On paper, a PLM is a workflow tool. It’s where tech packs get built, where sampling gets tracked, where approvals get chased, where the product moves from concept through to production with everyone involved able to see where it is and what it needs. That’s what it was designed to do, and that’s how most vendors still sell it.

Talk to the ops leaders actually running PLM implementations right now, though, and a different picture starts to emerge. The workflows are still happening. But they’re not necessarily happening inside the PLM.

What is happening inside the PLM is something more foundational. It’s becoming the product data master. The single source of truth for every attribute, every measurement, every material, every supplier reference, every sustainability data point, every piece of structured information that needs to exist on a product. And from there, it’s feeding everything else.

Shopify pulls from it. NetSuite pulls from it. Production trackers, sampling systems, allocation tools, reporting dashboards. All downstream. All dependent on the PLM having clean, structured, accurate data at the centre.

“We don’t really do the workflows in it,” one ops lead told us recently. “But if the PLM is wrong, everything else is wrong.”

Subscribe now

That’s the shift worth paying attention to.

For a long time, the case for investing in a PLM was built around efficiency. Faster sampling cycles, fewer email chains, better visibility across the product calendar. All real, all valuable, but all essentially productivity arguments.

The newer case is structural. A brand that can’t describe its own products in structured data can’t do any of the things that now matter most. It can’t populate a digital product passport. It can’t feed an AI model that’s trying to help with demand planning. It can’t integrate cleanly with a new channel partner. It can’t answer sustainability questions from a regulator or a retailer. It can’t scale without the data layer underneath it becoming the bottleneck.

The brands that are getting this right are the ones who’ve stopped thinking about their PLM as a project management tool and started thinking about it as infrastructure.

The distinction matters because it changes what you’re optimising for. If the PLM is a workflow tool, you care most about user experience, approval flows, notification logic. If it’s a data master, you care most about the structure of the data model, the quality of the integrations, the discipline around tagging, the governance of who can change what. Different priorities. Different implementation. Different definition of success.

One head of IT at a contemporary fashion brand put it simply when we spoke: “The PLM is where truth lives. Everything else is a view.”

Subscribe now

That framing is useful. It means the PLM doesn’t have to win the UX battle with a designer’s preferred sampling spreadsheet, or a production manager’s WhatsApp group, or a sustainability lead’s bespoke tracker. Those tools can all keep existing. What the PLM has to do is be the definitive, structured, trustworthy version of what the product actually is. Everything else orbits around that.

It’s a quieter role than the one the category was originally sold on. But it’s arguably a more important one.

For brands thinking about implementing a PLM, or re-implementing one that stalled, the question to start with isn’t “how do we move our workflows into this system.” It’s “what data needs to live here, and what needs to happen to make sure it stays accurate.” Those are very different projects. The first is about change management. The second is about architecture.

And the second is what the operators actually getting value from their PLMs are really doing.

Wednesday. Hallé St Peters in Manchester. 50+ brand-side leaders in one room. CTOs, Heads of Ecom, COOs, Digital Directors, CFOs, Merchandisers, Ops Leads, IT managers. People who have actual jobs at actual brands and actually had to take a day out of their week to be there.

We didn’t sell tickets to vendors. We didn’t have a sponsor parade. We had brands.

And it worked.

Chris Perrins on AI and safety

The session that removed (perceived) barriers to getting started with AI and building the confidence through your org to make the impact last. Chris Perrins from Gymshark on AI and safety. The kind of talk you’d expect to be either preachy or a bit over-whelming. It was neither.

Honest. Practical. Grounded in what’s actually happening inside of brands at scale. Where the risk really sits. What the team are actually doing about it. Where the speed-versus-safety conversations land. The bits that look easy but aren’t, and the bits that look hard but are.

Half the room had their notebooks out. The other half had their phones out. Nobody was on email.

That’s the bar for a keynote. He cleared it.

6 Roundtables, 6 honest conversations

After the keynote, the room split into six tables. Here’s what landed at each.

Ecom and merch. The Inverted Trade Pack. AI surfaces the trade headlines Monday first thing. The team spends the day doing the actual work and building the actions. The Monday meeting stops being a recap and starts being a decision. The line that stuck: “Get AI to mark the homework before we walk in the room.” Not write the trade report. Check it.

Tech. Internal knowledge bases. Documentation that lives somewhere queryable, without a human in the loop of every interaction. Updates that share themselves. The easiest AI project most businesses have not built. Funniest line of the day came out of this table: “Copilot is what your nan uses. ChatGPT is what your teenagers use. Claude is what the workers use.”

Ops. Cross-functional landed cost and approval intelligence. The tool that turns a dozen Slack threads into one answer.

Commercial. AI-generated product mockups for customisations, driven off user input. Visualisation at the point of want, not six weeks later.

Ecom and digital product. Product creation and launch readiness. The bit nobody posts about. Everybody needs it.

Data integrity, governance and workflow. The table where someone said out loud, “AI is a dirty word in our office,” and the rest of the table immediately recognised the pattern.

Each table did the same exercise. Spot the opportunity. Make it viable. Handle the objections. Action plan. Share back. By the end of the day, every table had something concrete to take home.

The bit that surprised us

How honest people were.

“We’re a bit of a shitshow on this.” “We’re racing without thinking about the foundation.” “I don’t know if anyone in this room has it figured out.”

You don’t get that at a normal industry event. You get it at OX1 because the room was curated, the format was right, and the people who turned up trusted each other enough to say it out loud.

That trust is the thing we’re building.

The next one

OX2 is being planned. This time, London.

If you want to be in it: https://operatorexperience.ai/

Thank you, Manchester.

Join OX2: LDN

Are we profitable internationally, by market? What are the real return reasons by country? What’s the duties and tax impact on our margin? How much revenue are we actually retaining through exchanges?

The data to answer these questions exists. It’s sitting somewhere in your stack right now. The problem is that it’s sitting in four different systems that don’t talk to each other, reconciled by nobody, and interpreted differently depending on which team you ask.

That’s the analytics gap. And it’s one of the more quietly damaging findings in our latest research.

Subscribe now

It comes up consistently across both returns and cross-border operations, and it matters because it’s not just a reporting inconvenience. It shapes decisions, or more accurately, it prevents them. When you can’t see international profitability by market, you can’t make intelligent decisions about where to invest, where to pull back, or where margin is leaking fastest. When you can’t reconcile exchange revenue cleanly, you can’t build a business case for improving the exchange flow. When duties and tax treatment lives outside the core systems, finance teams are always a step behind.

Our consultant Amy Washington put the buying and merchandising perspective clearly: “What you need as a buyer or merch is the qualitative feedback. To know your returns rate is high, that’s a KPI, that’s a marker of performance, but what you really need to know is why. What a buying and merch team would be looking for from a returns platform is: can we get the analytics out of what is being returned and why? That’s what’s important.”

The returns rate number is available. The why is buried in a system that doesn’t surface it cleanly, if it surfaces it at all.

The same pattern shows up on the cross-border side. Brands are running international operations with meaningful revenue in multiple markets, but their view of margin by market is either non-existent or assembled manually from sources that may or may not be consistent with each other. They know they’re profitable overall. They can’t tell you whether Germany is worth it or whether the US operation is subsidising returns costs that nobody has fully accounted for.

What operators want, when you ask them directly, isn’t more dashboards. It’s one trusted operational view. A single place where the numbers reconcile, where finance and operations are looking at the same data, and where decisions can be made with confidence rather than approximation.

Subscribe now

That’s harder to build than it sounds. It requires integration work that most implementations don’t prioritise, reporting architecture that cuts across vendor boundaries, and a level of operational discipline that’s difficult to maintain when teams are already stretched. But it’s also the gap that matters most, because almost everything else in the operation, routing decisions, exchange strategy, market prioritisation, platform evaluation, depends on having reliable numbers underneath it.

The brands that are investing in this now, building the connective tissue between their systems and establishing a single source of truth for operational data, are building a genuine competitive advantage. Not because the technology is sophisticated, but because most of their competitors are still working from four spreadsheets and a Slack thread.

Our full research report covers the analytics gap in detail as part of a broader look at where returns and cross-border operations are failing, and where the biggest opportunities for improvement sit.

Read the full report

In this series, we’ve been following West Park Rovers, the fictional club from Kieran Maguire’s The Price of Football, as they bring retail in-house.

So far, we’ve covered the business case for making the move, the operational reality of running what is effectively a small factory, and in the last piece, The 12-Month Apocalypse, the brutal economics of a catalogue that expires every year.

That piece was about pressure with a fixed horizon. Time you can see running out.

This one is about the opposite.

Moments that arrive without warning. A goal. A farewell. A signing that shifts the mood of a whole city overnight. Demand that exists right now and will be gone by Thursday.

Seasonal retail has a calendar. Sports retail has a pulse.

Catch up on this series:

• Part I - West Park Rovers

• Part II - The Business Case

• Part III - Your Club Shop is a Factory

• Part IV - The 12 Month Apocalypse

• PartV - Moments not Seasons

When clubs review retail performance, the conversation almost always starts with finished products. Shirts, training wear, special editions, limited runs. Which items sold through, which sizes caused problems, which launches exceeded expectations.

That focus is understandable. Finished goods are visible. They are what fans recognise and what revenue reports are built around. They are easy to explain to boards and easy to benchmark season to season.

But when clubs bring retail and e-commerce in-house and begin to operate it as a real business rather than a licensed afterthought, something changes. The further you move into the operational detail, the clearer it becomes that the most important products in the system are not always the ones fans see.

They are the components that sit underneath personalisation. Letters, numbers, badges, and patches. The alphabet.

These items rarely feature in planning decks. They are cheap, unglamorous, and operationally inconvenient. Yet they quietly determine whether large parts of the retail operation function at all.

Subscribe now

The product fans buy is not the product you manage

From a supporter’s point of view, a personalised shirt is a single product. You choose a size, add a name and number, and check out. The complexity is invisinble.

From the club’s point of view, that same order is not a product. It is an assembly process.

A base garment must be available in the right size. The correct letters must exist in the correct font and size. Numbers must match both the shirt size and competition rules. Patches may be required or optional depending on context. Placement rules must be respected, and not every combination is valid.

Every personalised order therefore consumes multiple SKUs at once. If even one of those components is missing, the product is no longer fulfilable in the way the fan expects.

What makes this difficult is that the system often does not fail immediately. The shirt still appears available. The checkout still accepts the order. The problem emerges later, when fulfilment is delayed, substitutions are required, or customer service has to intervene.

These failures are subtle, which is why they persist.

Why personalisation only works if components are treated as critical inventory

Personalisation is often positioned as a feature. Something that can be added on once the basics are in place.

In practice, it is closer to a manufacturing capability.

It increases average order value. It strengthens emotional attachment. It turns moments on the pitch into commercial outcomes. But it also introduces a dependency on component-level inventory that many clubs are not set up to manage properly.

When letters, numbers, and badges are treated as background supplies rather than first-class inventory, the system becomes fragile. Stock may exist, but not in the combinations required to meet demand. Availability becomes theoretical rather than real.

This is not a question of better picking or more careful packing. It is a question of whether the business actually understands what it is capable of selling at any given moment.

Clubs that scale personalisation successfully are not the ones that offer the most options. They are the ones that model availability accurately and constrain it intelligently.

The smallest stockouts cause the most damage

Running out of a popular shirt size is obvious. It is visible on the site, visible in reports, and usually prompts action.

Running out of a specific letter, number, or patch is different. The site often still looks healthy. Orders continue to place. The issue only becomes visible once fulfilment begins.

At that point, the cost is already incurred. Delays lead to service work. Service work leads to refunds or goodwill gestures. Goodwill gestures erode margin. Repeated experiences erode trust.

The irony is that these components are usually inexpensive. They are easy to replenish in theory. Yet because they are shared across many potential products, a single stockout can block an entire class of sales.

This is why letters and numbers matter far more than their unit cost suggests. They are not revenue drivers in isolation. They are revenue enablers. When they fail, revenue does not disappear dramatically. It leaks quietly.

Why this is not a warehouse problem

It is tempting to frame these issues as operational execution problems. Counting errors, picking mistakes, or warehouse discipline.

That framing misses the point.

The root issue is almost always upstream, in how availability is modelled and understood. If systems only track finished goods, they provide an incomplete picture of reality. They tell you how many shirts exist, not how many sellable configurations are possible.

Those two numbers are rarely the same.

Clubs that operate retail in-house for long enough begin to plan from components upwards. They understand that the true constraint in the system is not demand or even garment supply, but the weakest link in the component chain.

This shift is uncomfortable because it forces the organisation to confront complexity rather than abstract it away. But it is also where operational resilience comes from.

How the alphabet becomes a strategic bottleneck

The importance of components becomes even clearer when demand concentrates around specific moments.

Sport does not create evenly distributed demand. It creates spikes. One player, one performance, one narrative can dominate attention overnight.

When that happens, demand does not rise across the entire catalogue. It focuses sharply on a particular name and number combination. If the components required for that combination are unavailable, the moment is only partially captured.

Nothing crashes. There is no obvious failure. The club may even feel that it handled the moment well. But in reality, part of the opportunity was lost without being measured.

This is why many clubs underestimate how much revenue they fail to capture during peak emotional moments. The limitation is rarely visibility or traffic. It is fulfilability.

Subscribe now

Why this rarely reaches board-level discussion

Most retail reporting is designed to be legible rather than complete. Revenue by category, margin by range, sell-through by product. These metrics are useful, but they flatten reality.

They do not show how often demand was blocked by missing components. They do not surface how many orders were delayed or modified due to stock constraints at the component level. They do not capture the cost of quiet failure.

Without that data, leadership assumes the system performed as intended. From a distance, it looks like it did.

In practice, the system may have been operating below its true potential for months.

Treating the alphabet as infrastructure

Clubs that move beyond this stage stop thinking about letters, numbers, and badges as accessories. They start treating them as infrastructure.

They track them with the same seriousness as finished goods. They plan replenishment deliberately. They model availability at the configuration level. They design storefront logic to reflect what can actually be fulfilled.

Once that happens, personalisation stops being a source of operational risk and becomes a durable advantage. Moments are captured more reliably. Customer trust improves. Margins stabilise rather than erode quietly.

This is not about adding complexity for its own sake. It is about acknowledging where complexity already exists and managing it properly.

The short version

In sports retail, growth is not constrained by what you sell. It is constrained by what enables the sale.

Letters, numbers, and badges rarely get attention because they are not exciting. But they define the limits of the system more than any shirt ever will.

Clubs that recognise this build personalisation that scales. Clubs that don’t continue to lose revenue in ways that never quite show up on a spreadsheet.

In the final part of this series, we’ll look at what happens when this level of operational understanding is fed back into the wider organisation, and how retail data starts to influence decisions far beyond the shop.

Average AI confidence score across the group: 5.3 out of 10.

60% have no formal AI policy in place, including at businesses where AI use is actively encouraged. The most common blocker isn’t cost or capability or someone at the top saying no. It’s just not knowing where to start.

Subscribe now

The things people actually want to use AI for aren’t ambitious. Automating trade meeting notes. Cleaner forecasting. Product descriptions that don’t take half a day. Weekly reports that don’t eat a morning. Operational problems that have existed for years and now look like they might finally be solvable. Nobody is trying to reinvent anything. They just want Monday back.

The other thing that came through clearly: everyone is experimenting, and nobody knows what anyone else is doing. Someone in ecommerce has built something quietly. Someone in ops has a workflow running in the background. One respondent described their current situation as “fragmented and manual, with valuable knowledge spread across teams.” The shared learning that would make all of it faster and less risky isn’t happening because there’s nowhere for it to happen.

That’s what OX1 is for. But we should be straight about what it is and isn’t.

Everyone involved in building this aren’t standing at the front of the room with the answers. Nobody has AI solved right now. The technology is moving too fast for that to be true of anyone, regardless of what their LinkedIn bio says. What the people in the room have is practical experience, honest questions, and no commercial reason to pretend otherwise.

The question sitting underneath a lot of the survey responses was this: “is this replacing me or my team?” That question doesn’t get asked out loud very often. It should. It’s the room OX1 is trying to create.

If your confidence is sitting around a five, you’re in good company. If you’ve got something running in the background that nobody else in the business knows about, same. If you’re not sure whether your business is ahead or behind, you’re probably right not to be sure, because the reference points don’t really exist yet.

That’s not a reason to wait. It’s a reason to be in the room.

16th April. Hallé St Peters, Manchester. Brand-side operators only.

operatorexperience.ai

RSVP for OX1: Manchester

So many conferences, roundtables, and events you've been invited to this year was built around agencies, platforms, or vendors. The brand-side operator - the person running ops, owning the tech stack, holding the P&L together - is either an afterthought.

OX1 Manchester on 16th April is being built for a different kind of room.

What it actually is

Operator Experience is an invite-only community where brand-side leaders learn and build AI-era capabilities together. No agencies pitching. No platforms grandstanding. A closed network of people in the same seat, dealing with the same pressures, trying to figure out the same things.

OX1 is the first event. It runs from 12:30 at Hallé St Peters in Manchester, and the afternoon is structured around one practical challenge: what if your brand committed to automating 30% of all operational tasks by 2030?

Working in small squads, attendees will map end-to-end processes, identify automation opportunities, and leave with a concrete initiative - not a slide deck, not a framework to file away. Something you can actually take back on Monday.

The speakers worth knowing about

Chris Perrins is one of the confirmed speakers for the AI Fundamentals session. His take is worth paying attention to. The argument isn't that AI is coming - it's that it's already there, whether your business has formalised it or not. Data loss tools at Gymshark showed AI being used across the business long before any policy existed to govern it.

His framework is pragmatic rather than theoretical. Prioritise confidence over governance. Separate experimentation from production (vibe coding is not production-ready, and the "next Excel problem" - important tools nobody properly owns - is a real risk). And create a structure where ideas can spread safely, through champions in each function rather than controllers at the top.

The goal, as he puts it, is confident and responsible use, not perfect governance. That's a more honest framing than most AI content you'll read this year.

He's joined by Luke Hodgson and Lauren Muir as hosts, with the full agenda shaped by community members who've sat in the same roles as the people in the room.

Tables are filling. The event is running six ten-person tables and is already operating a waiting list for the afternoon session. Fashion photographer Nathan McDowell will be there to capture the day, so attendees leave with professional shots as well as a concrete plan.

After the workshop, the evening continues at The Castle on Oldham Street. A proper pub. A proper crowd. Strictly brand-side. No pitching.

That last detail matters more than it sounds. The conversations that happen when nobody in the room has a commercial agenda are a different category of conversation entirely.

Who's behind it

OX1 is free to attend for brand-side leaders. That's made possible by partners who've committed to a no-promotion, no-posturing environment. Headline sponsor Dema, event sponsors Loop, Torque, and Emfas, and community partners including Vervaunt, High Cohesion, Whanau, Centra, Resourced, and Commerce Thinking.

OX2 is already confirmed for July. This is not a one-off.

If you're the right person

Directors, heads of function, and senior leaders working brand-side in operations, finance, ecommerce, merchandising, buying, or product data at a fast-growing fashion, apparel, or consumer brand.

Places are limited and curated deliberately. If you want one, move quickly.

operatorexperience.ai

Open your inbox. Scroll through the vendor pitches that have landed in the last six months.

Count how many have a .ai domain. Count how many claim to be “the first AI-native [something] platform” or position themselves as the “AI-first” solution in their niche.

This is the current state of B2B software positioning, and it’s working.

Deals are getting done. They’re genuinely impressive in a demo, they move fast, and they’re being funded aggressively.

Fast growing businesses are adopting these new-breed tools and attempting to run critical operations on them.

But there’s a big, unspoken risk attached to such a move…

There are going to be far more venture backed businesses with .ai domains that fail in the next five years than succeed. That’s just how venture maths works at scale. And right now, fast-growing retail brands are considering building their operational backbone around these tools as if that reality doesn’t apply to them.

But it does.

The numbers behind the narrative

BLS (Bureau of Labor Statistics) data shows that 20% of newly created businesses don’t survive their first year. 50% are gone within five. Only 34.7% of businesses born in 2013 were still operating a decade later.

That’s not software-specific. But before you sign a multi-year contract with a company that was incorporated eighteen months ago and has a founder who’s never shipped an enterprise product before, it’s the right place to start.

The instinct is to dismiss this. “We’re not buying from some seed-stage startup. Our vendors have raised serious money.” That’s where the thinking tends to go wrong.

Venture economics are built on extreme skew. Only 1 in 140 VC-backed companies ever becomes a unicorn. The model assumes a small minority of portfolio companies become massive winners and a large majority don’t. That maths works fine for a fund managing across dozens of bets. It’s a terrible basis for choosing the system that runs your inventory, your finance, or your order orchestration.

Fashionable and well-funded is not the same as durable. The more capital floods into a single narrative, the more companies that narrative attracts - including a lot that won’t survive the consolidation that always follows a hype cycle. Funding is not a durability signal. It’s a signal that someone made a calculated bet on upside. Those are different things, and the retail industry keeps confusing them.

The question missing in the evaluation room

Being first to adopt an emergent tool in your marketing stack is a calculated risk worth taking. Being first to run your inventory, your finance, and your fulfilment on one is a different kind of bet entirely.

If it goes wrong, you’re not dealing with a campaign that underperformed. You’re dealing with an operational crisis, a migration under pressure, and a vendor that may not have the resources to support you through either.

The vendor evaluation process at most brands goes the same way. Features. Pricing. Integration complexity. Maybe a reference call with a brand two sizes bigger than yours. Nobody asks whether the company will still exist in three years. Nobody models what a forced migration would cost. Nobody asks what happens to their data if the Series B doesn’t close.

Those are the questions that should be leading the conversation. Especially now.

Subscribe now

Private equity isn’t necessarily the safe alternative you think it is

The counter-argument usually goes: “We don’t buy from scrappy startups. Our vendors are PE-backed.”

Moody’s data shows PE-backed companies defaulted at roughly double the rate of non-PE-backed peers between 2022 and 2024. Even among the largest, most established sponsors, default rates sat around 14%. S&P counted over 80 US PE portfolio company bankruptcies in 2024 alone.

What retail operators should actually be doing

The problem isn’t that brands are buying software from young or PE-backed companies. Some of the best tooling available right now comes from exactly those places. The problem is that the evaluation process doesn’t match the stakes.

If a vendor sits in the control layer of your business - finance, inventory, order management, replenishment - the bar for selection and ongoing management should be different from a tool running your email campaigns or your content.

For the control layer, the questions that actually matter:

On operational continuity. If this vendor ceased trading tomorrow, what’s your data extraction plan? How long would a migration take, and what would it cost? Have you modelled that scenario, or is it just uncomfortable to think about?

On financial durability. When did they last raise, and what does their runway look like? What’s their path to profitability?

Is it really worth the risk. When compared to established vendors in the space, do the feature and speed benefits really justify the inherent risks?

Most brands can’t answer those questions for their current vendors. That’s the real exposure.

The contrarian position

The standard advice is to avoid young vendors and stick to established platforms. That’s not quite right either.

The right frame is: match vendor maturity to workflow criticality. New, AI-native tooling can earn a place in your stack. Just maybe don’t put it at the centre of your entire operation just yet. Run your experimentation, your personalisation, your content tooling on the exciting new thing. Run your finance and inventory on something with a 15-year track record, a deep implementation ecosystem, and a balance sheet that’s survived at least one economic downturn.

The brands getting this wrong aren’t naive or foolish. They’re running an evaluation process that assumes software companies to be more durable than the data suggests they actually are.

Reliability is about to get expensive again. The vendors that survive the next round of consolidation will charge more for it, and buyers will pay, because they’ll have learned the hard way what the alternative looks like.

The uncomfortable truth

Somewhere right now, a founder is in a board meeting celebrating a new ERP go-live. The vendor they chose has eighteen months of runway and a Series B that hasn't closed. Nobody in that room knows. That's the problem.

But when they break, they break in exactly the same places.

That’s one of the more useful findings from our latest research, and it’s worth unpacking because it changes how you diagnose operational problems and where you look for the fix.

Most ecommerce stacks are layered. And the layers don’t fail evenly.

At the surface is what customers see: checkout flows, returns portals, localised pricing, branded experience. These layers tend to work well. They’re visible, well-supported, and heavily optimised for conversion and customer satisfaction. When brands invest in new tools, this is usually where the improvement shows up.

Beneath that sits the logic layer: rules, policies, routing decisions, market configurations, exchange logic, exceptions. This is where complexity starts to compound. Edge cases become day-to-day reality. And the assumptions baked into the platform start to show their age.

Underneath both is the operational core: inventory accuracy, landed cost, duties and tax treatment, margin attribution, financial reconciliation. This layer is the hardest to see, the hardest to change, and the least forgiving when something goes wrong.

When brands add new tools, they usually improve the top layer. When they scale internationally, the strain shows up in the layers below.

The friction points our research surfaced were identical across both returns and cross-border: ERP connections requiring manual reconciliation, WMS integrations that don’t reflect real-time inventory, OMS workflows creating duplicate orders or missing data, Shopify Markets configurations that don’t align with operational reality. It doesn’t matter which part of the operation you’re talking about. The failure mode is the same.

And then there are the workarounds.

Every brand we spoke to has at least one spreadsheet, one Slack channel, or one person who just knows how to fix the thing that should work automatically.

As our co-founder Andy Slater put it: “Every brand has temporary fixes that eventually became permanent. You’ll find that last Black Friday something broke, so a plaster got applied to make it work, and then that plaster just becomes part of the business. No one questions it, no one documents it, and it only surfaces again when something else breaks or the one person who built it leaves.”

That’s not a technology failure. It’s an organisational one. And it’s almost impossible to see from the outside, or even from the inside, until something forces it into view.

The reason this framing matters is that it shifts where you look when something isn’t working. If returns feels broken, the instinct is to evaluate returns platforms. If cross-border margin is leaking, the instinct is to interrogate the cross-border provider. Sometimes that’s the right call. But more often, the problem is in the connective tissue: the integrations, the routing logic, the manual processes that have hardened into infrastructure.

Fixing a layer-one problem when the failure is in layer three doesn’t solve anything. It just gives you a better-looking surface on top of a fragile foundation.

Our full research report maps this in detail across both returns and cross-border operations, including where the shared failure points show up most consistently.

Read the full report

That’s the dangerous bit. ERP can fail without a big bang and an error message. It fails quietly, then becomes the reason teams keep a separate spreadsheet and route workflows around the system instead of through it.

If you’re in that limbo, don’t default to blaming the software. In practice, the problems can be self-inflicted. You can have a great system and a great SI, and still end up with something that technically works but nobody trusts because it doesn’t reflect how the business actually runs.

Subscribe now

The quiet failure pattern

The first sign you’re off track is when the project starts optimising for sign-offs instead of outcomes. Milestones get ticked off, demos look slick, but nobody can walk a real order through the full flow without wincing.

The internal tell is confidence. If every test reconciliation is throwing up surprises, and confidence in the system is evaporating, that’s not teething issues. That’s a project that might not recover.

This is why brands without an in-house ERP operator get caught out. Each partner can genuinely deliver their slice, but if you haven’t got enough internal expertise you don’t know how to challenge it.

And the ugly truth is the in-between bits are where things break. Orders moving from your ecommerce platform into ERP, being picked in the warehouse, and landing back into finance is the chain where issues hide. A missing field, a misaligned mapping, a forgotten checkbox, and suddenly your operation breaks in a place nobody owns.

So if your project status looks green but your team’s behaviour is screaming red, listen to the behaviour. Workarounds aren’t coping mechanisms; they’re your warning system.

The stuff that actually derails go-live

If you’ve effectively outsourced decision-making to your vendor and your SI, you’re already late. Training isn’t the same as confidence, and confidence doesn’t appear by magic. It shows up when someone inside the business has the authority and experience to push the system further, and to keep pushing when it gets uncomfortable.

The pitfall here is thinking you can buy capability. You can’t. You either bring in an in-house operator early, or you choose partners who transfer knowledge rather than hoard it, and you set that expectation from day one.

The next sign is capacity theatre. Everyone says they don’t have time, but the real issue is you haven’t made room for the work.

You can’t implement something like ERP by squeezing it in around the edges. That’s not a time problem. That’s a capacity problem.

If your finance lead is meant to redesign the close, run UAT, and still hit month-end with a skeleton team, you’re not running an implementation. You’re setting yourself up to cut corners, miss issues, and then pay for it later.

Then there’s data. Not data in the abstract. Your product data, your SKUs, your attributes, your categorisation, the stuff merch and product dev live and die on.

If data is incomplete, outdated, or inaccurate, the outputs are useless. And if you’ve got no clarity on who owns that data across the business, you’ll end up duplicating entry, arguing about definitions, and migrating a mess with a nicer UI.

The pitfall is leaving product data clean-up as an IT workstream you’ll get to later. You won’t. You’ll hit cutover, discover items don’t behave properly, and you’ll be trying to fix it at the exact moment the business needs stability.

Timeline pressure is the other classic. Vendors will happily sell you speed, and the business will happily attach go-live to a milestone because it feels decisive. The cost of that decisiveness lands on you when quality gets traded away to hit a date.

Missing a vendor’s deadline is frustrating. Missing your own readiness is catastrophic.

If you’re hearing “can we go live in 90 days,” you’re asking the wrong question. The right question is what needs to be true for you to go live in 90 days.

That shift matters because it forces the awkward conversation. If sponsor attention, capacity, integrations, and data readiness aren’t there, the answer isn’t cross your fingers. The answer is to move the date.

Testing is where all of this gets exposed. If UAT is mostly happy-path clicking, you’re setting yourself up for the sort of failure that doesn’t even throw an alert.

Integrations can fail quietly. Orders can stop flowing, postings can be incomplete, and you won’t know until someone reconciles properly and realises things don’t tie out. That’s when the panic starts, because you’re now debugging while customers are waiting and finance is trying to close.

Subscribe now

That’s why end-to-end testing beats vendor sign-offs when you’re testing whether your business can trade. You’re not testing modules, you’re testing whether your business can trade.

And don’t kid yourself that some functions can join later. If finance isn’t engaged during implementation, then relies on the ERP for reporting, you can end up with the data wrong because the system didn’t meet their needs and nobody caught it in testing.

By the time you realise, it’s not a quick fix. With a large volume of inaccurate records, you’re into a long slog that could’ve been avoided by involving the right teams from the start.

Why talking to someone who’s done this matters

The difference between a successful project and a rescue mission usually comes down to whether someone in the room knows where the skeletons are buried.

Not theoretical skeletons. The actual ones. The integration that looked fine in the demo but breaks under volume. The data migration that passes validation but lands finance with a reconciliation nightmare. The cutover plan that works on paper but collapses when your warehouse team is still picking orders.

This is what Commerce Thinking’s ERP practice exists for. We’re former brand-side operators from companies like Gymshark, Liverpool FC, and Burberry who’ve run these projects, lived through the mistakes, and know what good looks like when the pressure’s on.

We’re not here to sell you more config or extend timelines. We’re here to tell you what’s actually going to break before it does, challenge the bits that don’t make sense, and make sure your team can own the system when we leave.

Because the goal isn’t dependency. It’s capability. If you need us six months after go-live to run a routine process, we’ve failed.

If you’re mid-project and starting to feel that unease, or you’re about to kick off and want to avoid the traps, talk to us. We know what a good project looks like because we’ve built them. And we know what a struggling project looks like because we’ve rescued them.

If you’ve got that sinking feeling

ERP doesn’t solve problems. People do.

If you’re mid-project and you’ve got that feeling, treat it as a signal. You’ve still got time to fix the fundamentals, make room, clean the data, and force end-to-end ownership before the system becomes a very expensive spreadsheet generator.

You don’t need deeper config. You need broader context, the kind that understands how your teams actually work, not how the process map says they should. If you don’t bake that context into design and testing, you’ll go live and still be arguing about why merchandising missed PO lock dates or why a landed cost update has just blown up reporting deadlines.

None of this is about perfection. It’s about reality.

That’s a different kind of problem. And it’s the one our latest research puts a sharp point on.

15 of 24 brands we spoke to use a dedicated cross-border platform. Of those 15, only three have ever migrated to a different provider. The frustrations are near-universal: cost opacity, rigid workflows, slow iteration, roadmaps that feel disconnected from operational reality. The satisfaction scores reflect it too, with Global-e sitting at 5.6 out of 10 in our research.

And yet the switching rate is remarkably low.

The reason isn’t loyalty. It’s architecture. Cross-border platforms sit deep in the stack. They touch checkout, duties calculation, currency conversion, fraud, logistics, and returns. Ripping one out means re-integrating all of it, often during the same peak periods when the pain is most acute. The risk feels higher than the reward, so brands tolerate. They work around. They accept margin leakage they can’t fully quantify because the alternative feels worse.

One operator captured the sentiment clearly: “They are the market leader, thus regarded as best in class, but we do not regard them as a pragmatic partner.”

That’s a telling quote. It points to a gap between market position and actual operational value that is becoming harder to ignore as international revenue becomes more material and finance teams start asking market-level questions the tooling can’t answer cleanly.

The parallel with returns is striking. Two or three years ago, returns looked exactly like this. High dissatisfaction, low movement. Operators knew something was wrong but couldn’t justify the migration cost. Then platforms like Loop, Swap, and Reveni made it easier to move, and the market shifted. Switching behaviour that looked structural turned out to be situational.

Cross-border hasn’t had that moment yet. But the conditions for it are building.

Macroeconomic pressure is forcing greater scrutiny of international margin. Currency volatility, rising logistics costs, and tighter cash positions are making previously tolerated leakage harder to ignore. At the same time, operating models are shifting. Brands that once ran a single global storefront are experimenting with multi-store setups, regional entities, and more complex Shopify Markets configurations. Those changes surface limitations in cross-border platforms that were designed around simpler assumptions.

The question for 2026 is whether cross-border follows the same arc as returns, or whether the integration depth keeps brands locked in longer. We don’t have a definitive answer. But the brands that are already stress-testing their cross-border setup, mapping their actual margin by market, and building internal clarity on what switching would involve, are going to be better positioned when that moment arrives.

Our full research report covers the cross-border landscape in detail, including adoption patterns, satisfaction scores, and the specific friction points operators flagged most consistently.

Read the full report

In this series, we’ve been following West Park Rovers, the fictional club from Kieran Maguire’s The Price of Football, as they bring retail in-house.

So far, we’ve covered the business case for making the move, the operational reality of running what is effectively a small factory, and in the last piece, The 12-Month Apocalypse, the brutal economics of a catalogue that expires every year.

That piece was about pressure with a fixed horizon. Time you can see running out.

This one is about the opposite.

Moments that arrive without warning. A goal. A farewell. A signing that shifts the mood of a whole city overnight. Demand that exists right now and will be gone by Thursday.

Seasonal retail has a calendar. Sports retail has a pulse.

Catch up on this series:

• Part I - West Park Rovers

• Part II - The Business Case

• Part III - Your Club Shop is a Factory

• Part IV - The 12 Month Apocalypse

Most retail businesses plan in seasons.

Spring drop. Summer sale. Autumn launch. Clearance.

Sports clubs plan in interruptions.

A late winner. A derby goal. A record broken. A captain leaving. A shock signing. A manager sacked on a Tuesday.

These moments do not arrive politely. They do not wait for a range review. They do not care what your production lead time is.

Fans react instantly. Your shop usually doesn’t.

That gap is where the money leaks. It is also where the relationship weakens.

Because in sport, the moment is the product.

Subscribe now

The problem with “seasonal” thinking

When retail is outsourced, the calendar wins.

A licensing partner plans ranges months in advance. Products get approved, produced, shipped, and sold. Everything is built for predictability.

That’s fine for the fundamentals. New kits. Training ranges. Evergreen accessories.

It falls apart the second something unexpected happens.

A goal that goes viral.
A farewell that turns emotional.
A promotion that changes the mood of a whole city.

Fans want to mark it now. They want something that says “I was there” while the story still feels alive.

Outsourced retail rarely moves at that speed. It moves at the speed of contracts, approvals, and shared priorities.

By the time it reacts, the moment has already moved on. The emotion has cooled. Fans have spent elsewhere. Or they have bought something unofficial.

In-house doesn’t just mean “make new products faster”

This is where most clubs get the wrong idea.

They hear “moments” and jump straight to “drops”.

Limited edition tee. Special graphic. Quick turnaround print run.

Sometimes that is the right move. Often it isn’t.

Because the most valuable part of moments-based merchandising is not always new product creation.

It’s attention control.

If a player has a standout performance, you do not always need a new item. You need to make it easier for fans to buy what already exists.

That means:

• Moving relevant products higher up collection pages

• Updating homepage modules while traffic is spiking

• Changing default recommendations to match what fans are searching for

• Making search forgiving when names are hard to spell

• Ensuring the “obvious” products are not buried behind generic navigation

This is not hype. It’s friction removal.

You are not inventing demand. You are stepping out of its way.

That is why bringing retail in-house changes everything. It gives you the ability to re-merchandise in hours, not weeks.

Not loudly. Not desperately. Just intelligently.

Why moments matter more than campaigns

West Park Rovers start by looking at what actually drives spikes in their store traffic.

They expect it to be planned launches. It isn’t.

The biggest surges come from unpredictable events.

A big win.
A controversial result.
A player suddenly becoming a hero.
A goodbye nobody expected to hurt as much as it did.

The club’s current retail setup can’t capture those surges properly because the store isn’t designed to respond to the club’s reality. It’s designed to sell a pre-agreed range efficiently.

That is the difference.

A club does not need a better calendar. It needs an engine that can turn emotion into action.

The line between celebration and cash grab

This is the delicate part.

Moments merchandising can feel cynical if it’s handled badly.

Fans can smell opportunism. They know when something is a cheap grab. They know when a moment is being squeezed for revenue.

But that is not a reason to ignore moments altogether. It’s a reason to build a better approach.

The best moment-based merchandising does three things:

1. It feels like it comes from the club
   The tone, design, and framing matter. Fans are buying belonging, not fabric.

2. It respects timing
   If you show up too late, it feels performative. If you show up too aggressively, it feels exploitative.

3. It offers something real
   That could be a limited item, a numbered run, a commemorative piece, or simply a more thoughtful presentation of existing products.

The point is not to monetise emotion. The point is to give fans a way to participate in it.

If the club does not provide that, someone else will.

The operational truth behind the romance

This is not a creative challenge. It’s an operational one.

If you want to respond to moments, your operations need to be built for volatility.

That means:

• Stock strategy that supports late-stage customisation

• Clear processes for prioritising what gets promoted and when

• Fast changes to storefront merchandising without breaking everything

• Systems that can handle sudden spikes without collapsing

• A decision-making model that doesn’t require six sign-offs

Moments do not wait for internal alignment.

West Park Rovers realise that “being agile” is not a marketing aspiration. It is a capability. And capabilities are built, not wished into existence.

You either have the muscle to move quickly, or you don’t.

Subscribe now

Why outsourcing struggles here

It’s not because third parties are incompetent.

It’s because they are optimised for a different game.

A licensing partner is built to deliver a stable operation at scale. They plan ranges, manage inventory, fulfil orders, and protect margin through predictability.

They are rarely built to:

• Drop everything for one club’s unexpected moment

• Take creative risks on short-run products

• Re-merchandise a site in line with match-day emotion

• Deal with chaos without renegotiation

Their job is to run a retail business efficiently.

Your job is to run a club that generates constant, unpredictable stories.

When the store isn’t connected to the story, it becomes generic. Fans feel it. And they shop accordingly.

Treat moments as signals

The real shift for West Park Rovers is how they think about their retail calendar.

They stop treating it as fixed.

Instead, they plan the season like a framework with space for signals.

Signals like:

• A player spike in popularity

• A sudden increase in searches for a specific name or item

• A viral clip driving traffic to the store

• A surge in demand for a certain product line after a match

The club does not need to guess perfectly.

It needs to be able to notice, respond, and capitalise without panic.

That means building the store to be rebalanced continuously, not refreshed occasionally.

The short version

Sports retail is not seasonal retail.

It’s reactive retail.

The most valuable moments are unpredictable. That is exactly why they are valuable.

Clubs that run commerce in-house don’t win because they can print a t-shirt quickly. They win because they can move attention quickly. They can make the obvious products easier to find. They can respond to demand while it still exists. They can show up in the moment without making it feel like a scheme.

Fans will always buy around moments.

The only question is whether they buy from you.

That’s the data point our latest research leads with in its chapter on switching behaviour, and it’s worth sitting with for a moment. This isn’t normal SaaS churn. Most software categories don’t see this level of movement at this stage of market maturity. What it tells you is that the returns platform market is still finding its shape, and that the fit between platform and operator breaks more often than anyone building the original business case probably expected.

The more interesting question, though, is why.

When we dug into the reasons behind platform switches, it wasn’t feature gaps or platform failures doing the damage. It was context shifts. The business changed shape, and the platform didn’t change with it.

New markets were the most common trigger. A platform that worked cleanly for UK-only operations doesn’t necessarily scale to EU or US without significant friction. Additional warehouses created similar pressure: routing logic that made sense when inventory lived in one location starts to break when it’s spread across multiple sites. Finance scrutiny drove switches too, particularly when leadership started asking margin questions the current setup couldn’t answer. And volume growth exposed the difference between a platform that works at 5,000 returns a month and one that holds up at 25,000.

There’s an implicit promise in most SaaS relationships: implement once, optimise occasionally, move on. Returns doesn’t work that way. Every new market, warehouse, or carrier relationship introduces friction. Every peak season stress-tests the setup. Every finance review surfaces questions the platform wasn’t designed to answer.

Operators aren’t disloyal. They’re pragmatic. When the fit breaks, they move.

What’s also interesting is the direction of movement. Most switching isn’t happening between modern platforms. It’s happening away from older, more rigid setups. And once that first move is made, appetite for further switching drops significantly. Platforms don’t need to be perfect. They just need to feel directionally right.

This changes how competitive pressure shows up in the market. Modern platforms aren’t winning by pulling customers from each other. They’re winning by being the obvious next step once legacy setups break. That makes switching events rarer but more decisive. When a brand is ready to move, they usually know where they’re going.

For brands, the practical implication is straightforward: budget for re-evaluation. Returns infrastructure isn’t a one-time decision. It’s something that needs to be revisited as the business changes shape, not just when something breaks.

The brands that treat their returns setup as a living part of the operation, rather than a problem that’s been solved, tend to be the ones that don’t end up in crisis mode when volume spikes or a new market goes live.

Our full research report covers this in more detail, including the specific triggers that tend to drive migration decisions and what the switching patterns tell us about where this market is heading.

Read the full research

That changes now. And AI gives us the perfect reason to start.

The pace of change hitting back-office functions right now is unlike anything most brands are prepared for. The operators who figure it out early, together, in the same room, without a sales pitch in sight, are going to be miles ahead. That’s what Operator Experience is built for. And OX1 Manchester on 16th April is where it begins.

What’s Happening

12:30 – 17:00 | Hallé St Peters, Manchester M4 6BF

The afternoon is structured around a single, practical question: what if your brand challenged itself to automate 30% of all operational tasks by 2030?

Working in small squads, you’ll map end-to-end processes, hunt for automation opportunities, and identify the highest-value targets, prioritising quick wins you can actually take back to the office on Monday. You’ll leave with a concrete initiative in hand, not a deck to file away.

The session opens with AI Essentials and How to Use It Safely, covering real-world use cases across product, ecom, finance and ops, alongside risk, governance, and compliance guidance you can apply immediately. Special guests will share AI-era org structures and practical safety frameworks.

What you need to bring: a charged laptop. We’ve got everything else.

17:00 – Late | The Castle, 66 Oldham St, Manchester M4 1LE

After the workshop, the evening continues at The Castle, a proper pub, a proper crowd. Strictly brand-side operators. No selling, no pitching. Just the kind of honest conversations you rarely get to have when everyone else in the room has a commercial agenda.

And fashion photographer Nathan McDowell will be capturing the day, so you’ll get professional shots afterwards.

You’re in Good Company

The brands already confirmed are ones you’d know; fast-growing, operationally serious, and brand-side through and through. But more than the names, it’s the people in the room that matter. Directors and heads of function dealing with the same challenges you are. No hierarchy, no agenda. Just people who get it.

Who Is This For?

If you’re a director, head, or senior leader in:

• Operations or supply chain

• Finance and back-office technology

• Ecommerce

• Merchandising or buying

• Product data or tech

...and you work brand-side at a fast-growing fashion, apparel, or consumer brand, this is for you.

The agenda has been built by the community itself: Lauren, Robbie, Ewan, Robin, Roberta, Kya, Paul and Luke. People who’ve sat in your seat and know exactly what’s useful and what’s a waste of time.

Made Possible By Our Partners

OX1 Manchester is free to attend for brand-side leaders and operators. That’s only possible because of our partners, who are here to support the community, not sell to it.

Headline: Dema
Event: Emfas, Loop, Torque
Community: Whanau, Vervaunt, Commerce Thinking, Centra, High Cohesion, Resourced

They’ve committed to making this a no-promotion, no-posturing environment. If you’re here, it’s because someone thought you’d fit.

How to Join

Places are limited. This is a curated event, size is deliberately kept small so the conversations are real and the workshop actually works.

To register or express interest in joining the Operator Experience community:

operatorexperience.ai

If you want to know more about what Operator Experience is building, a peer community for brand-side operators, events across the UK, and a space where the people doing the work can actually learn from each other, that’s the place to start.

Save the date: 16th April 2026. Manchester. See you there.

So why does the returns process still feel, for so many teams, like a constant source of pain?

That’s the question our returns and cross-border research puts at the centre of its second chapter. And the answer is one of the more useful distinctions in the whole report: the gap between tool satisfaction and operational confidence.

Operators like their platforms. They do not necessarily trust their returns operation.

That gap, between “I like my returns platform” and “I trust my returns process end to end,” is where margin leaks. It’s where manual workarounds live. It’s where the spreadsheet that someone built eighteen months ago becomes quietly load-bearing infrastructure.

The pain, when you dig into it, isn’t in the platform itself. It’s in everything the platform has to connect to.

Integrations with WMS, ERP, and OMS systems that half-work or require constant maintenance. International routing logic that doesn’t flex cleanly across markets, carriers, and cost structures. Reporting that’s fragmented across three systems and reconciles in none of them.

As one operator put it during our research: “The returns portal is great. The returns process is a nightmare.”

That sentence captures something important. It tells you that the product problem in returns has largely been solved. The platforms are good. The interfaces are clean. The customer-facing experience has improved significantly. What hasn’t improved at the same rate is the operational infrastructure those platforms sit inside.

Our research surfaces a consistent set of operational failures that show up not as edge cases but as default behaviour across brands. Returns being routed to the wrong warehouse because the logic doesn’t reflect actual cost or inventory. Duties being paid twice, or not reclaimed at all, on international returns. Exchanges requiring manual intervention despite available stock. Customer service teams overriding policies to unblock customers, creating inconsistency and margin leakage downstream. Finance teams reconciling returns and refunds weeks later, outside the core systems.

None of those are platform failures in isolation. They’re symptoms of systems that weren’t built to work together at this level of complexity.

The other area the research flags is recovered revenue. Almost every brand we spoke to is capturing some value through exchanges and store credit. Very few feel confident they’re maximising it. The mechanics exist in most modern platforms. The strategic clarity doesn’t. Teams aren’t sure what good looks like, how much they’re actually retaining, or where the biggest opportunities sit.

What this means practically is that if you’re evaluating your returns setup and the platform scores well, you’re asking the right question in the wrong place. The platform isn’t the bottleneck. The bottleneck is the system around it: the integrations, the routing logic, the reporting, and the manual workarounds that have quietly become part of how the business operates.

Good tools don’t fix broken systems. That’s the takeaway from this part of the research, and it’s a useful frame for any team that’s wondering why their returns operation still feels harder than it should.

Read the full research

That’s the thing that keeps coming up when you talk to mid-market ecommerce operators about international revenue. There was no board-level decision, no market entry strategy, no dedicated team assembled to manage cross-border complexity. There was a product that went viral in Germany. A TikTok that landed in the US. A marketplace that opened up in France. And then, somewhere between the excitement of the numbers and the reality of the fulfilment queue, 40% of orders were shipping overseas.

The growth was real. The infrastructure to support it was not.

In our latest research, 71% of brands reported that 30% or more of their revenue comes from international sales. At the sharper end, six brands sat between 70% and 90% international. These aren’t global enterprises with regional teams and dedicated localisation budgets. They’re mid-market operators, often with lean teams, domestic systems, and workflows that were designed for a single market.

The phrase that keeps coming back is this: technically global, operationally domestic.

It’s a useful way to think about where a lot of brands currently sit. The revenue is international. The returns logic, the customer service setup, the finance reconciliation, the carrier relationships, the duties handling: mostly not. And for a while, that works. Orders go out. Customers receive them. Revenue goes up.

Then the cracks appear.

Returns from international customers become more expensive and harder to route. Cross-border costs start eating into margin before anyone is measuring them properly. Finance teams begin asking questions that operations teams can’t answer. And suddenly the gap between revenue growth and operational capability becomes impossible to ignore.

The important thing to understand is that this isn’t a failure of ambition or planning. It’s a structural pattern. Paid social doesn’t respect borders. Organic demand doesn’t wait for you to build a cross-border strategy before it shows up. The platforms that brands use to grow, Shopify, Meta, TikTok, don’t have a “domestic only” setting.

What this creates is a very specific kind of operational challenge. The brand looks international from the outside. The infrastructure, if you dig into it, is still built around assumptions that made sense when 90% of customers were in the same country.

The first step toward fixing it is recognising it. And for many brands, that recognition is still catching up with the reality of their revenue mix.

We spoke to 24 fast-growing ecommerce brands for this research, primarily in fashion and consumer goods, with revenues ranging from approximately £10m to £100m. The full report is worth reading if you’re running or advising a brand at this stage of growth. Not because it has all the answers, but because it makes visible a set of problems that most teams are quietly navigating without a shared language for them.

If your international revenue is growing faster than your operational capability to support it, you’re not alone. You’re the norm.

Read the full research

If you’re a parent, you know what that means. The same episode of your child’s favourite show. On repeat. Over and over again.

In our house it’s Bluey, The Sign. I’ve lost count how many times we’ve watched this 30 mins episode.

In global commerce, it’s tariffs. And the three year old is the US President.

Trump is not considered a serious person by most of the world. But the consequences of his actions are.

Another threat. Another headline. Now a proposed 10% global tariff.

Rather than write another long explainer, here’s the short version.

If you’re an overseas brand selling into the US, these are the decisions you should make now. And stop procrastinating on.

Subscribe now

1. Model the Numbers Properly

Not high-level margin. Not blended averages.

Build a SKU-level landed cost model under both a 10% and 25% scenario. Know your true duty exposure, margin by channel, break-even thresholds, and cash impact.

If you haven’t modelled it, you’re operating on hope.

2. Import at Cost, Not Retail

If you’re shipping DTC cross-border and importing at retail value, you are voluntarily inflating your duty base.

Fix your transfer pricing, importer of record setup, and merchant of record structure.

3. Move Inventory into the US (If Volume Justifies It)

Stop treating the US like just another international market.

Bulk import at cost. Hold inventory in a US 3PL. Ship domestically.

Lower duty base, faster delivery, lower friction, more control. This isn’t just tariff mitigation. It’s operational maturity.

Subscribe now

4. Clean Your Product Data

Tariffs expose weak foundations.

You need accurate HS/HTS codes, clear country of origin, a clean item master, and SKU-level cost integrity.

If your ERP can’t tell you margin by SKU by market with confidence, that’s a real problem.

5. Build Pricing Agility

Decide now where you absorb, where you pass on, and what your threshold triggers are. Static global pricing is fragile in a volatile political environment.

6. Decide What the US Actually Is to You

Is it a strategic growth market, a profitable channel, or just incremental revenue?

If it matters, operationalise it properly. If it doesn’t, stop pretending.

7. Stop Delaying

Every time tariffs are mentioned, brands say the same thing: let’s wait and see.

We’ve been waiting and seeing for years.

Subscribe now

The bottom line

You can’t control Trump’s next outburst. He has the temperament of a three-year old.

You can control your landed cost model, your operating structure, your pricing logic, and your data integrity.

Stop kidding yourself.

Just like Bluey will play on my TV, this tariff episode will play again. So stop procrastinating on big decisions, just get on with it.

Subscribe now

• mostly fashion

• mostly UK-headquartered

• mostly between £10m and £100m

• and asked questions about returns, cross-border, margin visibility

• and the gap between what tools promise and what teams actually experience

We expected to find brands at various stages of deliberate international expansion. What we actually found: international revenue has already arrived, and for most it wasn’t a carefully thought out strategy.

67% of the cohort reported 30%+ of revenue from international sales. Six brands sat between 70% and 90%+. These aren’t global enterprises with regional teams. They’re operators with domestic systems and workflows scaling beyond a single market.

The growth is real. The margin pressure that follows is also real.

What the report covers

The full report goes deep across six areas. Here’s enough to frame the conversation, not enough to replace reading it.

Returns tooling has improved dramatically. Returns operations haven’t. The gap between “I like my platform” and “I trust my setup” is where the margin leaks live.

16 of 24 brands have switched returns platforms at least once, and the triggers are rarely about the tool itself.

Cross-border remains the problem everyone tolerates. Universal frustration, almost zero switching. The report explores whether that’s about to change.

And underneath all of it: the same integration pain, manual workarounds, and analytics gaps showing up in every conversation regardless of category or vendor.

Why now

Tighter margins, rising logistics costs, and sharper finance scrutiny are making previously tolerated leakage harder to ignore. The window where international complexity could be treated as a future problem has closed.

The report maps where that pressure is showing up, and what operators are actually doing about it.

Read the full report

Independent research. No vendor sponsorship. First in an ongoing series. If you’d like to participate or just tell us we’ve got something wrong: research@commercethinking.com

But what starts as a physical extension of the brand quickly becomes a merchandising, ops and stock management headache. And that’s the part nobody warns you about in a POS demo.

Here’s what digitally-native teams miss: a store isn’t a vibe. It’s inventory, data, people, and a stack of decisions that compound daily.

And the compounding is the killer.

Every physical location adds complexity to your operation. More SKUs. More touchpoints. More ways for stock to end up in the wrong place at the wrong time.

We’ve delivered these retail expansions in brand, from merchandise planning through to setting up the stock room and everything that breaks in between.

It’s never easy. But it is predictable.

Inventory stops being a number and becomes geography

In the main, ecommerce has the luxury of tidy pools of stock. Retail is different. The moment stock leaves the DC, it becomes harder to control.

Once you’ve pushed units into a store, you’ve turned them into a local bet. If you guessed wrong on depth, size curve, or timing, you can’t just fix it with a Shopify tweak.

This is why stores create strange new ways for it to break. Stock can sit in the wrong location simply because it isn’t that store’s pick day.

You start learning that allocation is never just allocation. It’s also delivery schedules, labour in the DC, and how fast you can move product without breaking the maths.

The trap for DTC brands is treating store stock like marketing inventory. Something you place beautifully and then let it do its job.

In practice, stock decisions are as strategic as store design, because the moment you go physical your stock stops being numbers in a system. It becomes geography, weather, timing, and people.

If you want a simple operator test, go look at how often you can rebalance stock today. Not theoretically, actually. If the answer is weekly or rarely, you’re about to feel it.

Omnichannel can act like a release valve, but only if you build it properly. The whole point is making stock fluid again, shipping from stores, fulfilling locally, and moving it at the pace of demand, not delivery schedules.

That’s also where some DTC teams get annoyed. They wanted a store. They accidentally signed up for a supply chain node.

Subscribe now

The boring data stuff will stop your store trading

Most DTC teams can tolerate messy product data for longer than they should. A dodgy size value here, a duplicate SKU there, a naming convention that only one person understands. Online, you can often ship anyway.

In a store, those little inconsistencies turn into hard stops. This can happen when the size format in the warehouse system doesn’t match the labels on the physical products. A shipment lands but can’t be received.

This is why store number two hits like a brick. The jump from one to two physical locations is less of a step and more of a chasm, like going from one child to two.

At one location, problems get fixed in real time. At two, they scale.

Brands end up thinking they need to switch POS or layer on another tool. Sometimes they do. But the harder truth is usually integration and someone owning it.

You need everything feeding into one view of stock everyone trusts, rather than stacking system on top of system. And you need to treat product data like an asset, not an afterthought.

If you’re firefighting at store two, store three is going to tip members of your team into madness. That line’s dramatic, but it’s also accurate.

The numbers start lying and you pay for it in working capital

Multi-location retail doesn’t just add admin. It messes with your confidence in the numbers.

One location can be messy and still survivable. Add more nodes and suddenly you’re dealing with timing gaps, transfers, delayed receipts, and teams making decisions off different versions of the truth.

That’s when you see the quiet behaviour change. People start over-ordering, holding extra, and pushing problems into next month because nobody trusts what’s on the screen.

Finance and merchandising end up talking past each other. Buying can’t buy confidently if stock, inbound, returns, and sales aren’t landing in one view quickly enough.

Meanwhile the store team sees it the simple way. Customers want the thing, it’s not there, and everyone looks at everyone else.

Subscribe now

A store is also a people business and the first year isn’t glamorous

DTC brands are used to scaling with code, media, and a few more warehouse pickers. Stores scale with rotas, training, and managers who can run a tight ship on a Tuesday morning, not just launch parties on a Saturday.

This is where a lot of digitally-led teams get caught out. You can’t A/B test your way to consistent store operations if you haven’t built proper training and clear ways of working.

If you’re going to do this properly, you need staff who understand operations, not just merchandise. You need processes that connect physical and digital inventory, and data that can be trusted by finance, fulfilment, and commercial teams alike.

You also need patience. The first year isn’t glamorous. It’s systems work. It’s process work.

This is the bit that frustrates a lot of founders. They wanted a store to make the brand feel more real. Instead, they get a new category of daily work that doesn’t show up on Instagram.

Physical retail has physical failure modes

The biggest one is brutally simple. The cost base doesn’t care how good your brand deck looks.

Online, you can often dial spend up and down with demand. In stores, rent turns up every month, payroll turns up every week, and the inventory you bought to make the shop look full is now sat somewhere specific.

That’s why so many DTC store rollouts look fine in the first few openings and then start to get ugly. When the economics don’t hold, you don’t just turn it off; unwinding stores can be expensive and public.

So what do far too many DTC brands overlook? It’s not the shopfit. It’s the operating model.

They underestimate how quickly inventory becomes a local replenishment problem. They underestimate how frequently messy product data will block trading. And they assume that if one store works, two stores will work, when two stores is the point where the cracks widen.

If you’re going into stores, treat it like you’re changing the shape of the business. Because you are. Build the unsexy parts early, especially data discipline and integration, and you’ll give yourself a much better chance to make the fun bits pay off.

Subscribe now

Happy Monday everyone. Why not stick on some Joy Division and settle in for an uplifting article…

The efficiency case for AI in brands isn't debatable anymore. Automate repetitive work, reduce headcount, protect margin. The tools work, the ROI is clear, and your competitors aren't waiting for permission.

So brands will do this. They’re already doing it.

What nobody’s pricing in is the leadership deficit we’re sleepwalking into.

The tasks AI automates best sit disproportionately in entry-level roles. Weekly trade packs, routine reporting, first-pass copy. The structured, repeatable work that happens to be where people learn how brands actually operate.

Automate those tasks away and you don’t just trim payroll. You collapse the apprenticeship. And without the apprenticeship, you don’t get the next generation of leaders who know what good looks like and can actually scrutinise and manage the machines.

Subscribe now

The pipeline breaks quietly

The hypothesis is simple.

If the number of entry-level roles shrink. Hiring slows. The juniors you do bring in spend their time monitoring AI outputs instead of doing the actual work. They learn to approve, not to judge. To trust the model because they’ve never lived through what happens when the model is subtly, expensively wrong.

Give it 10 to 15 years and today’s assistants should be your managers. Today’s coordinators should be your directors. Instead you’re staring at a gap.

A shortage of people who can interrogate what the automation is doing. Who’ve built the pattern recognition and muscle memory that only comes from years of reps. Who know how the business works when the systems don’t agree and the customer does something annoying right on the cutoff.

That knowledge doesn’t come from courses or observing. It’s earned through doing the work, making mistakes, and gradually building an “in the bones” understanding.

With AI, we’re doing away with the opportunities to develop knowledge in the next generation of brand leaders.

Subscribe now

The competitive failure mode

Here’s where it gets worse. Every brand faces the same pressure to automate. So every brand will. And every brand will end up competing for the same tiny pool of people who actually came up doing the work.

Costs spike. Churn accelerates. You lean harder on consultants and external hires because the institutional knowledge isn’t sitting inside the business anymore.

Worse, you get slower and more fragile. You’ve got excellent tools and not enough people who can tell you when the tools are lying. When the forecast is off, the copy doesn’t sound like your brand, or the operational decision will create problems three steps downstream.

The more you automate, the more you need humans with deep context to scrutinise what you automated. But those humans don’t magically appear. They’re grown through years of doing the work you just removed from the business.

We’re trading short-term efficiency for long-term fragility, and we’re all doing it at the same time.

We’re not prepared for this

The brutal part is that brands will make the rational choice at every stage. Automate or fall behind. The market punishes hesitation faster than it punishes deferred talent problems.

But rational individual decisions can still create collective problems. And we’re sleepwalking into one.

In 10 to 15 years, we’ll have a generation of “leaders” who’ve never done the work they’re leading. Who can’t sense-check AI outputs because they never built the judgement. Who’ve only ever approved, never created.

Now the optimists will say “well the AI will be so good in 10 years that it’ll paper over the gaps in individual understanding”. Totally fair, but doesn’t that feel pretty bleak for the next generation of school, college or university leavers? Is that really the extent of our ambition?

Without new pathways and different approaches, we’ll break the talent market.

The leadership deficit is coming. Not because anyone intended it, but because nobody stopped to ask what happens when an entire industry automates away its training ground at the same time.

If we don’t intervene now - if brands don’t deliberately rebuild pathways for early-stage talent, create new apprenticeships, make space for real practice and feedback - we’ll get the efficiency today and the crisis tomorrow.

Our small attempt at not just being part of the problem

AI is moving too fast to navigate all the implications alone, and there’s nowhere dedicated to retail brands solving the adaptation problem together.

Whether that’s practical use of the tools, to managing the org design shifts it’s prompting.

So I’ve created an invite-only community where brand leaders and operators learn, share and build AI-era capabilities as a peer group.

We're deliberately mixing senior leaders with ambitious junior talent. Systems admins with merch managers, ops specialists with analysts. Different experience levels, same challenge: learning to work with AI without losing the fundamentals that made us good at this in the first place.

If you’re interested then apply at Operator Experience.

Between November and December 2025, I interviewed 40+ leaders and operators at consumer brands. I spoke with people at different levels within brands turning over £10m to £250m.

I asked really straightforward questions about AI: their attitudes, confidence levels, practical usage, and what pressure they’re feeling.

One trend dominated each conversation.

Nearly everyone, from CEOs to early-stage operators, was certain AI would reshape how they work day-to-day and how their business operates. But they didn’t know where to start. Most didn’t trust the noise they heard online. And most were trying to figure it out alone.

The pressure is real

There’s another layer to this. Nearly every operator I spoke to admitted feeling insecure about their competition moving faster on AI. The fear isn’t abstract.

It’s that a rival brand is already automating processes, improving margins, or accelerating decision-making while they’re still figuring out where to begin.

That stress compounds when owners, investors, and board members keep (unhelpfully) asking “what’s your AI strategy?”

It’s a question that sounds reasonable but often creates paralysis. Most operators don’t need an AI strategy, they need to identify specific tasks worth automating and start building capability around those. But when you’re being asked for a strategy deck while trying to run the business, the gap between expectation and reality gets uncomfortable fast.

This is exactly why isolation doesn’t work. When you learn with peers who are facing the same questions and the same pressure, you realise you’re not behind, just figuring it out like everyone else. And that shared context accelerates everything.

The L&D gap

AI literacy was a top priority for every People and HR leader I spoke with.

Then I asked the follow-up: what programmes exist to solve this? What are you actually doing about it?

The answer was almost always the same. Not much. A few vendor webinars. Maybe some online courses that nobody finishes. But nothing credible. Nothing structured. Nothing that actually builds capability across the business.

There’s a clear gap between what brands know they need and what’s available to them. And it’s not for lack of trying.

I think we need to approach AI literacy differently. Traditional L&D (courses, one-size-fits-all training) isn’t fit for purpose. AI is moving too fast. The day someone creates a definitive course on leveraging AI inside a fast-growth brand is the day it’s out of date. The technology shifts rapidly, and the applications are nearly infinite.

So the questions I keep asking myself are: How do we help operators understand where automation creates leverage in their specific role? And how do they learn to apply it alongside peers facing similar problems?

Right now, that doesn’t really exist at scale.

What we’re creating

Operator Experience is a small, invite-only community for people running complex parts of modern businesses: operations, finance, commercial, merch, product, tech, logistics, growth.

It’s launching publicly this month with the support of a group of brilliant partners. The group will meet four times per year. Our first workshop is in Manchester on 16th April, with London following in July and more cities through 2026 and beyond.

We’re intentionally keeping this tight-knit, capping the number of brands but encouraging each to bring multiple people from across their business. The biggest winners in the AI era will be brands that organise cross-functionally to automate and augment work. We’re practising what we preach.

If this sounds relevant to you or your team, get in touch with me directly on luke@commercethinking.com

More at operatorexperience.ai