Three routes to the same goal. How to choose by how your business actually works, not by the demo.
In brief
- There are three routes: buy a package ERP as it is, buy one and customise it, or build your own.
- No route is best in general. The right one depends on how different your processes really are from the industry norm.
- The costly mistakes usually come before the choice: unclear processes, no owner, and an all-at-once cutover.
- Start by writing down the workflows that make your business different. Everything else can follow the package.
Most legacy replacement projects start with a vendor demo. The demo is polished, the screens look modern, and the question quietly becomes “which package looks best?” Eighteen months later the project is late, the customisations have multiplied, and the people who run the business still keep their own spreadsheets.
The better starting point is not the software. It is a plain description of how your business works today, and which parts of that are worth keeping.
The real issue is not old software. It is old assumptions.
A legacy system is rarely a problem only because it is old. It is a problem because the business has changed around it: new products, new channels, new reporting needs, new rules. Staff fill the gaps with spreadsheets and manual steps. Over the years, those workarounds become the real process, and nobody has written it down.
If you replace the system without surfacing that real process, you either lose it (and people rebuild it on the side) or you pay to recreate it in the new system as customisation. Either way, the project cost is decided by something nobody measured at the start.
Three routes at a glance
| Package ERP | Customised ERP | Custom-built | |
|---|---|---|---|
| What it is | Buy a standard product and adapt your process to it | Buy a standard product and change it to fit your process | Build software designed around your process |
| Best when | Your processes are close to the industry norm | Most processes are standard, a few are truly different | Your process itself is the competitive edge, or no package fits |
| Strength | Fastest to start, proven, supported by a vendor | A balance of speed and fit | Exact fit, full control |
| Risk | Staff resist changes to how they work | Customisation cost grows and makes upgrades hard | Highest cost and longest timeline, and you own all maintenance |
| Watch for | Hidden licence and module costs | Customisation that quietly replaces the standard product | Dependence on a few people who know the system |
Four questions that choose the route
Ask these in order, and write the answers down before any vendor demo.
1. Which of our workflows are genuinely different?
List the five to ten workflows that matter most. For each, ask whether a competitor would do it the same way. Most will be standard (invoicing, stock, payroll). A few may be where you win business. Those few decide whether you need customisation or a custom build.
2. Who owns each process, and can they describe it?
A process nobody can describe cannot be configured or built. If the real process lives in one person’s head or a spreadsheet, document it first. This step often shrinks the scope on its own, because it shows what is no longer needed.
3. What must keep working during the change?
Decide what cannot stop: month-end close, shipping, payments. This sets how the cutover works: all at once, by module, or by business unit. Most failures come from choosing “all at once” because it looks cheaper on paper.
4. How much change can the organisation absorb?
A package asks people to change their habits. A custom build asks the business to define its habits precisely. Both take time from your busiest people. Be honest about their capacity before choosing.
A simple way to decide
- Processes close to the industry norm, limited capacity for change → package ERP, with a firm rule to adapt the process, not the software.
- Mostly standard with a few real differences → package ERP with a small, named set of customisations, each one justified in writing.
- The process is the product, or nothing fits → custom-built, started small with the one workflow that matters most.
- Unsure → run the existing workflow through a low-risk simulation first (see our article on simulating your ERP), then decide.
Five mistakes we see repeatedly
- Choosing from the demo. Demos show the standard path, not your exceptions.
- Customising to match the old screens. Familiarity is not a requirement. Each customisation raises upgrade cost for the life of the system.
- Leaving data migration to the end. Old data is usually messier than anyone expects. Plan it from the first week.
- No business owner. If the project belongs only to IT, the process decisions will be wrong.
- All-at-once cutover. Moving one module or one unit at a time costs more upfront and fails less often.
Our experience
We have built and connected back-office and operational systems for enterprise clients, and handled the data migration that goes with them.
We can explore which route fits your workflows, starting with the few that make your business different.
Which of your workflows would you least want to give up?
Choose the route that fits your real process
Package, customised or custom-built are all legitimate. The expensive choice is the one made before the business has described how it really works. Write the process down, find the few parts that are truly yours, and let that decide.
Talk to Balerion about mapping your current workflows and choosing a route for replacing your legacy system. Email us at talk@balerion.co.th.
