Omnichannel vs. multichannel in practice

The distinction in one paragraph

Multichannel means several channels operating next to each other. Shop, marketplace, retail, wholesale — each with its own targets, often its own stock, sometimes its own pricing and almost always its own reporting. Omnichannel means the customer, not the channel, is the unit: one product truth, visible stock, one customer record, and processes that cross channel borders.

There is a useful stage in between, usually called cross-channel: single bridges without a full redesign. Stock visibility in the shop, a voucher that works everywhere, a return that can be handed in at a store. Most companies that call themselves omnichannel are here — and for many that is exactly the right place to be.

How to recognise multichannel in daily work

The label on the strategy deck is unreliable. These symptoms are not:

• Product data is maintained twice. The same article has a different description, different images or a different name depending on where it is sold.
• Stock is split by channel. The shop cannot see store inventory, or a fixed contingent is reserved per channel to avoid overselling.
• Service cannot see the whole customer. A purchase made in a store is invisible to online support, so every cross-channel case becomes a phone call.
• Vouchers and loyalty stop at the channel border. Points earned online cannot be spent offline, which customers experience as a broken promise rather than a system limit.
• Reporting uses different definitions. Each channel counts revenue, returns and customers slightly differently, so the sum is not comparable to anything.
• Channels compete internally. Store staff receive no credit for an online order they triggered, so nobody triggers one.

What omnichannel actually demands

Four building blocks carry almost every setup. Product data in one place, so that every channel reads from the same source rather than from an export. Stock transparency, at least as information — showing availability is a much smaller project than reserving across channels, and it already solves most customer questions.

Then customer identity: one account, one customer number, a purchase history that does not depend on where the purchase happened. And finally the processes themselves — click and collect, ship from store, return anywhere. Each of these is a logistics and accounting question long before it is a software question.

The fifth block is the one that gets skipped: incentives. If a store is not credited for the revenue it triggers, and if a return handled in-store lands in the store's own numbers as a loss, the process will exist in the system and not in reality.

The most expensive mistake: technology before organisation

Systems can be connected — that part is well understood and, with modern commerce platforms, no longer exotic. Conflicting goals cannot be connected. A return accepted in a store touches inventory valuation, commission, and the question of who carries the margin loss. If those questions are unanswered when the feature goes live, the feature will be quietly discouraged at the counter.

This is why omnichannel projects that start with a platform decision so often stall. The platform is rarely the reason. The reason is that two departments are measured on numbers that move in opposite directions.

When multichannel is the right decision

Not every channel deserves integration. Marketplaces are a good example: they come with their own pricing logic, their own margin structure and limited access to customer data. Running them deliberately as a separate business — separate assortment, separate calculation, separate targets — is often healthier than pretending they are part of one seamless experience.

The same applies to companies with small teams. Integration costs are permanent: every interface needs an owner, a monitoring routine and a plan for the day the API changes. Choosing multichannel on purpose, with clear reasons, is a legitimate strategy. Ending up there by accident is not.

A realistic way in

Order matters more than ambition. Start with customer identity, because everything else refers to it. Then stock visibility as information. Then one — exactly one — cross-channel process, implemented completely, including the accounting side. Only then the reporting layer with definitions that all channels share.

One use case done properly beats five half-finished ones, both for customers and for the internal appetite to continue. The first working process is what convinces the organisation that the second one is worth the effort.