Retail has spent more than a decade connecting channels.
Websites talk to stores. Inventory moves across locations. Customers can buy online and return in store, collect orders, access loyalty accounts across touchpoints and increasingly expect the brand to recognise them wherever they appear.
Technically, much of the hard work has been done.
Organisationally, many retailers still behave as though the customer wakes up each morning and decides whether to be an eCommerce customer, a store customer, a marketplace customer or a loyalty customer.
They do not.
They are simply customers trying to buy something.
The gap between those two realities is becoming one of the biggest constraints on omnichannel growth.
We connected the technology before we connected the enterprise
The first phase of omnichannel was largely an infrastructure problem. Retailers needed common inventory, order management, customer identity, fulfilment capability, integrated payments and a digital layer that could interact with stores rather than compete with them.
That work was difficult, expensive and necessary. It also created a slightly misleading sense that once the systems were connected, the organisation itself had become omnichannel.
Often it had not.
The customer journey became horizontal while the company remained vertical.
eCommerce had its targets. Stores had theirs. Marketing had acquisition metrics. CRM had retention metrics. Merchandising managed margin and stock. Technology managed delivery. Loyalty often sat somewhere between marketing, customer experience and finance, depending on the organisation.
Each team could be performing well against its own measures while the customer experienced the seams between them.
I have seen this repeatedly in large retail environments. The systems may be capable of serving one customer across channels, but the incentives around those systems continue to encourage each function to defend its own piece of the transaction.
That is not a technology integration problem.
It is an operating model problem.
Channel P&Ls can make sensible people behave strangely
Channel accountability exists for good reasons.
A store leader needs to understand store productivity. An eCommerce team needs to understand digital conversion and profitability. Marketplace economics should not disappear into a blended number that makes poor performance impossible to diagnose.
The problem begins when those measures stop being diagnostic tools and become competing definitions of success.
Imagine a customer who discovers a product online, visits a store to try it, receives advice from an associate and completes the purchase later that evening through the app.
Who created the sale?
The digital team can point to the conversion. The store can point to the service interaction. Marketing may have paid to acquire the customer months earlier. CRM may have triggered the message that brought them back. Merchandising selected the product and pricing that made the purchase attractive in the first place.
Trying to award ownership of that transaction to one channel is administratively convenient and commercially artificial.
Yet organisations do it constantly.
The consequence is predictable. Teams optimise the part of the journey they are measured on rather than the economics of the customer relationship as a whole.
A store may hesitate to support an online sale because it receives little credit for the transaction. Digital teams may prioritise online conversion even when store fulfilment or assisted selling would produce a better customer outcome. Marketing can chase new customers while the wider business underinvests in retaining profitable ones.
None of these behaviours require bad leadership.
They are often rational responses to badly designed incentives.
Attribution tells us where the sale happened, not necessarily where value was created
Retail has invested enormous energy in attribution because executives understandably want to know what produced the result.
The problem is that attribution becomes less reliable as the journey becomes more connected.
A sale can be easy to attribute technically and still difficult to explain economically.
The last interaction may have closed the transaction, but something else created the intent. A store visit may have built confidence. A loyalty benefit may have reduced hesitation. A service interaction months earlier may have protected the relationship. The customer may have first discovered the brand through a channel that never appears anywhere near the final conversion.
This is why arguments about “which channel owns the customer” increasingly feel outdated to me.
Channels do not own customers.
At best, they contribute differently to a relationship.
That distinction matters because executives make investment decisions based on what they can measure. If the organisation systematically over-credits the final transaction point, it will naturally over-invest there and under-invest in the capabilities that created the customer value upstream.
The answer is not to abandon attribution. Businesses still need to understand channel economics, marketing effectiveness and operational productivity.
But attribution needs to sit underneath a broader view of customer economics.
What is the cost of acquiring the customer? How frequently do they return? How profitable is the relationship across channels? Does store engagement increase subsequent digital spend? Does loyalty change frequency, retention or margin? Does easier fulfilment increase lifetime value rather than simply shifting where the transaction is recorded?
Those questions are harder than assigning a sale to a channel.
They are also much closer to how customers actually behave.
The store is becoming part of the digital operating system
One of the most persistent mistakes in omnichannel strategy is treating stores as the physical alternative to digital.
That may have been a useful distinction twenty years ago. It is increasingly meaningless now.
A modern store can be a showroom, fulfilment point, returns hub, service environment, acquisition channel, media surface and relationship-building platform at the same time.
The sales associate is part of that system too.
If an associate can see customer history, understand preferences, access inventory beyond the four walls of the store, order products from another location and continue the relationship after the customer leaves, the store is no longer simply a place where physical transactions occur.
It has become another interface into the enterprise.
This changes how store technology should be evaluated.
Mobile point of sale, clienteling, endless aisle, customer identity and cross-channel inventory are sometimes treated as store productivity projects. Their value is broader because they allow the customer relationship to continue regardless of where inventory happens to sit or where the eventual transaction is recorded.
A customer standing in a store should not hear, “We don’t have it here, but perhaps you can check online.”
From the customer’s perspective, “online” and “the store” are not two companies.
The associate should be able to solve the problem.
That sounds obvious. Yet making it possible requires technology, inventory visibility, payment capability, incentives, training and clear commercial ownership to work together.
This is why the next wave of omnichannel advantage will come less from adding another customer-facing feature and more from making the organisation behind the feature behave coherently.
AI will make organisational fragmentation harder to hide
AI is likely to accelerate this tension rather than resolve it automatically.
An intelligent customer layer can increasingly understand behaviour across browsing, transactions, service interactions, loyalty activity and product preferences. It can recommend next actions, predict intent and help employees serve customers with far more context than before.
But intelligence moving horizontally across the customer relationship will collide with organisations still managed vertically.
An AI system may identify that the best action is for a store associate to contact a customer who has been browsing online. The organisation then needs to decide whether the associate has access to that data, whether contacting the customer is permitted, whether the store gets credit if the eventual purchase happens online and who owns the commercial outcome.
The model can produce the recommendation in milliseconds.
The operating model can still spend three months arguing about governance.
This is where AI becomes useful as an organisational stress test.
It exposes the places where data is connected but authority is not, where customer context exists but employees cannot use it, and where the technically correct action conflicts with the incentive structure.
Retailers that treat AI purely as another digital capability may automate parts of the existing journey.
Retailers that use it to redesign how the enterprise serves the customer can change the economics of the relationship.
That requires a different conversation from adding copilots to functional teams.
Shared customer economics do not mean abandoning accountability
The obvious concern with organising around the customer is that everything becomes shared and therefore nobody is accountable.
That would be a mistake.
Stores still need productivity measures. Digital channels still need conversion, fulfilment and profitability metrics. Marketing must remain accountable for the quality and cost of demand. Merchandising must own assortment and margin. Functions need clear responsibilities precisely because complexity increases when journeys cross organisational boundaries.
The change is in what sits above those measures.
The enterprise needs a common economic view of the customer that functions cannot optimise against.
That might include acquisition cost, retention, frequency, lifetime contribution, full-price behaviour, cost to serve and the effect of omnichannel engagement on customer value. The exact measures will vary by business, but the principle is consistent: channel performance should be evaluated inside the economics of the relationship, not instead of it.
Decision rights then become critical.
Someone has to own cross-channel customer outcomes. Teams need to know when local optimisation should give way to enterprise value. Incentives have to reward behaviours that help the customer even when the transaction appears somewhere else.
This becomes especially important in stores.
If an associate spends twenty minutes helping a customer and the purchase is eventually completed digitally, the organisation needs a way to value that contribution rather than teaching employees, through compensation, that helping the customer finish elsewhere is bad business.
The same principle applies in reverse. Digital should not be rewarded for shifting transactions away from profitable store behaviour simply because its own revenue line becomes larger.
An omnichannel operating model therefore does not remove channel accountability.
It puts it in the right hierarchy.
Customer economics first. Channel economics underneath.
The goal of omnichannel should be to make omnichannel disappear
There is an irony in the word itself.
The better a retailer becomes at omnichannel, the less the customer should notice that channels exist.
Inventory is simply available. The relationship is remembered. A return works. An associate can help. The order arrives where it should. The customer can start somewhere and finish somewhere else without being asked to understand the company’s organisational chart.
That simplicity is difficult because the enterprise carries the complexity on the customer’s behalf.
Retailers have already invested heavily in the technical foundations required to do this. The next advantage will come from aligning incentives, decision rights, data access and customer economics with the journey those systems now make possible.
That is a harder transformation because it touches organisational power.
Channel structures are visible. Budgets sit inside them. Careers have been built around them. Revenue gets reported through them. Changing the technology can be uncomfortable, but changing who receives credit for the customer is often more uncomfortable.
Still, the direction is difficult to avoid.
Customers will continue moving across stores, apps, marketplaces, social platforms and whatever interface comes next. AI agents may soon become another participant in that journey.
The enterprise can keep reorganising the customer into the boxes it already understands.
Or it can accept that those boxes were always ours, not theirs.
The customer never shopped in channels.
The next omnichannel transformation is finally organising the business as though we believe it.