Likely they have a standard procedure to test from manufacturer distant to manufacturer caused, following a decision tree, that is not a learning system.
Such a procedure needs constant (automated) monitoring for the frequencies of found failure causes, and move tests for these causes forward. Thus skipping tests that ultimately would lead there eventually anyway reduces consumer/client frustration as well as support costs.
However, a young company can not set up everything perfectly at once, so I’m willing to concede them some slack for their first 1…3 years (depends how quickly issues become visible and how easy it is to change SOP).