ERP is where mid-market companies go to spend a year discovering their own processes. Sometimes that is exactly what they needed. Often the project is a symptom of a different problem and the software cannot solve it.
The signal that you actually need one
Not growth. Not headcount. The real signal is reconciliation: when a meaningful part of somebody's week is spent making two systems agree about the same number. Stock in the warehouse against stock in the accounts. Invoices raised against work delivered. Payroll against project cost.
When people are the integration layer, you have outgrown your tools. Until then, better use of what you own usually beats a migration.
Why these projects fail here specifically
- The processes were never documented. Implementation becomes an eighteen-month discovery exercise billed at implementation rates.
- VAT and local compliance are treated as configuration. UAE and Saudi requirements, particularly around e-invoicing in KSA, are not an afterthought and not the same between markets.
- Multi-entity and multi-currency arrive late. A group with entities in several territories has this as a founding requirement, not a phase two.
- Nobody owned it internally. An ERP without a full-time internal owner is a purchase, not a project, and it will drift.
What to do before you shortlist anything
Document the ten processes that actually move money. Not the ideal versions. The real ones, including the spreadsheet somebody maintains privately and the approval that happens verbally. That document is worth more than any vendor demo, and producing it sometimes ends the project honestly and early, which is a good outcome.
The regional shortcut worth knowing
Full-suite implementations are not the only option. A great many Gulf SMEs are better served by strong accounting software, a proper inventory or operations system for their specific trade, and a small amount of integration between them. It is less impressive on a slide and it delivers in months rather than years.
The industry has an incentive to present this as a compromise. For a company under a few hundred staff it is frequently the better system.
If you do proceed
Phase it by pain, not by module. Start with the reconciliation that hurts most, prove it in one entity, then extend. A go-live that covers one real process completely beats one that covers nine partially, because partial coverage means people keep the spreadsheets and you have bought the software without retiring the problem.
