Buy when the process is standard for your industry and a product covers it well. Build when the process is your competitive advantage, when no product covers the requirement, or when licence plus workaround costs exceed the build over three years. Most companies should buy, and integration is usually the real need rather than a new system.
We build custom software, and we talk several enquiries a year out of it. A $200-a-month tool that fits beats a $40,000 build that fits slightly better.
The evaluation window
Often the real need
Not just build cost
The right default
Per-seat pricing compounds. 50 users at $40/month is $24,000 a year — which changes the comparison.
Hours spent monthly working around a tool that nearly fits. Rarely measured, often the largest number.
$15,000–$75,000 typically, plus 15–25% annually for maintenance and enhancement.
Vendor lock-in, data portability and the cost of migrating if the product changes direction.
Buying is weeks. Building is months. For an urgent problem this often decides it alone.
$1,500–$8,000 to connect existing tools — frequently the actual solution to a 'we need custom software' brief.
| Build when | Buy when |
|---|---|
| The process is your competitive advantage | The process is standard for your industry |
| No product covers the requirement | Several products cover it well |
| Licence plus workarounds exceed build cost | Subscription is cheaper than building |
| You must own the data and rules | Vendor lock-in is acceptable |
| Integration between systems is the need | One tool would replace several |
| Requirements are stable and understood | Requirements are still moving |
| Buy | Build | |
|---|---|---|
| Year 1 | Licences + setup + training | Build cost + deployment |
| Years 2–3 | Licences + seat growth | Maintenance at 15–25% of build |
| Workaround time | Often significant | Should be near zero |
| Integration work | $1,500 – $8,000 each | Built in |
| Switching risk | Real, and rising | None |
| Typical 3-year total | $20,000 – $90,000 | $25,000 – $110,000 |
The comparison is closer than most vendors on either side admit. What usually tips it is workaround time — the hours your team spends every month compensating for a tool that nearly fits.
Buy the tools and build only the integration and the specific gap — usually a fraction of the cost of either extreme.
A large share of enquiries that arrive as 'we need custom software' are actually this. It is cheaper, faster and lower risk, and we recommend it regularly.
Run a genuine trial with real data and the people who will use it, not a sales demo. Most 'it cannot do X' conclusions come from a demo rather than a trial, and a surprising number turn out to be configuration.
Yes — budget 15–25% of build cost annually. Bought software has that in the licence fee. The difference is that you control the roadmap rather than a vendor deciding what to build next.
That is a good problem and a normal path: buy now, build later when the requirement is proven and understood. Building first for a future you have not reached is how budgets get wasted.
Yes, and we do several times a year. We would rather do a $6,000 integration that works than a $40,000 build you did not need — the first one leads to more work, the second leads to a bad reference.
How we think about this work, in more depth.
Comparison
A 30-minute call, then a written proposal with scope, price and timeline within two to three working days. No retainer required to get a real number, and no obligation if the answer is that we are not the right fit.