Guide · Retail
The omnichannel playbook.
Omnichannel is sold as software. It is really a seam: the join between the team that built your website and the team that runs your stores. They were built as two companies, so the join was never anyone's job. Here is how to fix it, in the order the customer feels.
This guide, in 5 parts
Most retail is still physical, and the channels do not know each other.
The store is not dead. In Q1 2026, e-commerce was 16.9% of US retail, which leaves roughly 83% still happening in a physical place. The customer moves freely between the two, looks online and buys in store, or the reverse, and barely thinks about the line.
The business thinks about little else. The omnichannel customer shops about 1.7 times more often than a single-channel one, and separate research puts multi-channel shoppers at roughly 30% higher lifetime value. That is the prize. The problem is that only about 12% of retailers can recognise the same shopper across online and offline.
Meanwhile about 79% of people expect consistent interactions across a brand. So the gap is plain: the customer is already one person, the brand keeps treating them as two.
« The customer is already one person. The brand keeps treating them as two. »
It is not a missing platform. It is an unowned seam.
The reason this is hard is organisational before it is technical. The website team and the store team were almost always built as two companies, with two budgets, two systems and two definitions of the customer. The join between them was never written into anyone's job. So it does not happen.
Software is sold as the answer, and a platform can hold one stock truth and one customer. But the platform does not decide who owns the seam. That is why a retailer can buy the suite and still feel like two companies to the person standing at the till with a phone in their hand.
How it was built
Two companies
- One team owns the website, another owns the stores.
- Two stock ledgers that reconcile overnight, if at all.
- The customer is an email on one side, a loyalty card on the other.
- An order belongs to the channel that took it, and stays there.
- Returns bounce: bought online, refused in store.
Result: the seam is nobody's job.
How it is felt
One customer
- One brand, one expectation, wherever they are standing.
- One stock answer: is it here, now, or not.
- One identity that remembers them on both sides.
- One order they can start anywhere and finish anywhere.
- One return, through whichever door is closest.
Result: the seam is the whole experience.
The fix is to make the seam someone's job, then build the connections that close it. There are four.
Four joins turn two channels into one experience.
One experience is not a vibe, it is four concrete connections. Each has a pass-test you can check in plain language, no dashboard required. If a join cannot pass its test, it is not done, whatever the platform says.
- Inventory · one stock truth. The website and every store read the same live quantity for the same item. No second ledger, no overnight sync that lies by lunchtime. Pass test: the count a customer sees online matches what the shelf actually holds, right now.
- Identity · one customer. The same person is the same record whether they buy on the site or hand over a card at the till. One profile, one history, both directions. Pass test: a shopper served in-store is recognised online next morning, without re-introducing themselves.
- Order · one order, any channel finishes. An order started anywhere can be completed anywhere. Buy online and collect in store, reserve in store and ship home, all on one order, not two systems pretending to talk. Pass test: a single order can begin on the web and finish at a counter, with one number tracking it throughout.
- Returns · any door. Whatever was bought, wherever it was bought, comes back through any door and lands as one event the customer never has to argue about. Pass test: a web purchase is accepted, refunded and restocked at any store without a phone call to head office.
Fix them in the order the customer feels them.
The four joins are not equal, and they are not independent. Each leans on the one before it, and the customer notices them in a fixed order. So fix them in that order: stock, identity, order, returns. This is a Hub Map call, sequencing the work by what the customer feels first.
Stock first, because empty shelves are felt instantly
The fastest way to break trust is to promise stock that is not there, or hide stock that is. Make one quantity true for the website and every store before anything else. Until the count is shared and live, every other join inherits a lie about what exists and where.
In shortOne live count, read the same by the site and the shelf.
Identity next, so the customer stops being a stranger
Once stock is honest, give the customer one identity that carries across channels, so a face served in a store is the same record that logs in online. This is the join that turns two audiences into one customer, and the one most retailers quietly skip because it crosses the hardest internal border.
In shortOne person, one record, recognised on both sides.
Order third, so any channel can finish the sale
With stock true and the customer known, let a single order move between channels: started online, completed in store, or the reverse. This is where buy-online-collect-in-store and reserve-in-store-ship-home stop being marketing words and become one order that two channels can both touch.
In shortOne order any channel can begin and any channel can finish.
Returns last, because they ride on the first three
A return is just an order run backwards through an identity, against a stock count. Build it once the other three hold, and it falls out almost for free; build it first, and it has nothing solid to stand on. Done right, the customer brings anything back through any door and never feels the seam.
In shortAny door takes any return, as one clean event.
Frequently asked.
We bought an omnichannel platform. Why does it still feel like two companies?
Because the platform is the tool, not the seam. The software can hold one stock truth and one customer, but it does not decide who owns the join between your website team and your store team. That ownership gap is what the customer feels as two companies. The fix is operational before it is technical: name the join, give it an owner, then make the four connections hold in the order the customer notices them.
Where do we actually start?
Start with stock, because empty or wrong shelves are felt instantly and break trust fastest. Make one quantity true for the website and every store before you touch identity, orders or returns. Each later join leans on the one before it, so fixing them out of order means rebuilding. Stock, then identity, then order, then returns, in the sequence the customer experiences.
Can our existing website and store teams do this themselves?
Often, yes, and they should where they can. If you have engineers who know both your e-commerce stack and your store systems, the four joins are a clear programme they can run. We are useful when the two sides were built as separate companies and nobody owns the join, when one side is strong and the other is not, or when you want it shipped quickly without a long hire. We have shipped both sides, which is rarer than it should be.
Is buy-online-pickup-in-store worth the effort?
For most multi-store retailers, yes. More than 68% of US shoppers used buy-online-pickup-in-store at least once in 2024, and click-and-collect is forecast at around 11% of US e-commerce in 2026. But it only works on top of the joins: it needs true stock so the promise is real, identity so the pickup is smooth, and the order join so one order can start online and finish at the counter. Build the joins and it follows; bolt it on alone and it disappoints.
How is this different from a digital transformation programme?
Scope and honesty. A broad transformation tends to boil the ocean and stall. This is four named joins with a pass-test each, sequenced by what the customer feels, so progress is visible and arguable in plain terms. It is the how-to under our flagship offer, where we build and stay until the seam holds, rather than hand over a roadmap and leave.
Much of this you can run yourselves. If you have engineers who know both your e-commerce stack and your store systems, the four joins are a clear programme to hand them.
Bring us in when the two sides were built as separate companies and nobody owns the seam, when one side is strong and the other is weak, or when you want it shipped fast without a long hire.
We are the rare team that has shipped both sides. This is the how-to under our flagship offer, The Whole Way, where we build and stay until the seam holds. On sequencing the work, see the Hub Map.
Feel like two companies to your customer?
No deck, no gate. A working session on the real stock, identity, order and return flows, with the operators who would close the seam.