Build vs buy: when custom software is the cheaper option
Key takeaways
- Compare the total cost over three years: licences, add-ons, setup, maintenance and the hours spent on workarounds.
- Per-seat licences grow with headcount, while custom software costs mostly upfront.
- In our 40-user worked example, the custom build pays back around month 19.
- Buy for solved problems like payroll and accounting; build where your process is your competitive edge.
Buy off-the-shelf software when your process is close to how everyone else works. Build custom software when the process is how you win, when per-seat licences grow faster than your revenue, or when your team spends hours a week working around the tool. Compare total cost over three years, not the first invoice.
Why the first quote misleads
A SaaS subscription looks cheap in month one. A custom build looks expensive in month one. Both impressions are wrong, because neither includes what happens next.
Software you buy keeps charging per seat, per module and per year, and the price usually rises. Software you build has a large upfront cost, then a smaller ongoing one for hosting, maintenance and changes. The question is where those two lines cross.
How to compare the real cost over three years
Here's a worked example with round numbers. Swap in your own.
| Cost over 36 months | Buy (SaaS) | Build (custom) |
|---|---|---|
| Licences: 40 users × $60/month | $86,400 | none |
| Add-on modules and integrations | $14,400 | none |
| Setup or build | $8,000 | $70,000 |
| Hosting, maintenance and changes | included | $21,600 |
| Workarounds: 10 hrs/week × $25 | $39,000 | none |
| Total | $147,800 | $91,600 |
Two lines decide it. Licences scale with headcount, so a growing team pays more every year for the same tool. And workarounds (the copy-pasting, the shadow spreadsheet, the "export, fix, re-import" routine) are real costs that never appear on an invoice.
In this example the custom build pays back around month 19. With ten users instead of forty, buying wins easily. That's the point: the answer depends on your numbers, and it can be calculated.
Signs custom software will be cheaper
- Your process is your edge. How you price, quote, schedule or deliver is the reason customers pick you, and the software forces you to do it the generic way.
- Licence costs track headcount. Every new hire adds a seat, and the tool's value per seat isn't rising.
- People are working around it. If there's a spreadsheet that "fixes" what the system can't do, you're already paying for custom software, just badly.
- You're stitching five tools together. Each one is fine. The integrations between them are where errors and hours go.
A real example: quotes in four hours instead of five days
Forge & Field Manufacturing in Pune made industrial components to order. Engineers priced every quote by hand in spreadsheets, and a quote could take up to a week. No off-the-shelf quoting tool understood their materials, tolerances and finishing options.
We built a quoting engine with a product configurator, connected to their ERP and CRM. Quotes now go out in about four hours, and faster responses lifted the quote win rate by 31%. The software is theirs: the code, the data and the roadmap.
When to buy, without second thoughts
Payroll, accounting, email, video calls and standard HR. These are solved problems with good products. Building your own is almost never worth it, however unusual your company feels.
The middle path most businesses miss
You don't have to choose all or nothing. Often the cheapest option is to keep your off-the-shelf systems and build a thin custom layer on top: a portal, an integration, an automation that removes the workaround. Kettle & Crane Coffee kept Shopify and Xero and added a wholesale ordering portal plus automated syncing. It saved 26 hours a week without replacing anything.
If you want help running these numbers for your own systems, that's the first step of our custom software development work. Book a free consultation and bring your current software bills.
