How to choose an ERP without regretting it in year two

A feature list is the easiest thing to compare and the least useful. Every system in this market will tick inventory, invoicing and reports. What separates a system you still want in year two from one you are fighting is rarely on that list at all.

Four questions tend to decide it.

1. Do two screens that should agree actually agree?

This is the question that matters most and gets asked least. Ask to see the stock valuation report and the inventory account balance in the general ledger, on the same date, in the same demo. Then ask for the customer statement and the receivables control account.

If the numbers differ, the system is keeping two sets of books: one for operations and one for accounting, reconciled by a person at month end. That reconciliation is a permanent job you are about to hire for. If they agree, ask why. The answer you want is that stock movements post to the ledger as they happen, not that a nightly job copies one into the other.

Ask the same question about cost. When a purchase price changes, does the cost of the items already in stock change, and does that change reach the ledger? A system that recalculates the report but not the journal entry will show you a profit figure no accountant can defend.

2. What does a change cost, and who can make it?

Your business will change its pricing rules, its approval chain and its branch structure within two years. The real question is not whether the system supports a change, but what the change requires.

• If adding a discount rule needs a developer and a deployment, you will stop adding discount rules.

• If a new approval level needs a custom build, approvals will migrate back to WhatsApp.

• If adding a branch needs a new installation, you will run two systems and reconcile them by hand.

Ask to be shown the change, not told about it. Add a second price list for wholesale customers, now, in this demo, is a fair request, and the answer is informative either way.

3. Who owns the data, and can you leave with it?

Ask for an export during the trial rather than after you commit. The answer is more honest while you are still a prospect. Three things are worth establishing: whether every list exports, whether a full database export is available, and whether leaving costs you anything beyond the notice period.

A vendor who cannot produce an export of your own data is telling you something about the next five years. This is also the cheapest thing in the entire evaluation to verify, which is why it is worth doing first.

4. Does it work in the language your staff actually use?

In this region that usually means Arabic, and it means more than a translated menu. Check the printed documents, the error messages, the reports and the mobile apps. Arabic that stops at the interface and reverts to English on an invoice is a system your customers will see in English.

Check the direction too. A right-to-left layout that was retrofitted shows up as misaligned tables, numbers in the wrong column and dates that read backwards. Small things individually, and a daily irritation for everyone who uses it.

What to do with the answers

Write the questions down before the demo, and insist on seeing rather than hearing. A vendor who shows you the ledger beside the stock report has already solved the hard problem. One who moves on quickly has told you where to look later.

And weigh an honest no higher than a vague yes. A system that says plainly it does not do something is easier to plan around than one that says it does and reveals the limits in month three.