The End of ERP as We Know It | Newsletter

Have you ever tried to get a grip on your production with an ERP system?

Then you know how it goes. The demo looks great. The implementation takes longer than promised. The integrations cost more than budgeted. And in the end you’re working with workarounds because the system doesn’t quite do what your business needs.

I’ve been helping manufacturing companies with their systems for ten years. And the whole time I’ve seen the same thing. Companies paying tens of thousands per year for software that doesn’t understand their processes. Integrations that depend on one consultant at the vendor, and when they leave, there’s no one left. Features on the roadmap, next quarter, or the quarter after that, or never.

I call it false security. You think you’re mitigating risk with a big name and a support contract. In practice, you’re dependent on people you don’t know, at an organization you have no control over, for systems you can’t access.

I’ve believed in the power of best of breed for years. Not a monolithic ERP, but an architecture where the best tools for each function work together on a Unified Namespace. But implementing that at companies at scale was difficult. Too expensive. Too much development work. Too dependent on specialists.

That game has changed.

The shift: your company builds its own software.

Not because you’re becoming an IT company. But because with vibe coding you can build what you need, when you need it. You learn it. You do it yourself. On a solid foundation.

Over the past year I’ve been helping clients with this. And what I see surprises me. How fast it goes. How strong the solutions are. How little prior knowledge is needed. Especially for SMEs, the smaller your company, the faster you can move.

One of my clients, a metalworking shop with forty people, built his own MES. Shopfloor, time tracking. His MES subscription? Cancelled. His ERP? Going out soon too. Savings: tens of thousands per year. Features he waited years for from vendors, he built himself in weeks.

But there are conditions. Anyone who starts without a foundation is building tomorrow’s legacy.

Three things determine whether this will work for you:

Start with infrastructure, not applications, without a data hub it becomes spaghetti
Know what you integrate and what you replace, not everything can, not everything should
Have a vision for your digital future, this is strategy, not a side project

1. Start with infrastructure, not applications

I see the same pattern at every company.

You start with an ERP. That works. Then another package is added for planning, or for quotes. Then a new machine, different brand, different software. Another integration. And another.

And then it becomes chaos.

Your laser connects to your ERP, but also directly to your planning, because that interface in the ERP doesn’t work well. From there another line to a dashboard. You have integrations via XML, via CSV, via APIs, all mixed together. Five vendors, each with their own consultants, who don’t understand each other’s systems and point fingers when something breaks.

Dozens of connections. No overview. Data that doesn’t add up. People entering the same information three times. And yet someone’s standing there with Excel on the side.

Every new integration costs tens of thousands and months of lead time. And if something changes, new machine, different package, the circus starts over.

That’s spaghetti. And what if the consultant who built one of those connections leaves? Then you’re stuck with integrations no one understands anymore.

There’s a different way.

A Unified Namespace breaks this pattern. Here’s why:

You prevent spaghetti. The UNS is your backend, one central place where all your data comes together. Your ERP talks to the hub. Your machines talk to the hub. Your planning talks to the hub. Never directly to each other. No more dozens of connections, everything through one hub, one format, one structure. New machine, different brand? Plugs into the same hub.

You get real-time visibility across your entire business. Only then do you see where you’re losing money. Not after the fact in a report, not next month when the accounting is done. Now. Which machine is idle. Which order is running late. Where the margin is disappearing. You can’t improve what you can’t see.

You no longer need ERP as the central hub. The UNS plus your own apps do what your ERP did, but the way your business actually works. The apps become disposable, the data and the architecture stay. No roadmaps. No implementation projects. No dependency.

2. Know what you integrate and what you replace

But what about maintenance? And scalability? Can you really build everything yourself?

No. Not everything can be built yourself. Not everything should be built yourself.

The question is where the boundary lies. I look at three criteria.

Legally required. Your accounting needs to meet legal requirements. You don’t want a self-built system you have to defend during an audit. You buy that, from a party that keeps up with tax regulations.

Technically fundamentally dependent. Your CAD/CAM is intertwined with your machines, with file formats, with years of development. That’s proprietary, not necessarily because the vendor wants it that way, but because it grew that way technically. Replacing it today isn’t realistic.

Mission critical or liability. Machine controls, network security, you don’t want self-built experiments there. Though you can go further than you think. Linux runs on most industrial systems. Open source firewalls are mature technology. But you need to know what you’re doing, and where the boundary lies of what you can take responsibility for.

Everything that doesn’t fall into those three categories? You build it yourself.

And that’s more than you think. Entire flows for purchasing, invoicing, quotes, at most SME companies that’s massive overkill in an ERP. You’re better off assembling that yourself with a few prompts and connecting it directly to the network. Look at SCSN, the Smart Connected Supplier Network. You can just connect to it, without your ERP in the middle.

Those boundaries are also shifting. DNC, file transfer to CNC machines, is starting to decouple. My next project is an open source CAM interface system. I built Eryxon MES, open source, specifically for job shops in metalworking. Step by step, more becomes possible.

But today, for most companies: planning, shopfloor, time tracking, workflow, purchasing, quotes, that’s where the value is. That’s business logic. How your company works. And that’s exactly where ERP failed, because it can’t model how your specific business operates.

3. Have a vision for your digital future

And what about strategy? You can’t just start building without knowing where you’re headed.

This isn’t a side project. This isn’t “we’ll also do some digitalization.” This is a fundamentally different way of looking at how your business works.

You need to know why you want to connect your machines. Why you want to integrate systems. What you want to see, know, improve. Where you want to be in three years.

That also means: looking at your IT landscape. What licenses are you paying for and why? What about security if you’re building yourself? Who manages what? Who learns what?

This is digital strategy. Not calling a vendor and asking what they can deliver. But deciding for yourself what you need, and then building or buying what fits.

What has changed

Why couldn’t this be done five years ago?

AI writes usable code. The models from the past year are a tipping point. Someone who knows their processes can now describe what they need. The AI builds it. In days or weeks, not months.

The bottleneck was always technical capacity, programmers who are scarce and don’t understand your business. That bottleneck is gone. What remains is domain knowledge. And you already have that.

Low-code was the stopgap. That era is over. Why work within the limitations of a platform when you can have the code generated yourself?

The Unified Namespace exists. Event-driven architecture that scales and where everything comes together. No spaghetti of point-to-point integrations, but a central hub where everything connects.

This isn’t theory. This is what I’ve seen working at clients over the past year.

When should you choose ERP?

I’m not dogmatic. There are scenarios where ERP is still the right choice.

A) You don’t need much. You’re a trading company, a small welding shop, you do simple things. Then a simple system is enough. Choose simplicity, look at Odoo, keep it light.

B) You have international obligations. Shareholders, compliance, reporting for multiple countries. Then you can’t avoid robust financial systems, Business Central or above. But remember: separate production and finance. Don’t get pulled into running your entire production in an expensive, complex ERP. The financial side can be complex, the production side shouldn’t be.

C) You have no digital ambition. You’re happy with the status quo. You don’t expect to change dramatically in the next ten years. This can be a deliberate strategy. But if you don’t have an ERP yet, ask yourself why you’d get one.

D) You work in a group. You already have three companies running on the same systems, company four and five are joining. Then it can make sense to pull them into the same administration. But even here I’d strip the ERP down, use it for what it does well, build the rest yourself.

Conclusion

The game has changed.

ERP was the solution when you couldn’t build it yourself. Now you can.

Your company builds its own software. Not because you’re becoming an IT company, but because with vibe coding you can build what you need, on a UNS as foundation, exactly the way your business works.

Three conditions:

Start with infrastructure, without a UNS it becomes spaghetti
Know what you integrate and what you replace, accounting, CAD/CAM and mission critical systems stay, you build the rest yourself
Have a vision, this is strategy, not a side project

There are exceptions. If you need very little, choose simple. If you have international obligations, choose robust for finance, but separate production. If you have no digital ambition, stick with what works. If you work in a group, standardize smartly.

But for most metalworking SMEs? I can no longer in good conscience recommend a complex ERP implementation. A solid accounting system. Integration with your CAD/CAM. And build the rest yourself on a solid data hub.

The smaller you are, the faster you can move.

Want to know what this looks like for your business?

Most companies need guidance during this transition. Not to do it for you, but to ask the right questions, make the right choices, and train your team so you can do it yourself.

Check out the Masterclass Paperless Factory, I’ll map out your situation and create a concrete action plan.

Or start with the UNS Workshop if your team is ready to learn hands-on.

Get in touch, tell me what’s not working, and together we’ll figure out where to start.

Talk soon,

Luke