Once a company realizes that it's not magical unicorns, they're too deeply embedded. Is it better to spend X every year to just make the Salesforce solution wo

Once a company realizes that it's not magical unicorns, they're too deeply embedded. Is it better to spend X every year to just make the Salesforce solution work (thereby not having to admit to a mistake, and mitigating personal risk), or spend 3X in one year to write something in house (that requires maybe 1/10th X to maintain/extend from then on, but which is riskier, and unlikely to show up on future evals such that they can claim credit and get a raise)? Corporations seem especially vulnerable to sunk cost fallacies.