Internal Tools: When Building Beats Buying

The question behind the question

“Build or buy” is really “is this process ordinary or is it ours?” Payroll is ordinary. Your specific way of scheduling engineers across regions might not be.

Buying an ordinary process is nearly always right. Buying something that encodes your differentiator means bending your advantage to fit someone else's model, which is a slow, expensive way to become average.

The real cost of building

• Build cost, which is the number everyone estimates
• Maintenance at roughly 15–25% of build cost annually, forever
• Dependency and security updates, which are not optional
• Support for internal users, which lands on someone
• Onboarding cost when the person who built it leaves

A £40,000 internal tool is a commitment of roughly £70,000–£90,000 over three years. That is often still the right call — it just needs to be the comparison being made.

The real cost of buying

Licence fees that grow with headcount, configuration work that is often substantial, integration with your other systems, data trapped in a format you do not control, and a roadmap you do not influence.

The migration cost is the one people discover late. Ask before signing: can we export everything, in what format, and what happens to the data if we leave? A vague answer is a red flag.

The four signals that favour building

1. The process is genuinely unusual and it is why you win business.
2. Licence costs scale with headcount and headcount is growing.
3. You have evaluated three products and each needs heavy configuration plus workarounds.
4. Integration is the whole point and the products available do not talk to your systems.

The four signals that favour buying

1. It is a solved, standard problem — accounting, payroll, helpdesk, email.
2. Compliance obligations come with it that you would rather someone else maintained.
3. You need it working in weeks rather than months.
4. Nobody internally will own it long term.

The middle path that usually wins

Buy the commodity layers and build the thin part that is actually yours. Use an off-the-shelf CRM, database and authentication, and build the specific workflow layer that encodes your advantage on top.

This gets you differentiation where it matters and someone else's maintenance burden everywhere it does not. It is less satisfying than building everything and considerably cheaper to own.

Frequently asked questions