Almost every organisation reaches the same point. The spreadsheets have grown into a maze, approvals live in chat groups, the accounts tool knows nothing about the warehouse, and leadership waits days for numbers that are already out of date. Someone says the word ERP, and a harder question follows: buy a ready-made product, or build a system around how we work?
There is no answer that is right for everyone. There is a reliable way to decide.
Start with the work, not the software
Before comparing products, write down how the work actually moves today. Not the organisation chart: the path a purchase request takes, how stock moves between branches, who approves what and how a month is closed. Two pages are enough.
This exercise does two things. It shows which processes are ordinary, the same as in thousands of other organisations, and which are yours: the way you price, the way you serve customers, the way your branches work together. And it gives every vendor the same picture, so their answers can be compared.
When an off-the-shelf ERP is the right choice
A ready-made product is usually right when:
- Your processes are standard. General ledger, payables, receivables and payroll work the same way in most companies. Mature products do these well.
- You can adapt to the product. If people are willing to change how they work to fit the software, implementation is faster and cheaper.
- The product is supported locally. Someone near you must understand the product, Ethiopian tax and the fiscal year that starts in Hamle, and be reachable when month-end goes wrong.
- The licence cost stays reasonable as you grow. Price the product for the number of users you expect in three years, not today.
If all four are true, buy, implement it properly and spend your effort on training. A good technology partner will tell you this, even when it means less work for them.
When a system built for you makes sense
A custom system earns its cost when at least one of these is true:
- Your process is your advantage. How you route orders, price for different customers or coordinate field teams is part of why you win. Forcing it into a generic product loses that.
- The system must fit deeply with what you already run. Core banking, a hospital’s equipment, a factory’s machines, a government registry: when the value is in the connections, a system designed around them is simpler than a product bent to reach them.
- Licences grow faster than value. When hundreds of staff, agents or partners need access, per-user licences become the largest line in the budget.
- Local requirements run throughout. The Ethiopian fiscal year, local languages, local payment providers and local reporting are not add-ons for you; they are the system.
The costs nobody puts in the proposal
Whichever way you go, these are the costs that decide whether the project succeeds:
- Data migration. Moving years of records from spreadsheets, Peachtree or an older system, and reconciling totals before the old system is switched off. Plan time for it.
- Training. People use what they understand. Budget for training in the language your staff work in, and for a second round a month after go-live.
- Change after go-live. Your organisation will change. Ask how a change request is handled, how long it takes and who can make it.
- Ownership. With a custom system, insist that the source code, data and documentation are yours, written into the contract. With a product, insist on a full export of your data at any time.
A safe way to start: one module, live, then the next
The biggest risk in any system project is a long build that delivers nothing until the end. The safest approach is the same for both paths: choose the one process that hurts most, put it live in weeks, and let people use it before anything else is built.
For a distributor that might be orders and stock across branches. For a hospital, registration and billing. For a microfinance institution, loan applications and approvals. When the first module is in daily use, the next decision is made with real experience instead of a sales presentation.
Questions to ask any vendor
- Who owns the code, the data and the documentation when the project ends?
- Can we export all our data at any time, in a standard format?
- What does a change cost after go-live, and how long does it take?
- Who supports us at month-end, and how fast do they respond?
- Can we see a system you built running, not slides about it?
If a vendor cannot answer these plainly, keep looking. If you would like a second opinion on a decision you are weighing, tell us about it; we will say what we would do, even if it is not to build.