So far, one beta tester out of 500 has reported the problem. That extrapolates to about 400 users. That works out to about $48k of annual loss if they all quit. In corporate-speak, that is acceptable loss compared to enginerring/design shortcut cost savings.