Internal platforms — the dashboards, workflow tools, and admin systems that run operations — are where engineering time quietly disappears. A short framework keeps that spend deliberate.
Start with the question you are actually answering
“Build or buy” is really three questions: does an off-the-shelf tool fit the workflow closely enough, is the workflow a source of competitive advantage, and what is the total cost of ownership on each side over three years — not just the licence or the first sprint.
If a tool fits and the workflow is generic, buy it and move on. Building a worse version of Notion or Retool is not a strategy.
Build when the workflow is the moat
Some internal processes are the business — pricing logic, fulfilment orchestration, a proprietary review pipeline. When the workflow is a differentiator, owning it in code is worth the investment, because every improvement compounds and no vendor roadmap is in the way.
The tell: if you find yourself fighting a SaaS tool's assumptions every week, that friction is information.
The third option: compose, don't build from scratch
Often the right answer is neither a pure buy nor a ground-up build. A thin custom layer — an integration, a small service, an internal API — on top of bought components gives you the fit of a build at a fraction of the cost and maintenance.
This is most of what good platform engineering looks like in practice: not heroic in-house systems, but a small, well-chosen amount of glue.