Your ERP is old. But replacing it scares you more than keeping it.
I build custom management software around your company's real process: we start from the piece you use every day, migrate without stopping anything, and in the end the code and data are yours. And if a standard ERP is enough for you, I'll say so in the first call.
30 minutes · If custom isn't worth it I say so · Reply within 24 hours
The signs your ERP has had its day.
- The ERP is ten years old, runs on a server nobody wants to touch, and whoever wrote it is long gone.
- The vendor replies when they feel like it, and every change costs as if it were the first one.
- Half the real work happens in Excel, outside the system, because it doesn't fit inside.
- Every year you pay a licence for modules nobody has ever opened.
- Changing scares you more than staying: "there are twenty years of data in there".
The fear isn't the cost. It's getting stuck halfway.
When a business owner puts off replacing the ERP, it's almost never about the quote. It's the picture of a Monday morning where the new system doesn't work, the old one is already off, and nobody can invoice.
It's a reasonable fear, and it's why the method matters more than the technology. Old and new coexist as long as needed. One piece moves at a time, gets verified against real data, then the next. There is no D-day where everything is switched off and you hope.
The other half of the fear is the history: "there are twenty years of data in there". That data is the starting point, not an obstacle — the export from the current system tells you more about how you really work than any meeting. I've done exactly this on a legacy invoicing system with over 1,000 users, migrated to microservices without interrupting operations.
A system that follows you, not the other way round.
Built on the process you actually have
We don't start from the software, we start from how you work: what goes through whom, where the errors come from, what gets done by hand because the system has no place for it. Then we build that, and only that.
Your code, your data
Version-controlled, documented code that you own. Data where you decide, in Europe. No licence that keeps growing, no vendor who can hold you still.
The migration doesn't stop the company
Old and new coexist as long as needed. One piece moves at a time, gets verified, then the next. No weekend where everything is switched off and you hope.
Twenty years of data don't get thrown away
History gets migrated, verified and stays queryable. It's the company's asset, not baggage: almost always it's also the best way to understand how you really work.
It talks to whatever stays
E-invoicing, e-commerce, CRM, shop-floor machines: the new system talks to what you keep, instead of forcing someone to retype data by hand.
One person, not an agency
I build it: the same person who decides the architecture writes the code and answers for it. No account manager, no handover to a junior after you sign.
When custom is NOT the right call.
If a standard ERP is enough for you
If your process is everyone's process — invoices, stock, deadlines — an off-the-shelf product costs less and does the job. I'll be the first to say so in the call, for free.
If the problem is organisational
Custom software photographs a way of working. If the way of working is the problem, new software crystallises it instead of fixing it: sort that out first.
If you're shopping for the lowest quote
It isn't the cheap choice in year one: it's the one that stops charging you per licence and per customisation. If your horizon is short, it isn't worth it.
Four steps, without switching anything off.
A 30-minute call
You tell me what isn't working. If custom isn't worth it, I say so right there.
Audit of the current system
What you actually use, what happens outside in Excel, what migrates. Written report, yours either way.
Start from the core
First the piece you use every day, in production. Then the rest, by priority.
Side-by-side migration
Old and new coexist as long as needed. What stays at the end is code, documentation and access.
Step two is the same method as the technical audit. If the piece that chafes is the sales side, look at the custom CRM; if instead it's about building something that doesn't exist yet, the path is building a software product.
The questions I get about custom ERP.
How much does a custom ERP cost?
It depends on scope, and scope gets defined after the audit — before that it's hot air. What I can tell you is how it's kept down: you don't rebuild everything, you start from the 20% you use every day plus the integrations that matter. The price is agreed in writing before we start, and from there on there are no per-user licences.
We have twenty years of data in there. Do we lose it?
No — it's the starting point. History gets migrated and stays queryable; the export from your current system is also the best way to understand how you actually work. Nothing gets thrown away: it gets freed from the platform.
Do we have to stop the company during the migration?
No. Old and new coexist as long as needed, and one piece moves at a time, verified before moving on. I migrated a legacy invoicing system with over 1,000 users to a microservices architecture without interrupting operations: the real work isn't writing the new thing, it's switching off the old one without hurting anybody.
And if you're not around anymore?
That's the right question, and it's why I hand over the way I do: code you own, version-controlled, documented, with tests where they matter. Any serious developer can pick it up. The stated goal is to not be indispensable — the opposite of the situation you're leaving.
You or a software house?
Depends on size. If it needs a team of ten for two years, a company makes sense and I'll say so. For an SMB's management software, usually not: with an agency you talk to a salesperson, then a project manager, and someone else writes the code. Here the person who decides the architecture is the same one who writes it and answers if something breaks.
The system should work for you.
30 free minutes: tell me what isn't working and I'll tell you whether custom makes sense — and if it doesn't, which path does. Either way you leave with a clear direction.
30 minutes · No commitment · Reply within 24 hours