An ERP implementation in a small or medium business takes three to nine months and costs, all in, several times more than the licences. The most expensive part is not the software but your people's time — and that is also why half of all ERP projects disappoint.
The good news: it can be done well, as long as you start from the real problem and accept that an ERP does not fix bad processes, it makes them visible.
Signs that the time has come
Don't look at turnover, look at the symptoms. Every department has its own spreadsheet and each one turns up at the meeting with a different number. Month-end close takes two weeks. Nobody can say, without a day of calculations, which product actually makes a margin.
If you recognise three out of four, the ERP conversation is legitimate. If you recognise only one, you probably have an isolated problem that can be solved more cheaply.
When it is too early
Below a certain level of complexity, an ERP is a coat three sizes too big. If you have a single sales flow, fewer than ten people and an accountant who copes fine, the money is better invested in tidying up the data you already have.
An ERP bought too early does not speed the business up. It slows it down, because it adds steps to a process that used to run on intuition.
The real cost, broken down
Licences are the tip of the iceberg. Underneath them sit implementation consulting, migration of the old data, integrations with what you already have, training and, the item most often forgotten, your team's time pulled away from day-to-day work for several months.
Ask for a quote itemised along these lines, with a separate budget for "what we discover along the way". Projects that don't plan for this don't come in cheaper, they just stop halfway.
How to reduce the risk
Start with a single flow — usually sales and stock — and get it fully up and running before moving on. An ERP delivered in stages shows you early whether the choice was right, while you can still change direction.
And appoint someone in the business as project owner, with officially allocated time. Without that person, the decisions get made by the vendor, and the system will end up looking like somebody else's business.
Frequently asked questions
Cloud or on your own server?
For small and medium businesses, cloud in most cases: no server to maintain, updates included. Your own server makes sense when you have strict data requirements or heavy local integrations.
Can we migrate the data from our old software?
As a rule, yes, but migration is the moment of truth: dirty data in the old system becomes dirty data in the new one. The clean-up happens before, not after.
How much should we customise?
As little as possible at the start. Every customisation is paid for again at every update. First use the system as it comes, then customise what genuinely gets in your way.
Who should lead the project?
Someone from operations, with technical backing. An ERP project led exclusively by IT ends up technically correct and unusable in practice.
