Off-the-shelf vs custom ERP: a selection guide for SMEs

Every SME that decides to change its management software faces the same question: is it better to go with an off-the-shelf product, already battle-tested by thousands of companies, or a system built to fit your own processes? The answer seems simple, but in practice almost nobody approaches it the right way. The reason is always the same: people start from the software instead of the problem.

This article is not a feature comparison. It is a line of reasoning — the same one we follow when a client asks for help — on how to reach a decision that holds up over time, without wasting months and budget on something nobody ends up using.

The real starting point: what is broken today

Before looking at any software, you need an honest answer to a precise question: which two or three processes cause the most problems today? Not in theory — in the actual working day. Where is time wasted? Where do data errors happen? Where is work duplicated?

In almost every SME we have worked with, the answer involves the same points: orders managed half by email and half in spreadsheets, stock that does not add up, invoicing that requires double entry, customer information scattered across three different places. The current ERP — or the lack of one — is not the cause: it is the symptom. The cause is that the workflow was never designed explicitly.

This step sounds trivial, but it is the most important one in the entire journey. If you do not know what must concretely improve in the next ninety days, any software — off-the-shelf or custom — risks becoming a cost with no return.

When off-the-shelf is the right choice

A ready-made product has one enormous advantage: someone else has already solved thousands of problems before you. The interface is tested, regulatory updates arrive automatically, support covers common scenarios. For a company whose process is relatively standard — buy, sell, invoice, manage stock — an off-the-shelf product covers about eighty percent of the need through configuration alone.

The trouble starts with the remaining twenty percent. That twenty percent is almost always where your company's specific way of working lives: the pricing rule that differs for a long-standing client, the approval flow that goes through the owner only above a certain amount, the report the salesperson needs in a particular format. With off-the-shelf software, that twenty percent is handled through workarounds, expensive and fragile vendor customisations, or — most commonly — Excel on the side.

If that twenty percent is not critical to your business, off-the-shelf works very well. But if that twenty percent is precisely what sets you apart from competitors or what eats most of your time, then you are buying a suit you will have to alter continuously.

When custom makes sense

Custom software makes sense in one specific case: when the process you want to manage is your competitive advantage, or it is so specific that no market product covers it without forcing it. It is not a matter of company size — some fifteen-person SMEs have more complex flows than hundred-person firms.

Here is a concrete example. A company that produces and sells chemical reference standards has a catalogue of thousands of items, each with certificates of analysis, expiry dates, substance compatibility rules and different storage conditions. The quoting process is not "find product, add to cart": it is a technical negotiation that requires checking availability, suggesting compatible substitutes, and respecting regulatory constraints. No off-the-shelf ERP handles this flow natively. A custom system, built around that process, eliminates hours of manual work per day and reduces errors that in that sector can have serious consequences.

The risk of custom is the opposite of off-the-shelf: it is not that features are missing, but that the project can grow without control. Every department wants its feature, every user has an idea, and if nobody says "this does not enter the first release", the project never finishes. That is why the concept of MVP — minimum product that works and can be measured — is essential: not to do things halfway, but to do them one at a time and verify they actually help.

The problem nobody wants to face: the data

Whether you choose off-the-shelf or custom, there is one step that everyone underestimates and that alone can sink the project: data migration. Customer records full of duplicates, inconsistent product codes, prices that live in a manually updated spreadsheet maintained by one person who "knows how it works". No software, however well built, produces good results from bad data.

Data clean-up is not IT work. It is a matter of business truth. It means sitting down with the people who know the clients, the products, the operating rules, and deciding: is this record correct? Is this field needed? Do we merge this duplicate or delete it? It is tedious, it takes time, and there is no shortcut. But it is the difference between a system that delivers reliable information from day one and one that perpetuates the same errors as the old spreadsheet — just with a nicer interface.

Excel: not the enemy, the signal

In almost every company we have worked with, alongside the official ERP there is a parallel ecosystem of Excel sheets. The "real" price list kept by the sales rep, the production order tracker, the file with deadlines managed by the secretary. These sheets are not a laziness problem: they are the symptom that the official system does not cover a real need. Every parallel spreadsheet is a spontaneous answer to a gap.

The new system — off-the-shelf or custom — must absorb those needs, otherwise the parallel sheets will reappear within a month. That is why, before choosing any tool, it is worth doing an inventory: which Excel files exist in the company, who uses them, what they serve, how often. That inventory says more than any software brochure about what is actually needed.

Adoption is everything: unused software is waste

The most critical moment is not choosing the software, nor building it. It is the month after go-live. The system is there, it works, it has been paid for — but people keep doing things the old way. It is a human problem, not a technical one, and there is only one way to solve it: start small.

Pick one department, one flow, a small group of users. Train those people on their real cases — not on generic scenarios, not with two-hundred-page manuals, but on "Monday morning client X's order arrives, here is how you handle it now". Appoint an internal champion, someone the team already respects — not necessarily the most technical person, but the most credible. That champion becomes the contact point for the first weeks: answering questions, collecting issues, escalating what needs fixing.

Measure two things: how long the process takes compared to before, and how often people revert to the old method. If after thirty days the time has not improved or people avoid the system, adding features is not the answer: understanding where the flow is uncomfortable and fixing it is. Often it is a detail — an extra step, a mandatory field that should not be, a piece of information not at the user's fingertips.

The vendor: partner or dependency?

Whether it is an off-the-shelf product with a local reseller or a custom project with a development agency, the vendor relationship is the invisible factor that determines medium-term success. The question to ask is not "how much does it cost" but "in a year, will I be more autonomous or more dependent?"

A good vendor leaves you owning your data, documents the configurations, teaches you to handle simple changes yourself. A bad vendor creates dependency: every small change goes through them, data sits in a proprietary format, documentation does not exist. This applies to off-the-shelf (where lock-in is in data formats and customisations) and to custom (where lock-in is in code and system knowledge).

How to decide: a practical framework

After years of projects, we have found that the choice becomes clear by answering five questions:

1. Is your critical process standard (buy-sell-invoice) or specific to your sector?
2. How many parallel Excel sheets do you have, and how vital are they for daily operations?
3. Do you have an internal person who can lead adoption for at least three months?
4. Are you willing to start with a reduced scope and expand only after seeing the numbers?
5. Does the vendor give you data in an open format and document everything?

If the answers to 1 are "standard" and to 5 "yes", off-the-shelf is probably the better choice. If the process is specific, parallel sheets are many and critical, and you have someone internal who can follow the project, custom will deliver more value over time — provided you keep the scope under control.

What really matters

The choice between off-the-shelf and custom is not a technology choice. It is a process and organisation choice. The best software is the one your team uses every day without going back to Excel, that gives you reliable information for decisions, and that a year from now you can evolve without starting over.

Start from the process, not the catalogue. Clean the data before migrating. Begin small and measure. Choose a vendor that makes you more autonomous. The rest — cloud or on-premise, React or SAP, monthly or perpetual licence — comes after, and it comes much more easily when the underlying reasoning is sound.

If you are evaluating a change of management system and want an outside perspective on which path makes more sense for your specific case, we can sit down together and look at the numbers. No obligation, no feature checklist: just an honest conversation about how to work better.