Does an SME need an ERP?

The question usually arrives after a difficult year: the figures no longer match between departments, the monthly close takes three weeks, and somebody has said the word “ERP” in a meeting. It is a good question. In most cases the honest answer is not yet.

This article gives the four signals showing the moment has genuinely come, what to settle beforehand, and the alternatives that usually suffice. Headcount appears nowhere: it is not the number of employees that decides, it is the complexity of the flows.

What an ERP does, in one sentence

An ERP replaces several separate databases with a single one, shared between departments. Sales, purchasing, stock, production, invoicing and accounting stop being islands exchanging files: they read and write in the same place.

Everything else — the modules, the dashboards, the alerts — follows from that. And so do all the problems: if your processes are not stable, a single database freezes the mess rather than resolving it.

The four signals

1. The same data is entered several times

An order arrives by email, someone re-keys it into a spreadsheet, then into the invoicing software, then into the stock tracker. Every re-keying is a chance to make a mistake and time lost, and the total is rarely measured because it is spread across several people.

The test: follow one order from arrival to payment, and count how many times a piece of information is retyped. Beyond three, the signal is clear.

2. Two departments give two different figures

Sales announce one order total, accounting announces another, and nobody can settle it without half a day of reconciliation. This is not a discipline problem: it is the mechanical consequence of two databases that do not talk.

3. The monthly close takes longer than it should

If answering “where are we this month?” takes several days of work rather than a look at a screen, the information exists but is not usable. This is the most expensive signal, because it slows every decision, not only accounting.

4. One person knows how it all fits together

Someone maintains the central spreadsheet, knows the formulas, knows which file is authoritative. The day that person takes leave, everything slows; the day they leave the company, part of the knowledge goes with them. That signal is not technical but organisational — and it is the most urgent of the four.

What to settle first

An ERP does not create order: it applies the order you describe to it. If your processes vary with whoever performs them, the configuration will freeze them in an arbitrary state, and you will discover the mess at the point where getting out of it is expensive.

Before opening a selection process, write down how the three or four main flows actually run — not ideally: order to delivery, purchase to payment, quote to invoice. That work is worthwhile even if you then drop the ERP: it almost always reveals two or three free fixes. Our article on digitising an SME describes that order of march.

The alternatives, and their limit

In many situations a few specialised tools wired together cover the need for a fraction of the cost and delay: invoicing software, a sales tool, stock tracking, connected by automatic exchanges.

The approach has two real advantages — you start within weeks, and you can replace one tool without questioning everything — and one limit that eventually surfaces: every connection is a breaking point, and their number grows faster than the number of tools. Beyond five or six connected systems, maintaining the connections costs more than the single database you were avoiding.

Compare on total cost over three years — licences, connections, maintenance, internal time — not on licence price. That is the only frame in which the comparison means anything.

The real cost of an ERP

Item	What it covers	How often forgotten
Licence or subscription	The only item present in every proposal	Never forgotten
Configuration	Translating your processes into the tool	Often underestimated
Data migration	Cleaning and importing what exists, duplicates included	Very often underestimated
Training	Every department, several times, and new joiners afterwards	Often forgotten
Internal time	Your teams in workshops, in testing, in double entry during cutover	Almost always forgotten

The last line is the one that tips projects over. It appears on no invoice, which is precisely why it is not budgeted.

Being least wrong

Three precautions are worth more than a long comparison of vendors.

• Start on a reduced scope. One flow, one department, one cutover. The rest follows once the first works — and you will have learned what you did not know before starting.
• Require a full export of your data in a standard format, tested before signing rather than promised in a brochure. That is your exit, and its absence turns a choice into a permanent commitment.
• Check who can work on it. A tool only one integrator in the world can configure ties you to that integrator, whatever the quality of the software.

The right moment is not when you reach a size, it is when re-keying and diverging figures cost you more than the project itself.

If the answer is “not yet”

That is not a negative conclusion. The two or three years before an ERP are exactly the time to stabilise processes, remove the obvious re-keying and gather the knowledge one person holds. That work is not wasted: it later reduces configuration, which is the most expensive item.

Our IT Advisory & Digital Transformation and Web & Mobile Development pages cover scoping and, where no off-the-shelf product fits, custom tools.