Every growing business hits the same wall: the spreadsheets stop scaling. Inventory lives in one Excel file, invoices in another, the accountant keeps a third, and nobody's numbers agree. That wall is where the ERP conversation starts — and where a lot of money gets wasted, in both directions.
I build business software for companies across Egypt, Saudi Arabia, and the wider MENA region, and the honest answer to "should we build a custom ERP?" is: usually not all of it. Here's the decision framework I walk clients through.
What an ERP actually is
Strip the acronym away and an ERP is one promise: a single source of truth for the operational state of your business — what you own (inventory, assets), what you're owed and owe (receivables, payables), and what's in motion (orders, production, deliveries). Every module is just a view over that shared state.
That framing matters because the failure mode of ERP projects — bought or built — is the same: multiple sources of truth surviving the rollout. If the warehouse still trusts its notebook, you didn't implement an ERP; you bought expensive software.
When off-the-shelf is the right answer
If your processes are standard — buy goods, hold stock, sell, invoice, collect — an established system (Odoo is the common choice in MENA, and the big vendors above it) will be faster and cheaper than anything custom. Buy it when:
- Your workflows fit the tool with light configuration, not customization
- You can adapt your process to the software's defaults without losing what makes you competitive
- Your pain is discipline (data entry, reconciliation) rather than fit
A custom build to replicate standard accounting and inventory is money burned. Accounting especially: the ledger rules are solved problems — you want correctness patterns, not novelty.
When custom ERP development pays off
Custom wins when your operation itself is non-standard and that non-standardness is your edge. Real signals:
- The workflow is the business. A repair-on-wheels fleet, a multi-branch distributor with consignment stock, a factory with a peculiar costing model — when the process is the moat, forcing it into a generic ERP's boxes means either mutilating the process or drowning in customization fees that exceed a clean build.
- Off-the-shelf customization has already failed you. Many MENA businesses come to me after an implementation partner spent a year bending a generic system. Heavily customized deployments are the worst of both worlds: upgrade-hostile, consultant-dependent, and still not quite right.
- You need Arabic-first, bilingual operations for real. Arabic RTL across documents, approvals on WhatsApp-speed mobile UX for staff who won't sit at desktops, local e-invoicing compliance (Egypt's ETA, Saudi ZATCA) — these are first-class requirements, not plugins.
- The ERP must talk to your customer-facing systems. If you run a storefront, a marketplace, or field apps, the operational backend and the customer experience are one system. That integration is where custom architecture shines: one backend, shared state, every device.
The hybrid answer most businesses actually need
The best-value architecture I deploy is usually a custom operational core + standard accounting:
- Custom modules for the workflows that differentiate you — orders, dispatch, production, field operations — built exactly to how the business runs
- A boring, proven ledger/accounting layer (or an integration into the accountant's existing tool) for the parts where "standard" is a feature
- One PostgreSQL source of truth, with role-based web dashboards (Next.js) and mobile apps (Flutter) for staff in the warehouse, on the road, and at the counter
This keeps the build scope to months, not the multi-year death marches ERP is famous for.
How to scope an ERP build that actually ships
- Map the money and the goods first. Follow one order from placement to cash collected, and one product from purchase to sold. The systems that touch that path are phase one. Everything else is phase two or never.
- One module live at a time, real data from day one. An ERP rollout is a migration of trust, not just data. Each module must replace its spreadsheet completely before the next starts.
- Design for the phone. In MENA operations, the people creating the data — drivers, storekeepers, sales reps — live on mobile. If data entry needs a desktop, your single source of truth dies within a month.
- Reports are the product. Owners don't want screens; they want to know margin, cash position, and who owes them, today, from their phone. Build the owner's dashboard early — it's what keeps leadership pushing adoption.
What it costs
A focused custom operational core — 2–4 modules, web + mobile, bilingual — is a 3–6 month build with a senior engineer, not a seven-figure enterprise program. The discriminator is scope discipline: ERPs fail by trying to be everything in version one.
If you're staring at the spreadsheet wall — or at a generic ERP quote that made your eyes water — talk to me. I'll tell you honestly which parts you should buy, which you should build, and what phase one looks like. First call is free.