Open-source business software gets recommended for the wrong reason — that it is free — and rejected for the wrong reason — that it is risky. Both misunderstandings lead to expensive decisions.

What "free" actually means

The licence is free. The software is not. You still pay for hosting, setup, configuration, updates, backups, security patching and somebody who understands it well enough to fix it at an inconvenient moment.

For a business of any size, those costs are real and ongoing. What changes is the shape of the cost: mostly one-off and internal rather than a per-user fee that grows every time you hire someone.

When it is genuinely the right answer

  • Per-user pricing is becoming the dominant cost, because you have a lot of occasional users — common in retail, hospitality and education.
  • You need to modify how the software works, not just configure it. Commercial SaaS will usually say no; open source cannot.
  • Where the data lives matters to you — for regulatory reasons, or because you have been locked in before and would rather not repeat it.
  • You are planning to run this for many years, which is when a large one-off cost beats a growing recurring one.
  • A mature project already exists that does most of what you need. This is the important one.

When it is the wrong answer

  • You need it working in three weeks. Configuration and migration take longer than a signup.
  • Nobody, internally or externally, will own it. Unowned open source becomes an unpatched liability within a year.
  • The commercial alternative is genuinely cheap for your size and does the job. Independence is not worth much if it costs more than the lock-in.
  • You are choosing it primarily to avoid a subscription. That is a reason to negotiate, not a reason to change architecture.
  • The project is barely maintained. A quiet repository is a much bigger risk than a licence fee.
Open source is a good answer to "we need control" and a bad answer to "we need to spend less this quarter".

How we assess a project

Before recommending anything open source, we look at how recently it was updated, how many people maintain it rather than one person, whether there is a real upgrade path between versions, whether commercial support exists if you want it, and how painful it would be to get your data out again. That last question is the one people skip, and it is the one that decides how trapped you are in three years.