- Buy by default, especially for work every company does the same way, like payroll or email.
- Consider building when the work is how you win customers, or when your process has real rules that products can't handle.
- Compare costs over several years, including workarounds, not just license fees against a build quote.
- The best answer is often both: keep what you bought and build the gap around it.
- Start small. A first release that solves one painful problem tells you more than a long comparison.
Start by assuming you should buy
Payroll, email, accounting and basic CRM work about the same everywhere. Good products already exist, they keep improving and someone else handles the updates. Building your own version rarely makes sense.
Honest vendors and honest developers should both tell you this. If a developer never suggests an existing product, be careful.
When custom software earns its place
Building becomes worth considering when one or more of these is true:
- The work sets you apart. If the way you quote, serve or deliver is why customers choose you, a generic product may flatten what makes you different.
- Your rules are too specific. Your team spends real time working around the product, and the exceptions are the job.
- The work crosses several tools. Each product is fine, but nothing owns the steps between them.
- Per-seat pricing stops making sense. You need many people to use a tool lightly, and the bill keeps growing.
- You've already built it in a spreadsheet. A critical spreadsheet is a homemade system with no permissions, history or backup plan.
Compare the real costs
People usually compare a license fee against a build quote. That leaves out most of the cost on both sides. List these over at least three years:
- Buying: licenses, setup, training, connecting it to other tools, price increases and the hours your team spends on workarounds.
- Building: the first build, hosting, maintenance, future changes and what happens if the person who built it leaves.
Workaround hours are the number most teams never count. Ask the people doing the work to estimate them, then multiply by what those hours cost you.
What most guides miss
The question is usually framed as one or the other, but the strongest setups do both. Keep your CRM and accounting system, then build the small piece that connects them or handles the part they do badly.
The other thing that gets missed is ownership. Before you build, ask who owns the code, where the data lives and how someone else could take it over. If those answers are vague, the risk you were trying to avoid has only moved.
AI has also changed the math. Custom software can now be built faster than it used to be, which makes smaller projects worth doing. It hasn't removed the need to understand the process, test the result and maintain it after launch.
A simple way to decide
- Write down the one problem you most need solved, in plain words.
- Try the two best existing products on your real work, including the awkward cases.
- Note where each one breaks and what the workaround costs.
- If the gaps are small, buy. If the gaps are the job, get a scoped estimate for a first release.
How we do it
We start by following the work, and we'll tell you when an existing tool is the better answer. When a build makes sense, we build around what you already use, and you own the code. See custom software or start with a process audit.
