Build vs Buy: A decision framework for custom software

Quick answer: there are four genuine reasons to build custom software — your process is your moat, integration cost exceeds build cost, data sensitivity demands sovereignty, or long-term economics favour ownership. If none of these apply, default to SaaS or low-code. Most businesses are better off buying than building.

The build-vs-buy decision is one of the most expensive ones a growing business will make — not because either path is wrong, but because the wrong path for your situation gets expensive fast.

The four genuine reasons to build custom

Most of the time, SaaS is the right answer. We say that as a studio that builds custom web applications for a living. But there are four situations where custom is genuinely the smarter call:

1. Your process is your moat. When the way you do business is itself a competitive advantage, forcing it into a SaaS template erodes the moat.
2. The integration cost exceeds the build cost. Sometimes wiring 6 SaaS tools together via custom API integrations costs more than building one custom tool that does the whole flow natively.
3. Data sensitivity / sovereignty. Some data shouldn’t leave your control, full stop.
4. Long-term economics. SaaS-per-seat costs at 200 staff exceed the cost of owning the software outright.

If none of these apply, default to SaaS.

When low-code wins

Low-code tools — Make, n8n, Zapier, Airtable, Notion, Bubble — genuinely shine for:

• Internal workflows that change frequently
• Prototypes you’d be willing to throw away
• Departmental tools where the team owns the system

They start to creak when:

• The system becomes business-critical
• Performance under load matters
• Multiple teams need different views

We’ve migrated more than one team off a fragile Zapier or Make chain onto proper code-built integrations — the visible cost of code is usually less than the hidden cost of low-code at scale.

A scoring framework

For each candidate path (SaaS / Low-code / Custom), score 1–5 across:

• Fit to process — does it match what you actually do?
• Total cost over 5 years — including switching costs
• Data control — who holds your data?
• Replaceability — can you move off it without pain?
• Speed to value — when do you get the benefit?

The highest aggregate usually wins. The highest on any single dimension rarely should.

What to do next

If you’re seriously evaluating a custom build, the questions to answer first are:

1. What problem are we actually solving — in concrete terms?
2. What’s the cost of not solving it for another 12 months?
3. What does success look like in measurable terms?
4. Have we genuinely tested the SaaS / low-code paths first?

If you can’t answer these, you’re not ready to commission custom software. That’s not a criticism — it’s a save.

When you are ready, start a project and we’ll tell you honestly whether custom is the right answer for your situation. We turn down work that isn’t a fit.

Common questions

When should I build custom software vs buy SaaS? Build custom when one of four conditions applies: your process is a competitive moat, integration cost across multiple SaaS tools exceeds the build cost, data sensitivity requires you to control where it lives, or long-term per-seat economics make ownership cheaper. Otherwise, default to SaaS — it’s the right answer most of the time.

Is custom software cheaper than SaaS? At small scale (under 20 staff), almost never. At larger scale (50+ staff with process complexity), often substantially cheaper over five years once licences, integration cost, internal staff time, and migration cost are all counted. We’ve covered the full TCO comparison in detail.

What is low-code and when does it work? Low-code platforms like Make, n8n, Zapier, Airtable, and Bubble let teams wire up workflows without writing code. They genuinely shine for internal workflows that change frequently, throwaway prototypes, and team-owned departmental tools. They struggle when systems become business-critical or volume scales up.

How long does custom software take to build? For a focused first phase: typically 6–16 weeks depending on complexity. Larger systems are usually phased rather than built monolithically — phase one ships in 8–12 weeks, with subsequent phases informed by what was learned. The build-vs-buy decision should always include a realistic timeline.

What does custom software cost? A serious mid-size custom build typically costs $80,000–$300,000 to deliver, plus ongoing maintenance at 10–25% of build cost annually. The crossover point where custom beats SaaS economically depends on headcount, process specificity, and time horizon — usually somewhere between 50 and 200 staff for the systems your business genuinely runs on.

Should I trust the agency that says I need custom? Not by default. A serious agency will tell you when SaaS is the right answer — even when that means turning down a build. We turn down custom work where SaaS or low-code would serve the client better. If the agency only ever recommends building, the recommendation isn’t neutral.