Build or buy? Compare three years, not the first invoice
Custom software looks expensive up front and subscriptions look cheap. Over three years the picture often reverses, in both directions.
The build-or-buy decision is usually made on the first number anyone sees: the development quote or the monthly subscription price. Both are misleading on their own. A fair comparison looks at what each option costs and delivers over about three years.
What buying really costs
- Licences that grow with you. Per-user pricing that is cheap for 10 people can be significant at 50.
- Setup and migration. Configuring the tool, moving your data and training staff often costs as much as the first year's licences.
- Workarounds. When the tool doesn't match how you work, people build spreadsheets around it. That time is a real cost even if nobody invoices it.
- Price changes and lock-in. Check what happens to your data and your price if you leave, or if the vendor changes its plans.
What building really costs
- The first version. This is the number in the quote, and it is rarely the largest cost.
- Maintenance. Plan on 15 to 25 percent of the initial build cost every year for fixes, security updates and small changes.
- Hosting and monitoring. Usually modest, but someone has to own it.
- Key-person risk. If one developer understands the system, you depend on that person. Insist on documentation and a proper handover.
When building tends to win
Build when the process is a genuine competitive advantage, when no product fits without heavy workarounds, or when you need deep integration between systems you already own. Build small: the first version should solve the core problem and nothing else.
When buying tends to win
Buy when the process is standard for your industry, such as accounting, payroll, CRM or ticketing. Mature products have already solved problems you haven't hit yet. Your money is better spent adapting your process slightly than rebuilding what already exists.
A simple way to decide
Put both options in one table with three columns: year one, year two, year three. Include licences, setup, maintenance, internal time and the cost of workarounds. Then add one row for risk: what happens if the vendor disappears, or if your developer does. The answer is usually clearer than the first quote suggested.