Kroger had $108 B.
Walmart had $485.9 B.
When you're selling that kind of volume, you're not doing it without algorithmic help.
Absolutely agree that OP should have been able to more easily get an in-person override. But I understand why the original rejection happened: or rather, I understand why HD uses a data mining company to attempt to prevent theft.
PS: Side note, you'd be surprised how big of a chunk of theft directly involves receipts / returns in some way or another.
These companies have been profitable for a long time prior to now. Why is it that they suddenly need algorithmic help?
If you're running a < 5% net profit margin, as most of these businesses are, then 1.5% inventory shrink (approx average) is a pretty big deal.
In any case it’s ridiculous that a manager can’t look at someone’s history of buying lots of stuff and override a one-time return for a fraction of a percent of their recent history.
The right answer here is that the algorithm probably screwed up in this instance. There was a clear signal for a legitimate customer: large order, small return.
Saying that the algorithm screwed up is confusing a symptom with the problem: making business processes so inflexible that local staff cannot correct mistakes.
In a case I know of, LP colluded with the customer service manager to steal $10K cash, by shifting the camera focus from the money counting area. In this case, both the LP manager and the customer service manager got fired, and, of course, without any charges, since there is no proof.
At Fry's electronics, to return a product, LP has to key in his password as well. If the price of returned product is more than $300, LP and some other manager have to key in their passwords.