So what's a good approximate metric for the possible damage if something does go wrong? Some function of fixed operating costs and customer turnover? Driving away a customer that you'd expected to keep for another 10 years is a bit different than if you'd only expected to keep them for a month, and losing half your customer base is a bit worse if your fixed costs are 80% of revenue vs if they're 20%.
The example had two cases, at opposite ends of this. Are there generally accepted not-too-heavyweight ways to calibrate procedures for cases that are more in-between?