The hidden cost of the business software nobody wants to use anymore

The ERP has been there for fifteen years. It works. But there’s exactly one person who really knows how to use it, the interface is from 2009, and every new hire takes three months to become productive.

This isn’t an IT problem. It’s a hidden cost you don’t see in the books, but you pay every week.

In short:

• The real cost of legacy software isn’t the licence fee: it’s slow onboarding, hours lost in manual workarounds, and key-person risk (if the only person who knows it leaves, you’re stuck).
• Vendor lock-in is measured in time, not just money: how long would it take to export your data and run it somewhere else?
• You almost never need a big bang. Often it’s better to keep the core (accounting, e-invoicing) and replace the modules that hurt.
• Before deciding you need a baseline: how many hours that process costs you today. Without a number, every choice is an opinion.
• At Soraia, scoping a custom tool costs €2,000 (refunded if you proceed), build sprint €10-50k, first release in 4 weeks, your code from day one.

The cost you don’t see in the books

You see the licence fee. It’s a line, you pay it, you forget it. The rest, no.

In manufacturing SMEs I always see the same four hidden costs:

• Slow onboarding: a new employee takes weeks to become productive on a tool nobody has documented.
• Manual workarounds: the Excel export because the native report doesn’t exist, the double data entry because two modules don’t talk to each other.
• Errors from hostile interfaces: dense screens, ambiguous fields, wrong clicks. They cost corrections downstream.
• Key-person risk: I’ll cover this separately, because it’s the most dangerous.

None of these shows up as a cost line. All of them drain hours every week.

Key-person risk is the real alarm

There’s always that one person. They know where to click, they know the bugs, they know which field not to touch on Tuesdays. Everyone calls them when the ERP throws a tantrum.

The day they retire or change jobs, the knowledge leaves with them. It’s documented nowhere: it’s tribal knowledge inside one person’s head.

This is the strongest replacement signal of all. Not because the software is old, old works, but because you have a human single point of failure on a core process.

Lock-in: measure it in days, not euros

“Switching vendors is impossible” usually isn’t true. It’s just expensive in time.

The right question isn’t “how much does the new software cost”. It’s: how long would it take to get my data out and make it work elsewhere? If the answer is “I don’t know” or “months”, the lock-in is real and belongs in the maths.

It’s the same build vs buy reasoning I wrote about in the piece on custom software or SaaS. The four criteria (core process, cost over 2-3 years, lock-in, lifecycle) apply here too.

When NOT to replace

I’ll say it plainly, because replacing an ERP is a serious project:

• The core works and nobody hates it → don’t touch it. Accounting, e-invoicing, master data: if they run, they stay.
• The problem is only one module’s interface → don’t rebuild everything. Build the usable layer on top of the existing core.
• You haven’t measured today’s cost → stop. Without a baseline you don’t know if replacement makes sense, or you’re just buying the shiny new thing.

The “throw it all out and start over” big bang is the surest way to spend triple and freeze operations for six months.

When yes, and how to do it well

Replace when: key-person risk is high, onboarding costs more than you can afford, manual workarounds drain measured hours every week.

And it’s almost always better to start with one piece, not everything. With ILTEC, a B2B office technology firm in Biella, we didn’t rebuild their internal systems: we added the missing layer, an AI search optimized site in 3 weeks and a sales agent via QR code on their machine fleet. You start at the point that hurts most.

The same model applies to an internal custom tool: scoping €2,000 (refunded if you proceed), build sprint €10-50k, first release in 4 weeks, your code from day one. No perpetual subscription, no new lock-in replacing the old one.

And before writing code, often the fastest win is getting the tribal knowledge out of one person’s head: documenting processes and training the team. That’s what we do with AI Adoption.

The concrete first step

Measure what your ERP really costs today: onboarding hours, hours lost in workarounds, dependence on one person. That number tells you whether replacement makes sense, or whether modernising one module is enough.

If you want an honest outside view, let’s talk: 20 minutes, no pitch. Or start with the check-up.