Looking towards systems that are ‘good at what they do’

Listen to our podcast here (Dutch):

Wim Farasyn, who is synonymous with Lanark, based at the Beacon in Antwerp, is our guest today. Lanark operates as a project organisation, where supply chain engineers provide support in implementing potential solutions. The starting point is always the necessary experience of our own people, who take a hands-on approach to problem-solving. But the company also employs young graduates who, in the form of detachments in operational roles, with their feet firmly on the ground, slowly but surely gain the necessary experience.

Wim, an engineer by training, is more at home in this world. At Procter & Gamble, he entered the world of supply chain and gained a lot of experience. His personal milestone, however, was at battery manufacturer Duracell. Within a year, a complete supply chain turnaround was achieved. Finding chaos and bringing structure to it is Wim's life.

What trends do you see?

In the world of supply chain, there is never a dull day. Turbulence is inevitable, but Wim still sees a few major trends.

Multinationals that have ridden the wave of big data are now also thinking in terms of AI. Unfortunately, growing SMEs are still struggling to get the basics of their supply chain in order. They are still in the phase of implementing ERPs, the daily operational reality.

For the latter, “big data” is still uncharted territory. And yet these SMEs suddenly want to get started with AI. Today, there is an advantage for this type of company: the latest solutions for their supply chain now come with AI already built in. This is a unique opportunity to catch up in this way.

There is, therefore, a clear difference in maturity between these two groups. Volume will be the key here, but this comes with considerable overhead. Manufacturing companies with a good product will always have a reason to exist, no matter how difficult the future may seem. But they will have to invest, including in systems.

It is mainly family businesses that are hesitant, because the investments are substantial, so they are trying to continue to rely on their existing, but outdated, ERP infrastructure.

In these turbulent times, this is obviously not an easy decision. But in a crisis, you have to invest to come out stronger. As a smaller entity, you even have to dare to look for collaboration, possibly considering a merger. You can postpone this for a while, but ultimately, you will come to a standstill.

Where are we headed in the long term?

There will certainly be further consolidation in the distribution and transport sector, and the tangle of small companies will naturally find its way. They are also looking for a success ratio in the various systems in the supply chain, but this often remains below expectations. The companies that are consolidating are building large mastodons, systems that can literally control everything, right down to the coffee machine.

But we should rather look at systems that are “good at what they do”, a combination of micro-services that are part of the “best of breed”. We will have to live with the pendulum swinging between the monoliths and the micro-service solutions for a while longer.

And AI will certainly play a role there, but what role? Can we agree on a data model? Can we agree on “business requirements”? Can you rely on AI to come up with a working application? Where does that leave the big ERP systems of this world? The big ones among them are probably already incorporating this philosophy.

Whether fully monolithic or fully microservices, these choices will face difficult times. The answer will probably lie in a balanced combination.

But the basic starting point must always be the master data. Do we have it correctly in our hands? Are we going to do it ourselves? Or give control to external parties?

The answer is likely to be hybrid. The “architect” does not necessarily have to be internal. But he does have to understand your business. Then his know-how becomes added value. But such profiles are becoming rare.

Moreover, there is an additional constraint: the world of orders is very diverse, assignments are complex, data is difficult to map in the systems, a stock order and a transport order are two different worlds.

Where does Hyperfox stand in this world?

How does Wim view the solution offered by Hyperfox? He realises that order processing in most companies is still manual. The fax, for example, is far from obsolete. The effort involved in this is enormous. Not to mention the eternal debate about whether order management is the responsibility of Sales or the supply chain.

The Hyperfox solution is specifically designed to address the difficulties large companies encounter when processing orders. This is especially true if they still rely on older systems that do not provide a single, clear overview. Working with multiple screens is not practical, which is why the culture of Excel sheets has taken hold. Or why companies start building their own solutions around different systems. And where data is combined, extracted and re-imported.

By the time they discover Hyperfox, they have already passed this phase.

According to Wim, configurability is the key. This is important, as every company has its own specific fields. And then there is also diversity between specific orders. Training language models or those specific elements is therefore not unimportant. But then you also get good results in return.

A generic system must be language model independent. One model executes the command, the other critically reviews the response. It is a form of checks and balances. But above all, the architecture around it must be mature.

And the connotation with a niche, a transport order is fundamentally different from a stock order. That implies different workflows and interfacing with different systems. That is an insight that is often neglected.

And there's Excel again

And there's Excel again. You could almost say that without it, the world would stop turning. It's no different in the world of supply chain management. But it proves its usefulness. In terms of data, it's easy to connect and load, and the macros, although a nightmare to maintain, can prove their usefulness. And if you stay in the world of Microsoft, you'll see that Power BI is becoming increasingly popular.

It is becoming increasingly integrated into workflows, and integration with data is no longer just about “reading”; interaction is also possible. On the other hand, it remains complicated, and the necessary know-how is a prerequisite.

But Wim also sees an openness to building small applications. Not for crucial functions, but for niche flows that you cannot get rid of in an ERP, unless you want to pay a fortune.

When we talk about best practices.

There is such a thing as a “cost to service” analysis. What does delivering an order to a customer yield? Logical question? Many companies calculate once and arrive at an average cost that is entered into the model. And as a result, they lose sight of the actual cost. If you then try to look at the reality, which takes some effort, it often opens your eyes. It immediately becomes clear where you need to focus your sales efforts, especially on customers with special requirements.

Then there is another well-known example, the implementation of systems such as a WMS, which involves a lot of work. It is one of the few moments when you have to connect a physical flow with the data flow for the processes to work properly. It's not easy; you need to have a fade-out/fade-in plan, you need to have a plan for the transition moment, because if you can't send it after all. It doesn't have to be a failure; if you're well prepared, it should work.

Another example, the willingness to give the planner space. You need to make the tools and information he needs available. Because planning is like solving a puzzle, it's about freeing up space to create breathing room in the planning schedule. For example, planning based on capacity instead of just dumping orders into the “plant”.

Listen on YouTube