Excel or ERP: when should you switch?

The scene is a familiar one. Three Excel workbooks keep the company running: one for sales, one for inventory, one for cash. Only one person really knows how they are built. Each of them works — until the day you discover, while preparing the year-end close, that theoretical inventory and actual inventory no longer bear any resemblance to each other. This is not an accident: it is the sign that a great tool has been pushed beyond what it was made for.

What Excel does very well

Let us start by giving the spreadsheet its due. Excel is not a bad management tool: it is an excellent calculation tool, and many companies owe it their first years of serious management.

Its first strength is complete flexibility. No management software will let you model a margin simulation, a workload plan or a financing scenario in ten minutes. You add a column, you invent a formula, you test an assumption. No approval, no configuration, no development.

Its second strength is the cost of entry. The spreadsheet is already there, already paid for, already installed. For a company that is just starting out and issues a few invoices a month, setting up an ERP would objectively be out of proportion.

Its third strength is the learning curve. Almost everyone knows how to open a workbook, sort a column, add up a total. There is no training to schedule and no change management to organise.

These three qualities explain why Excel remains unassailable for one-off analysis — and why it will remain so even after an ERP is adopted. The problem only appears when it is given a role that is not its own: that of the company's official register.

The six limits that end up costing dearly

These limits are not flaws in Excel. They are the mechanical consequences of an individual calculation tool used as a shared management system.

1. Double, triple, quadruple data entry

The same customer order is re-entered in the sales file, in the inventory file, in the cash tracker, then sent on to the accountant. Every re-entry takes time and introduces the possibility of error. Above all, each copy ages independently of the others: after a few weeks they no longer tell the same story, and nobody knows which one is authoritative.

2. No reliable audit trail

In a workbook, an overwritten value disappears without leaving a trace. Who changed that price? When? Why? Nobody can answer. Files named `V2`, `V2_final`, `V2_final_corrected` are a clumsy attempt to reconstruct that history — and an additional source of confusion. Without an audit trail, it becomes impossible to explain a discrepancy after the fact.

3. No management of access rights

A shared file is an open file. The sales rep who looks up orders also sees the margins. The warehouse clerk who updates quantities can change purchase prices. And anyone can, with one unfortunate keyboard shortcut, wipe out an entire range. Sheet passwords are only a symbolic lock, not a security policy.

4. The fragility of formulas

This is the most insidious risk, because it is silent. A row inserted in the wrong place shifts a range. A copy-paste breaks a reference. A lookup returns an approximate value instead of an exact match. The file keeps displaying a result — a wrong result — and it propagates to every total that depends on it. A formula error does not announce itself: it gets discovered.

5. Inventory and invoicing out of sync

This is the most common breaking point. A sale is invoiced but inventory is not decremented, a goods receipt is recorded late, a return is never entered. Theoretical inventory drifts away from actual inventory, you sell what you no longer have and you tie up capital in what you thought had been sold. We detailed this mechanism in our article on the inventory management mistakes that undermine profitability. A spreadsheet, by design, cannot automatically link an invoice line to a stock movement.

6. Tunisian tax compliance becomes untenable

This is the hardest limit, because it is not negotiable. A compliant invoice requires continuous, tamper-proof numbering, the correct calculation of VAT, the fiscal stamp and withholding tax, and proper archiving. A payslip requires the exact application of the CNSS, IRPP and CSS scales, with brackets that change. And electronic invoicing run by the TTN now imposes a standardised, signed format transmitted to a platform: a spreadsheet simply cannot produce that.

The ten signals that show it is time to switch

Rather than a theoretical threshold, here are the concrete symptoms. If you tick three or four of them, the question deserves to be asked seriously. Beyond six, switching is no longer an improvement option: it is a protective measure.

1. You enter the same information more than once. The most universal signal. Every re-entry is time lost and a potential error.
2. Two people give you two different figures for the same question. Revenue for the month, a customer's outstanding balance, the stock of one product: if there are two answers, there is no longer a reference.
3. You discover stock discrepancies at the physical count. A stocktake should confirm what you already knew, not teach it to you.
4. Only one person "knows" how the files work. Their absence or departure paralyses the company. This is a major operational risk, and it is often underestimated.
5. You circulate files by e-mail or messaging app. As soon as a workbook travels as an attachment, several competing versions exist and nobody knows which one is authoritative.
6. You cannot say who owes you how much, today. If answering requires half a day of consolidation, collection is inevitably late.
7. The monthly close is an ordeal. When producing the month's figures takes several days of manual reconciliation, management becomes a rear-view mirror.
8. Your field teams work blind. Sales reps and delivery drivers have neither available stock nor the customer balance in front of them. They promise what they cannot deliver.
9. Tax obligations force you into workarounds. Numbering re-done by hand, withholding tax calculated separately, electronic invoices impossible to produce.
10. You put off decisions for lack of data. Opening a point of sale, hiring, discontinuing a product line: when the analysis is not available, you decide on instinct or you do not decide at all.

Note what this list does not contain: a headcount, a revenue figure, an invoice volume. It is not size that triggers the switch, but the number of people who have to share the same up-to-date information.

Excel vs ERP: the honest comparison

Comparing the two tools on every criterion would be dishonest: they are not aimed at the same use. The table below therefore compares what matters in order to run the management of a company — and Excel keeps its advantages there where they are real.

Criterion	Excel	ERP
Data entry	Repeated in every file concerned	Entered once, available everywhere
Data reliability	Depends on discipline and formulas	Built-in controls and consistency
Working as a team	Competing versions, editing conflicts	Single database, real time
Access rights	File open or closed, little nuance	Profiles and permissions per module
Inventory	Tracked manually, never linked to sales	Automatic movements, batches and expiry dates
Compliant invoicing	Layout possible, compliance not guaranteed	VAT, fiscal stamp, withholding, e-invoicing
Reporting	Rebuilt from scratch every time	Permanent dashboards
Mobility and field work	Poorly suited outside the office	Mobile apps, offline mode
Scalability	Degrades as the file grows	Modules and users added as needed
Cost	Zero to enter, high in hidden time	Visible subscription, less hidden time

On the two lines covering cost and flexibility, Excel keeps the advantage — and that is precisely why it will never disappear completely. On every line that concerns sharing and reliability, the gap becomes structural: these are not missing features, they are properties the spreadsheet was never meant to offer. If the notion of an ERP is still unclear to you, our guide "What is an ERP and why your SME needs one" lays the groundwork before going further.

What the switch changes in practice

Abstract arguments rarely convince. So let us compare the same day, before and after.

Before: a day in spreadsheet mode

A customer calls to place an order. The sales rep opens the inventory file — the one from last night — and quotes an availability he cannot guarantee. He writes the order in a notebook and will re-enter it in the evening. At midday, the owner looks for a customer's outstanding balance: he cross-checks the invoicing file with the bank statement, a quarter of an hour. Late in the afternoon, the warehouse clerk reports a stockout on a product the file still showed as available. The accountant, meanwhile, is asking for last month's documents, scattered across three folders.

After: the same day with an ERP

The sales rep checks actual inventory from his phone, commits an available quantity and creates the order on the spot. Stock is reserved immediately; the warehouse sees the line to be picked. The invoice is generated on delivery, with its numbering, its taxes and its compliant format. The customer balance updates by itself and the reminder is triggered on the due date. Inventory has been decremented with no intervention, and on his dashboard the owner reads the day's revenue, outstanding receivables and upcoming stockouts.

The difference is not that "it looks nicer". It is that the information was entered once, at the moment it came into being, and that it fed everyone who needed it on its own.

Getting the migration right in 6 steps

A failed switch almost never comes from the software. It comes from a rushed data migration or from a changeover that is too abrupt. Here is how to proceed.

1. Clean your data before migrating. This is the longest step and the most profitable one. Remove duplicate customers, standardise product descriptions, eliminate dead references. Migrating dirty data amounts to moving boxes of rubbish: you will pay for the transport and find the same mess on arrival.
2. Define your master data. Products, customers, suppliers: decide on a single naming and coding rule and apply it without exception. These are the foundations; everything you build afterwards will depend on them.
3. Carry over balances, not the whole history. You do not need to replay three years of invoices. Carry over the positions as at the switchover date: physical inventory counted, customer balances, supplier balances, bank balances. The history remains available in the old archived files.
4. Configure on your real cases. Taxes, document templates, units, warehouses, rights per profile: test the configuration with your real invoices and your real payslips, not with demonstration examples. Discrepancies always show up on the edge cases.
5. Train the teams before the switch, not after. Every user must have handled the tool on their own workstation before go-live. A well-targeted one-hour session per role is worth more than a three-hour general presentation.
6. Switch gently, with a short period of duplicate data entry. Choose a clean start date — the beginning of a month — and accept a duplicate-entry period of a few weeks at most, long enough to check that the figures match. Then stop the old files outright.

This sequence takes a few weeks of preparation, but it avoids the most destructive scenario: an ERP brought into service on wrong data, which nobody dares use a month later.

Mistakes to avoid

Four mistakes come up systematically, and all of them can be corrected upstream.

Migrating everything at once

Trying to start invoicing, inventory, purchasing, accounting and payroll at the same time multiplies the points of friction. Start with the most painful area — often invoicing and inventory — stabilise it, then add the following modules. A modular ERP is designed for exactly that kind of gradual rollout.

Keeping Excel running in parallel indefinitely

This is the most frequent mistake, and the most expensive. Keeping the old files "just in case" for six months brings back every problem the switch was meant to solve: duplicate data entry, diverging versions, permanent doubt about the source of truth. Duplicate entry is a temporary precaution, never an operating mode.

Neglecting the data migration

A rushed migration produces an ERP populated with duplicates and wrong balances. Users lose confidence in the very first week, and that confidence is very hard to win back. Devote to this step the time it deserves; it is what determines the credibility of the new tool.

Skipping training

"The tool is intuitive, they will manage" is a false economy. With no reference points, everyone invents their own method, the data degrades and the most reluctant quietly recreate their spreadsheet. Successful adoption is prepared: appoint an internal champion, train by role and provide support during the first weeks.

Once the decision has been made, the right tool still has to be chosen: our guide "Getting your ERP choice right: criteria, pitfalls and method" details the evaluation grid, and the comparison of ERPs in Tunisia helps you draw up a shortlist suited to the local context.

Frequently asked questions