The output of Data Science is harder for non-specialists to evaluate.
Not as management. You just have to see that other people's similar sites are not slow with the same resources, therefore it is possible for your site not to be slow. You don't have to know why you're failing to know that the totality of the people you hired were not as good as the people those others hired.
This is of course barring management failure; but if you're failing at management, that's about the same as saying that your engineers were under-resourced.
Engineering competence is largely composed of the skills to figure out what is causing problems e.g. slowness. If you can't figure out what is causing the slowness, your engineers aren't good enough to figure out what is causing the slowness, qed.
That's different than data science.
In other disciplines it is way more fuzzy. If you are in the conclusion business and there isn’t a clear path to test your conclusion in the short term you can bullshit away!