The numbers have moved
For a decade, the default answer to "should we build or buy?" has been "buy unless there's a strong reason to build." That default was correct until very recently.
The cost of building bespoke software has fallen for the kinds of organisations that were previously priced out of it. Small teams using modern tooling and AI-assisted engineering can deliver in weeks what would have taken months even three years ago.
You still should not build everything. The buy-by-default rule of thumb is no longer reliable, and decisions that were obviously buy in 2022 are now genuinely closer to even.
Where build is suddenly more competitive
- The "almost-right SaaS", when a 90% fit is costing you in workarounds and per-seat fees.
- Operational systems that encode your competitive advantage.
- Customer-facing software where the SaaS aesthetic is hurting brand.
- Integrations between systems that don't talk to each other.
Where buy still wins
- Commodity functions (payroll, accounting, identity, payments).
- Heavily regulated areas with mature, certified vendors.
- Anything where the SaaS is genuinely best-in-class for your scale.
Most of the time you still buy the commodity and build the differentiator. The line between those two has shifted, so it is worth looking again.
