Build vs buy: when custom software beats another licence

Start with the bill you already have

Write down what you pay today: seats, add-ons, implementation partners, admin time, and the hours spent copying between systems. That three-year figure is the baseline. A custom build only has to beat that baseline, not beat every SaaS product on the market.

Per-seat platforms look cheap until occasional users, contractors or branch staff each need a login. Distinctive workflows look free until a partner bills by the day to bend the object model.

Four answers, not two

Treat every expensive tool decision as one of four paths:

• Keep and configure: the product fits; you need better setup, training or integration.
• Switch SaaS: you need a CRM, helpdesk or ERP, but not this expensive one.
• Hire in-house: software is core to your product for years, and you can keep seniors busy.
• Build and own: the process is yours, seat pricing is the wrong shape, or vendors cannot meet ownership and residency rules.

When custom software usually wins

Custom wins when your process is a competitive advantage. Forcing it into someone else’s model degrades the thing you sell. It also wins when hundreds of light users each pay for a seat they open twice a week. It wins when you are already paying a partner to customise a platform that still needs spreadsheets beside it.

Common examples include an internal ops console and a client portal tied to your pricing. We also see scheduling or case systems sold as CRM packs. We see workflow layers that should sit beside a lighter off-the-shelf CRM rather than inside an enterprise suite.

When you should not build

Do not rebuild a standard sales pipeline, inbox or accounting ledger for ten people. Buy a cheaper tool. Do not chase a lower licence cost if the real job is unused seats and a renewal negotiation. That is an admin problem, not a software project.

A like-for-like clone of a major suite, including report builders, email and offline mobile for every role, is a product company commitment. Most businesses need the distinctive 30 percent built, and the boring 70 percent left on a simple licence.

Custom studio vs hiring developers

Hiring is right when you will ship software continuously for years and want people embedded in your team. For a defined internal platform, a studio is usually faster. It brings product, engineering and support in one engagement, a fixed first phase, and code in your repositories from day one.

Many clients do both. We build the first releases while they recruit, help write the role, and hand over a codebase with tests and documentation. The comparison is not agency versus ownership. You own the software either way.

A simple three-year sketch

Take annual licences and partners as line A. Take a phased build plus hosting and ongoing monthly support as line B. If line B is lower over three years and the process fit is materially better, build. If line A is lower and the tool is close enough, buy or switch. If neither is clear, pay for a short discovery that models both with your real seat counts and workflows before anyone writes production code.

Published SaaS list prices and custom quotes vary widely by region and discount. Treat any blog table as a starting point, not a tender. The useful artefact is your own spreadsheet with your renewals and your process.