This used to be an easy decision. Bespoke software was expensive and slow, so most businesses bought a platform and adapted themselves to it. The economics have shifted, and the answer is now more interesting.
Buy what is common
Accounting, payroll, email, document storage. Thousands of businesses need exactly the same thing, and established products do it well. Building your own would be an expensive way to arrive at something worse.
Build where you are different
Every business has a handful of processes that are particular to it: how jobs are quoted, how work is scheduled, how information reaches the people who need it. Off-the-shelf platforms handle these awkwardly, and the gap gets filled with spreadsheets, workarounds and people re-keying data. That gap is where a focused, bespoke tool makes sense.
The third option
It is rarely all or nothing. Often the right answer is to keep the platform you already have and build a small layer around it: a dashboard that pulls from three systems, an internal tool that covers the step the platform does badly, an automation that connects two products that do not talk to each other.
Questions worth asking
How much time does the team lose working around the current system? Would the business have to change a process it values in order to fit the platform? What does the subscription cost over five years, per user, as you grow? Who owns the data, and how easily can you leave?
Bespoke no longer means enormous
A focused application that solves one operational problem can now be built in weeks rather than quarters. The discipline is in keeping it focused: build the part that matters, and buy the rest.