The ERP customization trap
Every ERP customization adds cost for as long as you run the system. Why staying close to standard pays off, and how to decide what's worth customizing.
Every ERP consultant has heard "we're different". Most have helped customize a system to "match how we work". Years later that customization is the biggest barrier to upgrades, the largest line in the support budget, and sometimes the reason the company starts looking for a new ERP altogether.
Customization isn't bad in itself. The trap is that its cost is paid mostly after go-live, by people who weren't in the room when the decision was made.
How customization grows
The trajectory is familiar. In the first year a handful of custom workflows and fields get added to match existing processes. In the following years new managers bring new requirements, and customizations start to be built on top of other customizations. Then the vendor ships a major release. Every customization has to be checked, often rewritten, and retested, and the upgrade turns into a project of its own.
At some point the upgrade gets postponed indefinitely. The company is stuck on an old version: security patches stop, new integrations don't fit, and the people who wrote the custom code have moved on. The system wasn't outgrown. The customizations were.
Why the cost doesn't stop at go-live
- Upgrades. Every vendor update has to be tested against every customization.
- Support. Vendors and partners usually support the standard product. Custom code is your problem.
- Knowledge. Undocumented customizations live in the heads of a few people. When they leave, the logic becomes archaeology.
- Integrations. Changed data structures break connectors that expect the standard model.
- Onboarding. New employees can't use public documentation or training, because your system no longer matches it.
Five questions for every customization request
- Can we change the process instead of the system? Often the honest answer is yes. The process exists because the old system worked that way.
- Is it truly differentiating? Customize what makes your business different, like pricing logic or production specifics. Not how you receive purchase orders.
- Does a standard module cover most of it? Use it even if the fit isn't perfect. The remaining gap is usually less painful than the maintenance debt.
- Can it be an extension rather than a modification? Separate modules and external services survive upgrades far better than changes to core code.
- Who will maintain it, and at what yearly cost? If nobody can answer, the request isn't ready.
Extension over modification
When the business genuinely needs something the ERP doesn't do, keep the core standard and build the missing piece next to it. A separate application talks to the ERP through its API: a customer portal, a production planning screen, a pricing calculator, a sync with a marketplace. The ERP stays upgradeable, and the custom part has its own codebase, tests and release cycle.
That's exactly the kind of work that fits a custom web application rather than another layer of ERP modifications. It also gives you a clean exit: if you ever change ERPs, the extension needs a new connector, not a rewrite.
Common mistakes
- Rebuilding the old system screen by screen inside the new one.
- Letting each department request customizations directly from the implementer, with no central review.
- No registry of what has been customized and why.
- Custom code without documentation or tests.
- Customizing built-in reports for every request instead of feeding the data into a separate reporting tool.
Checklist
- A customization registry: what, why, who asked, who maintains it, upgrade risk.
- A single person or small board approving every new customization.
- Preference order agreed in writing: change the process, then configure, then extend, and modify the core only as a last resort.
- Every customization tested as part of each vendor update.
- A yearly review that removes customizations nobody uses.
Companies that start looking for a new ERP often haven't outgrown the system. They've outgrown their own customizations. Keeping close to standard is less exciting during implementation and much cheaper over the following years. For the timeline side of the same story, see ERP implementation in 90 days. And if you already run custom extensions around your ERP, they need the same care as any other product; that's what our web application support covers.
Weighing a customization request or untangling an over-customized system? Write to us on Telegram or MAX.
Got a project on this topic? Tell us the task — we'll come back with an estimate and a schedule.
Message on Telegram → Message on MAX