Build vs Buy: When a Growing SMB Should Build Custom Software

The build-or-buy question usually arrives at a growing business in a bad mood. Something broke: the tool everyone tolerated finally lost a customer's order, or the third pricing email of the year landed, or someone counted the spreadsheets. And the question gets asked the way software questions get asked, as a binary. Should we buy something off the shelf, or should we build our own?

Read the first page of advice on this and you will notice a pattern: everyone answering has a side. Development shops conclude, after careful analysis, that you should build. Software vendors conclude, with equal care, that you should buy. We build custom software for a living, so you know our seat in that argument, which is exactly why this article starts with the cases where you should not hire anyone like us, and why our prices and our disqualifying criteria are written down where you can hold us to them.

The honest answer to the binary is that the binary is wrong. Almost no growing SMB should build most of its software, and almost none should buy all of it. The useful question is where the line runs through your particular business.

Buy everything where your business is the same as everyone else's

Accounting, payroll, email, calendars, document storage, video calls: for these, your business is not special, and that is a compliment. These functions are commodities, the off-the-shelf products are polished by every other business that uses them, and every hour spent customising them is an hour of your differentiation budget spent on sameness. Nobody should build their own accounting package, including us, and a partner who lets you consider it is billing you for the consideration.

The same logic covers more than the obvious commodities. If a stock tool does ninety-odd percent of what you need and the missing part is a preference rather than a profit driver, buy it and adapt. Changing a habit is cheaper than maintaining software, and the discipline of asking "is our way actually better, or just ours" kills a lot of expensive projects that deserved to die. When the tool fits, the honest advice is the boring advice: buy it, configure it, move on.

Consider building where your process is the reason customers pay you

The signal that the line has been crossed is rarely dramatic. It looks like workarounds. Someone keeps a private spreadsheet that patches what the system cannot do. An experienced employee retypes data between two tools that refuse to talk to each other. A step exists because the software demands it, not because the business does, and everyone has stopped noticing. Each patch is small; together they are your operations team quietly staffing the gap between how the tool thinks and how your business actually works.

That gap is worth paying attention to when it sits close to the money. If the awkward remainder is precisely the thing your customers choose you for, the way you handle rush orders, the way your quotes work, the pipeline that gets your product out a day faster than the competition, then bending it to fit a generic tool erodes the difference you are selling. Off-the-shelf software optimises for the average customer of the vendor. Your edge, by definition, is not average.

The meter matters too. Subscription pricing follows your growth: more seats, more documents, more tasks, more money, forever. We did the arithmetic on both versions of that treadmill separately, in what another SaaS seat really costs and in what it costs to automate a manual process, and the short version is that at low volume the subscription wins honestly, while at growing volume the same meter becomes the argument for owning the thing instead. And some processes settle the question by themselves: if the workflow touches scanners on your local network, machines on your floor, or data that has to stay on your own server, the cloud products were never really candidates.

The position most SMBs should actually take

Put the two halves together and the sensible posture is hybrid: buy the commodity layer, and build the connective tissue. Keep the accounting package, the email, the stock tools that fit. Build the small, specific things that make them behave like one system shaped like your business: the integration that moves data so nobody retypes it, the tool that handles your weird-but-profitable edge cases, the dashboard that answers the question your team currently answers by exporting three reports. Connective tissue is small by nature. It is also, in our experience, where most of the payback lives, because it sits exactly where the spreadsheets and the retyping were.

This is what we build, so the claim comes with our numbers attached rather than a discovery call: most of these projects land between €8,000 and €30,000, fixed, typically built in 4 to 8 weeks, starting from €3,500 for the smallest useful version. One price, you own the code, no subscription, no per-user fees.

If you do build, build like a coward

The horror stories about custom software are mostly stories about scope and ownership, and both risks respond well to cowardice. Scope the first build small: one process, one bounded problem, finished in weeks, generating its verdict before you commit more. A partner who proposes a modest first project over the grand rebuild is showing you their incentives; a partner who needs six months before anything works is asking you to fund faith.

Insist on a fixed price, because a bounded problem can be quoted, and a partner unwilling to quote it is telling you the scope is not actually understood. Get the ownership in writing before the first line is written: the copyright assigned, the repository in your name, the build runnable by someone who did not write it. We covered exactly what that means and why a paid invoice is not enough in if your developer disappears tomorrow, what do you actually own. And prefer boring technology, because every exotic choice narrows the set of people who can maintain your system later.

For what it is worth, here is our own disqualifying checklist, the one that ends some of our sales calls early: if your volume is small, your process is standard, and a decent off-the-shelf tool exists, we will say so and point you at the tool, because a build that saves less than it costs is a bad product no matter who sells it. The fastest way to check which side of the line you are on is arithmetic, not opinion: your volume, your hourly cost of the manual work, and the prices involved. Our calculator does that multiplication in a few minutes, and if the manual work in question is currently a job ad, price the software option first.

If the arithmetic comes out interesting, the next step costs nothing: a free consultation where we tell you which side of the buy-build line your process sits on, even when the answer is "buy". If you want the full plan and a fixed quote in writing before committing to anything, that is the €1,200 audit described on the custom software page, credited in full toward the build if you go ahead.