Omnichannel retail in Portugal: the definitive guide for store chains

A customer walks into a footwear shop in Guimarães. She saw online that there was a size 38 pair in stock. She arrives, the salesperson looks, and the pair doesn't exist — it was sold two hours ago at another shop in the same chain, and the inventory only syncs at night. The customer leaves. The sale is lost. And nobody, at head office, will ever know it happened, because there is no record of a sale that never occurred.

This is the real cost of the lack of omnichannel: not the stockout, but the breakdown of trust between what the customer sees and what the shop delivers. The thesis of this guide is simple: omnichannel retail in Portugal is neither an e-commerce project nor a marketing project — it is a problem of integrating operational data across POS, warehouse, ERP and digital channels, and it is solved in the right order, starting with unified inventory.

Over the next sections we will take the problem apart at its operational root, map the Portuguese landscape honestly, compare the implementation models with their real costs, and lay out a 90-day path that the board can approve without betting the whole house. We write this from three decades on the shop floor and the warehouse floor in the North of Portugal — not from a slide deck.

1. The real operational problem

The word "omnichannel" reaches board meetings via marketing. It arrives clean, shiny, with customer-journey slides. And then it hits the shop floor and the warehouse, where reality is different: systems that don't talk to each other, stock levels that diverge, and a warehouse manager who has already seen three "digital revolutions" fail and isn't going to let go of the radio because of the fourth.

The problem is not conceptual. Everyone agrees that the customer should see the same stock everywhere. The problem is that, beneath the pretty concept, there are three or four databases that were designed in different years, by different vendors, with different priorities — and that were never designed to share a single truth.

Where the model breaks in practice

The most common fracture point in a Portuguese retail chain is not the digital shop window. It is inventory divergence. Each shop has its POS, often with a local database. The e-commerce runs on another platform. The ERP at head office consolidates — but with a delay of hours or a day. Result: three different truths about how many units of a SKU exist, and none of them reliable in real time.

For a footwear chain, the problem multiplies. A collection of women's shoes may have 800 to 1,200 active SKUs, along three axes — colour, size and fitting. Size 37 in one colour may exist; in that same colour, the 38 has sold out. A generalist ERP models this with difficulty. It models it with free-text fields, with nomenclatures invented by each shop, and with the predictable result: the online promise of availability does not match what is on the shelf.

The technical root is almost always the same. When the e-commerce was set up, someone replicated the catalogue to the online-shop platform with a separate item master. From that day on, each channel started to "think" it knows the stock. And when two systems think they know, neither knows. The divergence is not a bug — it is the logical consequence of an architecture in which the truth has no single owner.

The specific case of footwear and seasonal fashion

In Felgueiras, São João da Madeira or Oliveira de Azeméis, the collection logic makes everything worse. A women's pair in a specific colour and fitting may be literally unique in stock — there is no possible replenishment mid-season, because the factory has already moved on to a different production run. In this context, a synchronisation error is not a small delay: it is the difference between selling the only available black size 38 pair or promising it simultaneously to two customers on two channels.

The Portuguese footwear industry exports the overwhelming majority of its production, with Europe as its main destination, and international buyers visit the cluster twice a year. When those same manufacturers open an own-brand shop and sell online, they import the SKU complexity of production into retail — and most generic retail platforms simply were not designed for that depth of matrix.

The four symptoms that give away a non-omnichannel chain

• The customer sees "available" online and the shop doesn't have it — or vice versa, with idle stock that nobody knows exists.
• An item bought online can only be returned online, forcing the customer to package and post it instead of walking into the shop next door.
• The loyalty card accumulates points in the shop but does not recognise them on the site, treating the same customer as two people.
• Head office closes the month and discovers inventory discrepancies that nobody can explain without a full day of physical counting.

Omnichannel doesn't fail in the design of the experience. It fails in inventory synchronisation — the rest is consequence.

The invisible cost: sales that are never recorded

The visible stockout hurts and is measurable. The invisible one — the customer who leaves because what they saw online wasn't there — leaves no trace in the system. It is a ghost sale. It appears in no report, and that is why it is the one that most destroys margin over the year without the board understanding why sales per shop stagnate despite online traffic growing.

There is a second invisible cost, even more insidious: the erosion of trust. A customer who travels to a shop on the basis of an online promise that is not kept does not come back twice. The first failure is the system's; the second is the brand's. And, unlike the day's lost sale, this damage accumulates and only shows up many months later, when it is already hard to link cause to effect.

Why the warehouse manager is suspicious — and is right

In the Lousada/Paços corridor, the warehouse manager of a distribution centre lives off the radio. Every minute away from it is a picking request stalled, a shipment delayed, a lorry waiting. When he hears "we're going to implement a new system", he doesn't hear innovation — he hears two days of training, a week of errors, and the certainty that, on the first month-end, the new screen will be slow and he'll take the blame.

This suspicion is not resistance to change out of laziness. It is memory. It is the accumulated experience of rollouts that promised a lot and delivered slowness. Any omnichannel project that ignores this reality — that treats the warehouse as an implementation detail instead of the critical axis it is — is doomed before it starts. Inventory is born or dies in the warehouse, and that is where omnichannel is won.

2. What exactly is omnichannel retail in Portugal

Omnichannel means that the sales channels — physical shop, e-commerce, marketplace, B2B sales force, telephone — share the same operational truth: one inventory, one customer, one history. The customer moves between channels without friction and without repeating information. The company sees a single reality, not silos.

Omnichannel is not multichannel

The distinction is operational, not semantic. Multichannel is having several channels that operate in parallel, each with its own stock, its own price, its own customer. Omnichannel is having several channels that share the same data infrastructure. Most Portuguese chains that call themselves omnichannel are, in fact, multichannel with a pretty site on top.

The test is simple and merciless: if a customer buys online, receives at home, and then goes to return it to the nearest shop — does the system recognise the purchase, accept the return, credit the customer and restore the stock in the shop as available? If so, it is omnichannel. If the salesperson has to call head office, fill in a form and the credit takes three days, it is multichannel with an omnichannel veneer on top.

The vocabulary you need to master

• BOPIS (Buy Online, Pick-up In Store) — buy online and collect in the shop. Requires knowing, in real time, which shop has the item.
• Ship-from-store — dispatch an online order from the stock of a physical shop, turning each shop into a mini-warehouse.
• OMS (Order Management System) — the layer that decides where each order ships from: central warehouse, shop A or shop B, depending on stock, proximity and cost.
• Endless aisle — the salesperson in the shop sells an item that is not physically there, ordering it from the warehouse or another shop for delivery to the customer.
• Click & Collect — a variant of BOPIS, with a dedicated collection point.
• Master inventory — the stock database that holds the single truth, against which all channels read and write.

What the framing of these acronyms hides

Each of these features looks like a button to switch on. It isn't. BOPIS requires the shop to know, at the moment the customer clicks "buy", that the pair exists, and to reserve it instantly — otherwise it will sell it over the counter to someone else in the next ten minutes. Ship-from-store requires the shop to have a packing, labelling and carrier-collection process, something a high-street shop was not designed to do. Each acronym is an operational contract, not a box to tick.

A brief history of why this got complicated

In the 2000s, e-commerce arrived at Portuguese companies as a separate department — often with its own warehouse and its own system, because integrating with the existing ERP was expensive and nobody wanted to touch what worked. That decision created the original sin: two inventories. Twenty years later, most omnichannel projects in Portugal are, in practice, an exercise in reunifying what should never have been separated.

There was a logic at the time. The ERP of the 2000s was transactional and heavy, made for the back office, not to serve thousands of stock queries per hour coming from a site. Separating the e-commerce protected the central system. The problem is that the technology evolved and the architecture did not. Many companies are still paying, in 2025, the price of an architectural decision taken when the iPhone did not yet exist.

The technical debt of Portuguese retail has a name: the day it was decided that e-commerce would have its own stock.

3. The landscape in Portugal today

E-commerce in Portugal has grown steadily, but from a lower base than the European average. According to Eurostat, the proportion of people who bought online in the previous 12 months has risen year after year, progressively approaching the EU average, though still below it.

Where Portugal is — and is not

INE data on ICT use in businesses show that a significant minority of Portuguese companies make sales through e-commerce, and that this proportion rises with the size of the company. The large chains are advanced; SMEs in specialist retail — optics, tools, regional apparel — are substantially behind.

This relative lag has a double reading. It is a risk — the SMEs that do not integrate their channels will lose ground to larger operators and to international platforms that already operate with unified inventory. But it is also an opportunity: whoever integrates now, with the right architecture, skips an entire generation of mistakes that the pioneers made and are still correcting.

The structural factor: size of shops and of the warehouse

A regional food-retail chain with 22 shops has a different logistical reality from a footwear chain with 8 points of sale. Food plays with high turnover, expiry dates and cross-docking. Footwear and apparel play with brutal seasonality — international buyers visit Felgueiras twice a year (men's footwear in August, women's in February) and the collection defines the whole year. Omnichannel has to accommodate both, and there is no single solution.

In regional food retail, product expiry enters the inventory equation. It is not enough to know there are 40 units of a yoghurt — you need to know that 12 expire this week. A good omnichannel architecture for food has to manage batch and expiry in the same master that serves the e-commerce, or the site will sell product that the shop cannot dispatch because it is out of date.

There is no "the Portuguese retail market". There is high-turnover food and the extreme-seasonality specialist — and confusing them when choosing software is the costliest mistake.

The regulatory pressure that accelerated POS digitalisation

Unlike many European markets, Portuguese retail has already been required by law to modernise the point of sale. Software certification by the AT, SAF-T reporting and, more recently, ATCUD, forced even the smallest trader to have a modern, connected POS. This is, paradoxically, an asset: the technological base of the Portuguese POS is more advanced than the omnichannel maturity would suggest. What is missing is not the POS — it is the link between the POS, the e-commerce and the master inventory.

The international buyer changes everything

For the footwear and apparel chains and brands of the North, the B2B channel is often larger than the B2C. A distributor in France or Germany orders by full grid, with negotiated payment terms, and expects a portal where they see stock, place orders and track shipments. This is omnichannel too — only turned to the professional, not the consumer. A good B2B order-portal solution handles this channel without duplicating inventory.

The classic mistake is to treat B2B as a world apart, with spreadsheets and phone calls, while investing in the shiny B2C site. But if the German distributor and the shop customer draw from the same physical stock, they have to draw from the same logical master. Otherwise, the brand ends up promising the same full grid to a distributor and to its own e-commerce, and someone is left without the product when it comes time to ship.

4. The implementation models

There are four realistic approaches to achieving omnichannel in a Portuguese chain. Each has a distinct cost, risk and time profile. The wrong choice costs years.

Model A — Point-to-point integration

Connect each system directly to the others: POS to e-commerce, e-commerce to ERP, ERP to warehouse. Cheap to start, unsustainable to maintain. Each new channel multiplies the connections. With four systems, that's six connections; with six systems, fifteen. It is the road that looks short and ends in a maze.

The Achilles heel of point-to-point is not the building — it is the maintenance. When the e-commerce vendor updates the API, three integrations break. When the POS changes version, another one breaks. And because each connection was made by a different person at different times, nobody has the complete map. The company becomes hostage to a set of fragile connections that only one person understands — and that person is usually the internal IT hero who one day retires.

Model B — ERP as the single core of truth

The ERP holds the master inventory. POS, e-commerce and B2B portal read and write against the ERP in near-real time. It is the most solid architecture for industrial SMEs that already have a vertical ERP. It requires an ERP capable of withstanding the transactional load of retail and a POS layer that synchronises well.

The strength of this model, for an industrial company, is that inventory is already born in the ERP — in the purchase of raw materials, in production, in warehouse receipt. Making the ERP the master is not adding a piece; it is recognising who already had the truth. Multi Connect resolves precisely the communication between branches and channels when the MULTI ERP is the core, without multiplying item masters.

Model C — Dedicated OMS as orchestrator

Introduce an Order Management System between the channels and the inventory. The OMS decides where each order ships from and manages ship-from-store with rules. Powerful for large chains with many shops and high volume. Excessive — and expensive — for chains with fewer than a dozen points of sale.

The OMS shines when there are many possible origins for the same order: a customer in Lisbon orders online, and the system has to decide whether to serve from the Porto warehouse, the Cascais shop or the Braga shop, optimising transport cost, lead time and stock balance. That decision, made hundreds of times a day with fine rules, justifies an OMS. Made ten times a day, it is a cannon to kill a fly.

Model D — Integrated retail platform

A solution that already brings POS, back office, e-commerce synchronisation, BOPIS and ship-from-store designed to talk to each other from the outset, connected to the ERP through a single point. This is the case of a platform such as MAXIRETAIL, especially when retail is the core activity and not an appendage of production.

The structural advantage is that the internal integration comes already built and tested by the vendor. You are not gluing together pieces from different origins — you are using a set designed to work together, with a single point of connection to the financial back office. The risk is the inverse of point-to-point: less flexibility for very specific cases, in exchange for much less fragility.

Model A is not an architecture — it is a debt taken on in the name of a fast start, with interest paid on every new channel.

The trade-off nobody wants to discuss

Real-time synchronisation costs money and complexity. Batch synchronisation is cheap but creates the ghost sale. The honest question is not "do we want real time?" — everybody does. It is "which items justify real time and which tolerate five minutes of delay?". For seasonal footwear with unique pairs, real time. For slow-turnover consumables, five minutes will do.

How to decide the granularity of synchronisation

The sensible answer is to segment by stockout risk and value. An item with stock of thousands of units and daily replenishment tolerates batches. An item with one or two units, seasonal and expensive, requires instant reservation at the moment of purchase. Defining this policy — which most companies never do explicitly — is half the work of a well-executed omnichannel project, and it is what separates an implementation that breathes from one that chokes at every traffic peak.

The mistake of underestimating merchandise returns

Everyone designs the sales flow carefully and forgets the reverse flow. Returns are, in much of fashion retail, a significant fraction of online sales. An item returned in a shop has to go back to the master inventory as available, with status control — and, in footwear and apparel, with verification that the box and product come back sellable. Poorly designed reverse logistics is where the margin of omnichannel disappears in silence.

5. How to assess whether your company needs it

Not every company needs full omnichannel tomorrow. It needs to know where it stands and what the next defensible step is. This diagnosis serves that purpose.

Signs that you are already losing money

• You receive regular complaints from customers who saw availability online and did not find the item in the shop.
• The month-end close includes an inventory-reconciliation effort between channels that consumes more than a day of work.
• Online returns cannot be processed in the shop, generating friction and reverse-logistics cost.
• You cannot answer, in a single report, the question "how much did we sell of this SKU across all channels this month?".

Step by step

1. Map the channels and their inventories. List each sales channel — shop, site, marketplace, B2B — and identify which stock database each one consults. If there is more than one, you have found the problem.
2. Measure the inventory divergence. Pick 50 SKUs and compare the quantity that each system reports at the same instant. The percentage of divergence is your honest starting point.
3. Quantify customer friction. Count, over a month, the complaints and the returns blocked by channel incompatibility. Assign an average lost-sale value to each.
4. Assess current certification and compliance. Confirm that the POS is AT-certified, that it reports SAF-T monthly and that invoicing complies with DL 28/2019 with ATCUD.
5. Define the anchor channel. Decide which channel represents the largest share of revenue and make it the starting point for unification — usually the physical shop in Portugal, not the site.
6. Estimate the ERP integration effort. Check whether the current ERP exposes APIs or integration points for inventory, customers and orders. The answer determines whether the path is evolution or replacement.

If you cannot measure the inventory divergence between channels, you don't have an omnichannel problem — you have a problem whose size you don't even know.

The Portuguese decision trio: CEO, CFO and the IT hero

In the family company of the North, the decision almost always passes through three people: the CEO, who wants growth; the CFO, who wants control of the investment and its return; and the IT lead, who is often a self-taught person with fifteen years of deep knowledge of the business and no diploma to certify it. Ignoring any of these three is to guarantee derailment. The IT hero, in particular, knows where all the bodies from previous projects are buried — and their buy-in is worth more than any vendor presentation.

Quick wins before the big project

• Enable a consolidated sales report by SKU adding up all channels — even if it is built at night by batch, it gives management immediate visibility.
• Allow returns in the shop for online purchases with a documented manual procedure, while the integration is not yet in place. It resolves real friction in two weeks.
• Unify the customer record and the loyalty card between shop and site, even if the synchronisation is daily. The customer stops being treated as two.

What NOT to do at the start

Don't try to solve all channels at the same time. Don't force real time on the whole catalogue out of technical vanity. Don't do the big rollout at the launch of the collection or at the sales peak — start in the dead months. And don't treat the project as an IT project: it is an operations project that uses IT. Whoever confuses the two things delivers a technically correct system that nobody on the shop floor uses.

6. What to choose and why

The right decision depends on two variables: the number of points of sale and whether retail is the core activity or a channel for selling off own production. A clothing factory that opens three own-brand shops has different needs from a food-retail chain with 22 shops.

Decision matrix by profile

For the industrial SME that also sells to the public

This is the most common case in the North: a footwear company or clothing company that manufactures for third-party brands and also sells its own brand in-store and online. Here, the vertical ERP has to be the core, because inventory is born in production. A MULTI ERP that already models the three axes of colour-size-fitting solves the structural problem that generalist ERPs cannot — and connects to the retail POS without duplicating the item master.

The subtlety here is that production and retail compete for the same stock. An item destined for a third-party brand order cannot be sold in the own shop. The vertical ERP has to know how to allocate: this batch is for Inditex, this one is for the own shop, this one is free. A pure retail system does not understand this distinction — only an ERP that was born in textile or footwear production models it natively.

For pure retail

When selling to the public is the business — not an appendage — the retail platform is the core and the ERP stays in the financial back office. Here the priority is the certified POS, synchronisation with e-commerce, and the convenience functions: BOPIS, ship-from-store, loyalty. BI keeps track of everything, and this is where a layer of Qlik Sense dashboards turns scattered data into purchasing and replenishment decisions.

The role of the warehouse in the choice

None of these choices survives a disorganised warehouse. If the physical stock does not match the logical stock in the central warehouse, omnichannel merely propagates the error to more channels, faster. That is why a WMS and picking solution that ensures accuracy in the warehouse is often the silent prerequisite of any serious omnichannel project. The truth of the master inventory is only worth as much as the physical count that feeds it.

The question is not "which omnichannel software to buy". It is "where is my inventory born" — and the answer defines the entire architecture.

7. Regulatory framework and applicable compliance

Omnichannel in Portugal does not live outside the law. Every sales channel that issues invoices is subject to the same obligations — and the unification of systems is often an opportunity to tidy up compliance that was scattered.

Invoicing, certification and SAF-T

DL 28/2019 governs electronic invoicing, requires ATCUD on each document and mandates the use of AT-certified programs. Portaria 195/2020 defines the monthly SAF-T reporting. In an omnichannel architecture, this means a critical decision: who issues the invoice, the POS or the ERP? Both have to be certified, and the numbering cannot collide between channels.

This numbering question seems technical, but it blows up in an audit. If each shop's POS and the e-commerce issue invoices in independent series without a clear policy, it creates a reconciliation nightmare and a compliance risk. Good practice is to define series by channel and point of sale, all certified, all reported in the monthly SAF-T, without overlap.

Customer data and GDPR

Unifying the customer record between shop, site and loyalty is exactly the kind of processing that the GDPR and Law 58/2019 regulate. Consolidating data from channels that were previously separate requires a legal basis, information to the data subject and coherent retention policies. It is not an obstacle — it is hygiene. But ignoring it in the integration phase creates liability.

There is a frequent trap: cross-referencing shop purchase history with online browsing behaviour for marketing purposes without an appropriate legal basis. The technical unification is legitimate; exploiting the unified data for purposes the customer does not know about is not. Document the purpose of each processing operation before joining databases — once joined, separating responsibilities is much harder.

Cybersecurity: when NIS2 applies

An omnichannel architecture increases the exposure surface: POS connected to the network, e-commerce exposed to the internet, permanent integrations. Depending on size and sector, NIS2 (Directive EU 2022/2555, transposed by DL 65/2025) may apply, imposing risk-management and incident-notification requirements. It is worth reviewing the framing in our guide to NIS2 compliance for industrial SMEs before exposing more systems to the internet.

• Certified and up-to-date POS and e-commerce — AT certification does not dispense with vulnerability management.
• Network segmentation between the shop environment, the warehouse and the financial back office.
• Tested backups of the master inventory — if the core falls, all channels fall.
• Qualified eIDAS signature for documents that require it, manageable in a document-management layer.

Every new channel is a new door. Omnichannel that does not think about cybersecurity is opening doors without counting how many locks it has.

Sustainability and traceability — what is coming

For fashion brands that also sell omnichannel, a new layer of demand is approaching. The EU Strategy for Sustainable and Circular Textiles points towards the digital product passport, which will link each item to its composition, origin and traceability. A well-structured master inventory, with batch and provenance, is the foundation of that future compliance. Whoever unifies inventory today is, without knowing it, preparing for what regulation will require tomorrow.

Funding: PT2030 and PRR

Retail digitalisation and systems-integration projects are frequently eligible for instruments such as PT2030, COMPETE 2030 and Norte 2030. What usually stalls applications is not merit — it is the poorly substantiated technical report and the absence of measurable result indicators. Defining metrics of inventory divergence and lost sales, as in the diagnosis above, strengthens the application.

The typical application mistake is to describe the purchase of software without linking that purchase to a measurable transformation result. An evaluator does not fund "buy a retail platform"; it funds "reduce inventory divergence from X to Y and increase cross-channel revenue by Z". Having the diagnosis numbers before submitting is what distinguishes an approved application from one returned for reworking.

8. How INFOS approaches this

We have worked for more than three decades with the industry and retail of the North of Portugal, and our conviction is born from there: omnichannel is solved in the order of inventory, not in the order of marketing. That is why we always start with the question of where the stock is born — in production, in the warehouse or in purchasing — and we design from that core.

For the industrial SME that also sells to the public, we connect the MULTI ERP to the POS and to the e-commerce with a single master inventory, integrating production and retail without duplicating items. For retail as the core business, MAXIRETAIL brings POS, back office, BOPIS and ship-from-store from the outset. Warehouse management relies on a WMS and picking solution, and the professional channel on a B2B order portal. All of this is read in BI dashboards that show consolidated sales by SKU and channel.

We don't sell revolutions. We prove with a proof of concept focused on a real problem — inventory divergence, typically — and we scale from what works. The warehouse manager who has already seen three projects fail is right to be suspicious; our answer is not to take him off the radio more than strictly necessary.

Why we start with a proof of concept and not the big contract

A focused proof of concept costs little, risks little and proves or disproves the thesis quickly. We choose a concrete problem — usually synchronising the inventory of a pilot shop with the e-commerce and measuring the reduction in divergence in six to eight weeks — and we show the number. If the number convinces the CFO, it scales. If it doesn't convince, a big investment in the wrong path has been saved. It is the honest way to work with those who have already been burned by big promises.

What we learned from projects that went wrong

We learned, sometimes the hard way, that the most common mistake is not technical — it is one of sequence. Companies that wanted to switch on BOPIS before having a reliable master inventory delivered a feature that generated more complaints than sales, because it promised stock that did not exist. The lesson was clear: first the truth of the inventory, then the convenience for the customer. Reversing this order is to accelerate towards the wall.

9. 30/60/90-day roadmap

An omnichannel project is not launched all at once. It is structured in phases with verifiable milestones, so that the board sees value before committing the full budget.

Days 1-30: single truth about the inventory

• Map all channels and the inventories each one consults, and measure the real divergence on a sample of SKUs.
• Define the anchor channel and the inventory that will become the master.
• Confirm invoicing compliance — AT certification, ATCUD, SAF-T — across all channels that issue documents.
• Implement the quick wins: consolidated sales report and documented in-shop returns.

Days 31-60: unification of the core

• Establish inventory synchronisation between the master and the anchor channel, defining which SKUs require real time and which tolerate batches.
• Unify the customer record and loyalty between shop and site, with a documented GDPR legal basis.
• Segment the network and review the cybersecurity framing before exposing more systems.
• Run a BOPIS pilot in a single higher-volume shop.

Days 61-90: scale and convenience

• Extend BOPIS to the set of shops and enable ship-from-store in those with suitable stock and logistics.
• Connect the B2B channel to the same master inventory, if applicable to the business.
• Implement BI dashboards with consolidated sales, inventory divergence and sales lost to stockouts.
• Document the measurable results — reduction in divergence, increase in cross-channel sales — for the funding application and for the decision to scale.

The milestones the board should demand to see

Each phase has to end with a number, not with a feeling. At the end of the first 30 days, the measured inventory divergence. At 60, the reduced divergence in the pilot shop and the number of in-shop returns processed. At 90, the cross-channel revenue attributable to BOPIS and the estimated reduction in ghost sales. If a phase ends without a number, the project has lost its bearings — and it is time to stop and ask why before spending more.

After 90 days you will not have full omnichannel. You will have something more useful: proof that it works and a number that justifies the next step.

After the 90 days

What comes next depends on what the numbers say. It may be extending the endless aisle to turn each shop into a shop window for the entire catalogue. It may be fine-tuning the synchronisation policy by item segment. It may be integrating marketplaces. The important thing is that each next step is justified with data from the previous step, and not with the next fad to arrive from marketing. Mature omnichannel is built this way — by accumulation of proof, not by leaps of faith.

Sources

• Eurostat — E-commerce statistics for individuals (online purchasing indicators by EU country).
• INE — Survey on the Use of Information and Communication Technologies in Enterprises.
• Decree-Law No. 28/2019 — invoicing regime and reporting obligations to the AT.
• Portaria No. 195/2020 — monthly reporting of the SAF-T (PT) file.
• Regulation (EU) 2016/679 (GDPR) and Law No. 58/2019 — protection of personal data.
• Directive (EU) 2022/2555 (NIS2), transposed by Decree-Law No. 65/2025 — cybersecurity of essential and important entities.
• EU Strategy for Sustainable and Circular Textiles — European Commission (official documentation of the strategy and of the digital product passport).
• Portugal 2030 / COMPETE 2030 / Norte 2030 — business digitalisation funding programmes (official documentation of the programmes).

Frequently asked questions

What is omnichannel retail and how does it differ from traditional e-commerce?

Omnichannel retail integrates all sales channels (physical shop, online, mobile) into a single experience, with unified inventory in real time. It differs from traditional e-commerce because it does not treat the online shop as an isolated channel, but as an extension of the same operation. The customer sees the same stock, prices and offers at any point of contact, and can buy online and collect in-store or return online purchases physically.

Why does omnichannel fail in Portuguese retail chains?

The failure occurs at the operational level, not the conceptual one. Most chains have legacy systems — a local POS in each shop, e-commerce on another server, ERP at head office — that were never designed to share data in real time. The inventory synchronises with a delay of hours or days, creating three different "truths" about stock. The root is architectural: without a single master data set, each system believes it knows the available stock, and none really does.

How does the inventory problem specifically affect the footwear sector?

In footwear, a unique SKU is defined by colour, size and fitting. A collection of women's shoes may have 800-1,200 active SKUs. If the system does not synchronise, the black size 37 pair may be available online while the 38 in the same colour is sold out — but the customer does not know it until they reach the shop. In seasonal fashion, a pair may be unique in stock, with no replenishment possible. Synchronisation errors result in double promises of the same item to two customers.

What is the real cost of a sale that is not recorded in the system?

It is a "ghost sale": the customer sees stock online, goes to the shop, doesn't find it, and leaves. It leaves no trace in the system because the sale never occurred. It differs from a visible stockout because it does not appear in reports, destroying invisible margin over the year. More serious still is the erosion of trust: the first failure blames the system; the second blames the brand. This damage accumulates and only shows up months later, when it is difficult to link cause to effect.

Why does local inventory in each shop create divergences?

Each shop has its POS with a local database, often disconnected from the central system. When an item is sold in a shop, that information takes hours or a day to reach the ERP and the e-commerce. Meanwhile, another customer buys online based on stock that no longer exists. The problem is compounded because the e-commerce and the ERP have separate item masters, creating different nomenclatures for the same product. Without continuous synchronisation, divergence is inevitable.

How does omnichannel returns work and why is it important?

In an omnichannel chain, the customer can return an online purchase directly in a physical shop, without the need for postal shipping. This reduces friction, increases the accepted-return rate (because it is easier), and improves the experience. Without omnichannel, the customer is forced to package and post, which discourages legitimate returns and harms trust in the brand. In-shop returns also feed the physical inventory more quickly, reducing discrepancies.

What is the impact of a non-integrated loyalty card?

If the loyalty card accumulates points only in the shop but does not recognise them online, the customer is treated as two different people. This fragments the purchase history, makes coherent personalised offers impossible, and frustrates the customer who expects continuity. In addition, it prevents unified analysis of customer behaviour, missing opportunities for cross-sell and retention. Loyalty integration is fundamental to turning scattered data into commercial intelligence.