ERP or custom software?

ERP is the idea of gathering all of a company's processes into a single system. Done right, it is powerful. Done wrong, it becomes the most expensive software mistake a company makes.

This piece is not an argument that ERP is bad. The argument is this: ERP is not a product choice, it is a scope decision, and in most companies that scope is drawn wider than it needs to be.

Why ERP projects struggle

The company gets bent to fit the system. ERP packages arrive with assumptions about how you operate. If your operation differs, you have two options: change the process or customise the system. The second gets expensive quickly.

There is an important distinction here. Some processes benefit from being standardised; nobody differentiates themselves from competitors by how they issue invoices. Other processes are the company itself. Bending those to fit a package means erasing the difference.

Every department is affected at once. Accounting, production, sales, and purchasing all migrate together. A problem in one place affects the whole company. A single module slipping moves the entire schedule.

Customisation locks you out of updates. Customisations made to a standard package collide with the vendor's next release. A few years later you are stuck with a system that cannot be upgraded. From that point on, every new requirement gets solved by working around the system.

Data migration gets underestimated. Moving years of inconsistent data out of the old system is usually the longest part of the project. Three different spellings of the same customer, old records never closed, and mandatory fields left empty all surface here.

No budget is set aside for training. Even if the system works technically, the project has failed if the people using it revert to their old methods. This item is missing from most proposals.

ERP is the right call if...

• Your processes run close to the industry standard.
• You have multiple locations or legal entities and need consolidation.
• Regulation requires an integrated record system.
• You have an internal team that can manage the change.
• Your audit and reporting obligations are heavy.

The last two are the most commonly skipped. ERP is as much a change management project as a software project. Without someone who can make decisions and get process owners around a table, the project finishes technically and goes unused.

Custom software is the right call if...

• Your competitive advantage lies in a specific process that does not fit a standard package.
• You need to fix one or two blocked processes, not the whole company.
• You do not want to replace your existing programs, only connect them.
• Per-user licence costs compound as the team grows.
• Your process has not settled yet and is expected to change over the next year.

That last point matters. Freezing a still-changing process into an ERP is expensive; every change means consultant hours.

The third path: combine them

This is what we recommend most often in practice, and it is the right answer for most mid-sized companies.

Stay off-the-shelf where you are standard. Accounting, e-invoicing, and payroll are mature areas. Rewriting them returns nothing. They also get updated as regulation changes, and you do not want to carry that burden.

Build your own where you are different. Production planning, field operations, and dealer management are typically in this group.

Connect the two with integration. Data is produced in one place and flows to the other, and duplicate entry disappears. What is API integration explains how that connection is built.

The advantage: you are not risking the whole company at once, and you can replace each piece independently. If you want to renew the production planning system in two years, you do not have to touch accounting.

The disadvantage, stated plainly: the integration layer needs maintenance. When one system changes its API, the bridge has to be updated. That cost is small but it is not zero.

Which process goes in which group?

There is a simple question that makes the decision easier: does your competitor do this the same way?

If yes, use an off-the-shelf solution. Issuing invoices, calculating payroll, and sending e-invoices are in this group.

If no, and the difference works in your favour, build your own. How you produce a customer-specific production plan, your dealer discount rules, or your field team's job acceptance flow are in this group.

If no, but the difference gains you nothing, changing the process is cheaper than changing the system.

Measure before deciding

For one month, count the following:

• How many separate programs do you enter the same data into?
• How many days does month-end close take?
• How many people do you have to ask to find one piece of information?
• How many reports are prepared by hand, and how many hours does each take?

Those four numbers go a long way towards telling you whether you need ERP, integration, or a single custom application. Often the answer is the third, and it is also the cheapest.

Compare budgets properly

ERP proposals usually lead with the licence fee. To see the real cost, add these:

1. Implementation and consulting days
2. Customisation development
3. Data migration
4. Training and the parallel running period
5. Annual maintenance and version upgrades

Do the same arithmetic for custom software: development, infrastructure, maintenance, and next year's development allowance. Put the two tables side by side and the decision usually makes itself.

---

Describe your processes and we can work out which you need. If we think you need neither, we will say so. Get in touch.