I've recently been reading/watching some videos and writings by Robin Diangelo on systemic racism - here's a great starting point: https://www.youtube.com/watch?v=h7mzj0cVL0Q. She also wrote the book "White Fragility".
Thinking about that, I'm just wondering how different it would be if one of those people who mistook the employee for the CEO instead turned to the CEO and said "I'm sorry, please excuse me for the instance of racism I just perpetrated against you, I promise it won't happen again." I realize how outlandish that may sound writing that out, but I'd propose that the fact that it does sound outlandish is the main problem. Everyone in the US was raised in an environment that inculcated certain racial ideas, subconsciously or not. We can't address them if we're so embarrassed by their existence as to pretend they don't exist.
It means your behaviour is informed by racial profiling of an individual.
It means that you're not treating someone as an individual, but rather based on membership of a racial group he happens to be born in, which has statistical characteristics (e.g. lower chance of being a CEO) that do not necessarily have any bearing on the individual at all.
We define treating people distinctly like that because of their racial membership, as racist. That's really just the definition. You can have a discussion about whether you think racism is justified or not, and make your own value judgement. You could say that even if membership of a group does not necessarily say something, odds are that it can be a good way to infer things. And that's true, group-membership (e.g. your ethnicity) has many useful correlations from which to infer things. But to say it's not racist simply isn't factual, it is racist according to how we define it. What's left open for discussion is whether racism is okay or not.
Of course as a society we have indeed had that discussion and fortunately decided that racism isn't justified, not okay, and should be prevented as much as possible. I'm happy about that. Because even if group membership (e.g. race, ethnicity, gender, sexual preference, religion etc) has correlations with all kinds of outcomes we may wish to approximate, as a society we agree that it is only fair to judge people on their individual merits, and not on the group that they belong to.
1. We make judgements on “group membership” all the time, and to the extent they’re correlated to the outcome (i.e. true) that’s not wrong or -ist. Example would be, do you avoid stepping in front of a car because you infer that the car might run you over, or do you just take your chances and hope that this individual car will stop. Ok, contrived example. How about this: people generally accept that men are more dangerous than women (e.g. when it comes to rape, stranger danger, domestic violence, courts, jails, ...) even though the vast majority of men aren’t violent at all! Is this sexist? Yes, to some extent, in particular when the state does that (e.g. always arresting men in case of domestic violence call). But at the same time, that kind of “prejudice” might just save your life; how wrong can that be?
2. Even if there’s no correlation to the outcome, not every inference is -ist. Example I mentioned downthread is, assuming that “Alex” is a man. Is that really wrong?
3. Which brings me to, you write that “treating people distinctly” is -ist. But what is treating? Again, people make inferences (that’s literally what intelligence is, short-term prediction engine). Sometimes that’s even embedded in the language (e.g. in Slovenian, you have to assume gender, unlike in English). But as long as we remain open to change, that’s fine! If the woman tells me that actually she’s “Alex”, the only actually sexists way would be if I refused to call her by her (masculine-ish) name.
TL;DR: if everything is racist, then racism cannot be immoral.
If it is based on ones favorite sports team, it is just stupid.