I joined H&M in logistics over three decades ago, when the company had just opened its fifth country. By the time I retired, 35 years later, the turnover had doubled every five years. I spent time as a country manager for garment production in Bangladesh and Romania, seven years living in Hong Kong as head of global inbound logistics, and then back in Sweden overseeing the same function for another decade. In the end, we were managing shipments across more than 400 trade lanes into warehouses in over 70 countries, across every mode of transport.

Throughout that growth, a core challenge persisted. The inbound shipping data existed, but we could not rely on it.

The real problem was not visibility

For a long time, many people in the industry spoke about visibility as the end goal. What I’ve learned throughout my career is that visibility without accuracy is not very useful. We had visibility tools and data feeds. Carriers were sharing container positions and estimated arrival times. But the quality of that information was often poor.

I saw this play out in numerous ways. A container was tracked accurately across an entire ocean voyage, but when it entered customs on a Friday afternoon, nothing was updated over the weekend. Or it was transferred to a feeder vessel from a major hub to a smaller port, and that leg of the journey was not recorded. When this happens, a warehouse team, trying to plan staffing levels, may stop trusting the data they are receiving. Once that trust is lost, it is very difficult to recover.

At H&M, the consequences were direct. The store team plans six months ahead which products should be on the floor and when. For a two-week campaign window, everything needs to arrive on time. If a collection of blue garments arrives a week or two late, the store has already moved into the next campaign. The yellow range is on the floor. The blue goods have nowhere to go at full price, so they get marked down, and that margin loss happens on every piece.

The warehouse team at H&M then faces a different version of the same problem. A typical distribution center carries around 60 to 70 percent of its workforce as permanent staff and brings in the remaining 30 to 40 percent on a flexible basis, based on what is inbound. When the data says a large volume of goods is arriving, the warehouse calls in those additional workers, but if the shipment is delayed or the figures were simply wrong, those workers stand idle. That costs money. And when it happens repeatedly, the warehouse stops trusting the data entirely and starts making its own judgments, which creates a different set of issues.

You Might Also Like: How to Improve Schedule Reliability in Ocean Freight.

Experience fills the gap (but doesn’t solve the core problem)

For a long time, the industry managed this through institutional knowledge. Experienced logistics staff could read the signals, anticipate problems, and flag delays before they became serious. But this approach did not scale, and it created its own problems.

The logistics team knew, more or less, when ships were arriving. But that knowledge did not always reach the people who needed it. The warehouse planned around one set of figures, while the buying team worked from another. Each department operated with a slightly different version of reality, and when something went wrong, the feedback rarely made it back to the source. Instead, it created silos, where we would only hear about a problem if it was serious enough to escalate. Smaller inaccuracies, the ones that accumulated into real costs over time, went unaddressed.

There was no single thread of information running from the production country to the store shelf. Each handover point was a potential break in the chain. Without a shared, trusted picture of where goods actually were and when they would arrive, it was very difficult to act as one organization.

My view has always been to make it simple. If a system is too complicated to trust, people will work around it. The answer is accurate, trustworthy data that everyone in the organization can plan against.

Related: The Strategic Buying Guide to Logistics Visibility Software in 2026.

What the logistics industry needs to get right

Many companies today are moving quickly toward supply chain orchestration and autonomous execution. The potential is significant, but I have seen firsthand what happens when you build complexity on top of an unreliable data foundation. You get automation that produces wrong answers, but more quickly.

The path to success is to get visibility and the data quality right and earn the trust of the people who depend on that information. Everything else builds on that foundation. What that requires, in practice, is not just more tracking feeds. It needs a way of evaluating which data to trust at each point in the journey to give the whole organization a source of truth they can actually plan against.

The technology is now catching up to that problem. For those of us who managed global logistics before these tools were available, that is a significant development. What I believe the industry is moving toward is a continuous thread of information from the production country all the way to the end customer. This would mean not just knowing where a container is, but knowing what action is needed next at every point in the chain.

For smaller companies, that capability is not optional. They cannot afford the internal departments that larger organizations rely on to absorb the uncertainty. But larger organizations will need it too, because at scale, the cost of poor decisions only compounds.

‍

Mats Samuelsson is a strategic advisor at Moddule and a former Executive Senior Vice President of Logistics at H&M, where he spent 35 years managing supply chain operations across 70+ countries.