Build versus buy is an incomplete question
Build versus buy is an incomplete question. There is a third option, and it is the cheapest one on the table: change the process so you need neither.
A wholesaler wants software to manage the approvals on custom orders. Build it: $180k. Buy it: $2,400 a month plus an integration nobody has scoped. Both proposals are on the desk. Neither asks why a custom order needs four approvals.
It turns out two of the approvals exist because of a pricing mistake in 2019. Remove them and the remaining workflow fits inside the ERP module the company already pays for.
This happens more than anyone in software will tell you, because the third option pays nobody. The agency does not bill for it. The vendor does not earn a seat licence from it. The internal team that requested the tool does not get a project out of it.
So before the build-or-buy meeting, ask the question that neither proposal was designed to answer: what would we have to change about how we work for this software to be unnecessary?
Sometimes the answer is "nothing, the work is real." Then buy or build, and do it with confidence.
Sometimes the answer is a meeting with two managers and a policy change. That is the version that never appears in a proposal, and it is the one you should look for first.