of where the sale occurred. Where this is missing, the observable symptom is agents working across several systems per inquiry, which raises handle time and produces inconsistent answers. The system that is supposed to hold that context is the helpdesk, and it usually holds the conversation rather than the order. Marketplace privacy rules complicate it, since some channels withhold customer contact details, so the design has to work from order identifiers rather than assuming a unified customer record.

The visible cost of multichannel lands here, and it lands on the person answering the email. On a discovery call with one global brand I heard the usual setup: a separate Shopify store for each market, so customer service in each market logs into a different Shopify to understand what happened after the purchase. It is all disconnected. Across our calls, teams toggle between three and ten systems to resolve a single ticket, and on the worst setups it is more. Every one of those switches is a place where the order context gets dropped and the agent starts again. Operational complexity is the silent growth killer, because it never appears as a line item. It appears as handle time, as inconsistent answers, and as a support team that grows in proportion to the channel count rather than to the revenue. Every order is a promise. Keeyu keeps it. A customer does not care which channel their order came through, or that it was placed somewhere your support tooling cannot see. What the connected version is worth is omnichannel benefits, and the operating model it belongs to is post-purchase operations.

That range is not a guess. It is what operators told us across the 786 pain points we mined from 270 customer call transcripts between May 2025 and May 2026, where toggling between three and ten or more systems to resolve one ticket was among the most repeated entries, and a single 'where is my order' investigation could open more than 40 tabs.

Returns multiply the channel problem

Returns multiply the channel problem, because the return path may differ from the sale path and each channel sets its own policy and refund timing. The operational questions are where a returned item physically goes, which stock pool it re-enters and when, who funds the return, and how refunds reconcile against a marketplace's own settlement process. Cross-channel returns, such as buy online return in store, add a further requirement: the returning location must be able to see and accept an order it did not sell. The wider process, including where the item physically re-enters stock, is returns management.

• Journal of Retailing. Verhoef, Kannan and Inman set out in the Journal of Retailing. Multichannel and omnichannel as different things, from the research.
• Council of Supply Chain Management Professionals. CSCMP supply chain glossary. Settled definitions for allocation and order management.
• Shopify. Shopify Fulfillment object. One channel's definition of fulfilled, against a marketplace's different one.
• GS1. GS1 GTIN. The same physical unit carrying one identifier across channels.
• Shopify Engineering. Building the Universal Commerce Protocol. UCP, co-developed with Google as an open standard for AI agents to connect and transact with any merchant.