Every growing company reaches this moment: the spreadsheet can no longer cope, and you have to choose between buying an off-the-shelf system or having one built to measure. The wrong choice is expensive on both sides — off-the-shelf software that does not fit becomes a crooked process forever, and an unnecessary custom system becomes an expensive project that reinvents what already existed for a fraction of the price.
The good news is that this decision is not a matter of taste. There is a criterion. And it starts with a question almost nobody asks.
01The question that decides almost everything
Is your process different because it is a competitive advantage — or just because nobody has ever reviewed it?
That distinction is the heart of the decision. If the way your company does that thing is what makes it better than the competition — how it prices, routes, serves, produces — then bending that process to fit generic software means giving up your own advantage. That is where custom is justified.
But if the process is different by accident — because that is how someone set it up eight years ago and nobody touched it — then the difference is not an asset. It is debt. And paying to automate debt is the most expensive way to keep it.
02When off-the-shelf is the right choice
- The process is a market standard: accounting, payroll, tax issuance, email marketing.
- There is regulation, and staying compliant alone would be a permanent project.
- The volume is still small and the business is being validated — what you need is to start, not to build.
- A mature tool already exists, with a community, support and continuous evolution paid for by thousands of customers.
- The licence cost is lower than the cost of keeping a team (or a supplier) maintaining your own.
It evolves without you paying for it. Every security fix, every change in legislation, every new feature is shared across all customers. With custom software, that bill is yours — in full, and forever.
03When custom is the right choice
- The process is the differentiator of the business, and adapting it to a generic system would destroy what makes the company win.
- No tool on the market covers the flow, and the company already lives on workarounds: export, paste, reimport, check by hand.
- The integration between the current systems is where the operation bleeds — and that is exactly where off-the-shelf does not reach.
- The cost of licences, multiplied per user and per year, has already passed the cost of building.
- The data is the asset, and keeping it hostage to a supplier is a risk the company does not want to take.
04The most common mistake: comparing purchase price with build price
The calculation most people make is between the annual licence and the project budget. It is the wrong calculation, because it ignores half the costs on both sides.
On the off-the-shelf side, add: the customization it will demand (and it almost always does), the implementation consulting, the data migration, the training, the per-user cost as the team grows, and the price of leaving — because one day you will want to leave.
On the custom side, add: the project, the infrastructure, the maintenance after delivery and the evolution over the years. Deployment is not the end: it is where most software projects are abandoned, and precisely where the real cost shows up.
It is in the third year that the two curves cross — and it is where they cross, not the initial price, that answers the question.
05The five questions before signing anything
- Is this process a competitive advantage, or just a habit? If it is a habit, review the process before buying software.
- How many hours a month does the team spend today working around the absence of this system? If you do not know, measure before deciding.
- If I want to change supplier three years from now, what exactly do I take with me — and what stays behind?
- Who owns the data, and in what format can I get it out?
- What happens if the supplier disappears, is acquired, or changes the price?
The last three are about lock-in, and they are the ones nobody asks in the middle of the excitement. They are also the ones that hurt most later.
06There is a middle path, and it is underrated
In practice, the answer is rarely “all off-the-shelf” or “all custom”. The most efficient arrangement is usually: buy what is commodity, build what is differentiating, and integrate the two.
Nobody should build their own payroll system. And almost nobody should outsource the software that runs the heart of the operation. What sits between the two — the integration — is usually where the biggest gain is, and it is what fewest people see as a project.
07How to start without betting the company
If the answer points to custom, start small. An MVP — the minimum slice of the system that already solves a real problem — usually takes 4 to 16 weeks. It exists to answer the most important question before the investment grows: does this actually solve it?
It is the same method we apply to our own products: build the core, put it in use, learn, and only then expand. A project that only delivers value at the end is a project that bets the company on an assumption.
And if, midway through the assessment, it becomes clear that an off-the-shelf tool solves your case — that is what we are going to tell you. A supplier who can only recommend what they sell is not helping you decide.
Next step
Want to discuss this applied to your case?
The first conversation and the preliminary assessment cost nothing — and if the answer is that you do not need to hire anything, that is what we will tell you.