Build vs. Buy: The Framework We Actually Use With Clients

Clients rarely bring us a clean “build or buy” question. They bring us a problem, and somewhere in scoping it, the question surfaces on its own. It’s one of the more consequential calls in a project, and it’s also one of the easiest to get wrong by defaulting to whichever answer is more familiar — “we always build” or “there’s definitely a SaaS for that” — rather than actually working through it.

Three questions before anything else

Before we talk technology at all, we ask three questions, roughly in this order:

• Is this core to how you compete, or is it plumbing? If the workflow is genuinely part of what makes the business differentiated to its customers, owning it usually pays off. If it’s a well-understood back-office function every business in the sector needs, that’s a strong signal to buy.
• Does an off-the-shelf tool solve 80% of it cleanly? The dangerous middle ground is a SaaS product that covers most of the requirement but needs heavy customization to cover the rest — that’s often more expensive and more fragile than either a clean off-the-shelf fit or a purpose-built system.
• Can you actually own the five-year cost, not just the license fee? Whichever path you pick, someone has to own upgrades, integrations, and the knowledge of how it works, for years, not just the first year.

The hidden cost of “just customize the SaaS”

Off-the-shelf software is genuinely the right call more often than engineering-led teams like to admit — for well-solved problems, it’s usually cheaper, better tested, and comes with a support team you’re not staffing yourself. The trap is heavy customization layered on top of it: every custom field, workflow rule and integration you add is a piece of debt that has to survive the vendor’s next major release, often on their schedule, not yours.

A SaaS platform customized past a certain point stops being “buy” and quietly becomes “build,” except now you’re building on someone else’s foundation with none of the control.

The hidden cost of building it yourself

Custom software has the opposite failure mode: it’s easy to underestimate everything that isn’t the visible feature — authentication, backups, monitoring, the second and third round of edge cases that only appear once real users touch it, and the ongoing cost of a team that keeps knowing how it works after the people who built it move on. We build custom software often, and it’s the right call often — but only when the client goes in clear-eyed about owning it for years, not just funding the first release.

The hybrid pattern we use most often

The answer that comes up more than either pure option: buy the commodity layer and build the differentiator. Use established platforms for accounting, identity, or payments — the parts every business needs and gains nothing from reinventing — and invest custom engineering only in the specific workflow that’s actually unique to how this client operates, connected to those platforms through clean integrations rather than replacing them.