Build or buy: when off-the-shelf software is enough, and when it isn't

When a small or mid-sized company struggles with its software, the technology is the first suspect. Too slow, too old, too complicated. So the company evaluates, migrates and replaces — and a year later it's frustrated by the same things, just with a new logo.

The question usually asked is “Is this software good?” The more important question is “Does this software fit this company?” Those are two very different questions. An off-the-shelf product is built for an imaginary average business. Your company isn't average. It has its own processes, its own exceptions and its own reasons for doing things a certain way.

Buying is right when your problem is a standard problem. That's true in more areas than software developers like to admit: accounting, payroll, email and calendars, a common CRM for a classic sales process. In these areas, processes are largely the same across companies, often defined by law, and the available products are mature.

A good test: does your company differ from its competitors in this area? If not — if your accounting simply has to be correct and no customer chooses you because of it — then buy. Custom software for a standard problem is wasted money, and we say so when we're asked.

It's different where your company does things its own way. A few signs that an off-the-shelf product doesn't fit:

Employees keep lists outside the system because important cases don't exist in it. Data gets copied between several tools by hand. You pay licenses for a large system and use a small part of it, while the part you actually need is missing. Every change to a process requires an outside consultant, or isn't possible at all. Or the process that doesn't fit the tool is exactly the one that makes you money.

That last point is the decisive one. If the way you work is your edge over the competition, an off-the-shelf product forces you to give up exactly that edge.

For a long time, the answer was still often “buy,” for a simple reason: custom software cost several times as much and took months or years. That has shifted. AI-assisted development significantly reduces the effort of building your own software, as long as experienced people are steering it. What used to be a major project is often manageable today. That's why the build-versus-buy debate is back, after years in which a SaaS product was considered the answer to every problem.

In practice, the answer is rarely either-or. Usually it's a mix: standard products for the standard areas, custom software for the core, and clean interfaces between them so nobody has to copy data by hand.

The point isn't to build everything yourself. The point is to make fit a deliberate decision instead of leaving it to whatever the software market happens to offer.