In retrospect, I should have mentioned a fifth stage - CCF upload to actual posting. I was unexpectedly delayed by twenty or so minutes due to being asked to urgently buy some groceries.
Of course, it's not much of a write-up anyway, and the photos suck - the coin was apparently really unphotogenic...
As I have already mentioned, I don't actually collect most series by date actively enough to look for the keys.
In the few cases where I do, as often as not, I end up saying "nope, those are too rare to bother with" and just not counting the ultra-rare dates in the first place.
But I see no immediate reason why
the key date for, say,
this type should be 1999 (the mildly scarce date, which I have - think 2009 nickel for relative scarcity) and not 2001 (attested in several known examples, with an estimated mintage of at most a few thousand).
I mean, I don't have the 2001 coin, naturally, but if I did end up stumbling on one, I probably would have included it in my set.
(Or sold to get rich, more likely, but that's beside the point.)
Anyway, with Russian series, which I do actively collect (some of them), I usually already have all the normal dates and have nearly no chance of ever getting any of the ultra-rare dates; in any case, the mintages are unknown, so it's not like I could meaningfully say which date is the key (ignoring the ultra-rares, anyway).
I do collect some non-Russian series by date; again, the mintages are often unknown, and the purchases are usually sporadic, and in any case if the key is significantly more expensive, I usually don't bother looking for it unless I already accidentally have it.
Of course, more often than not, if I think I accidentally have the key, that just means it's not the
actual key, and the official mintages are misleading.
Notably, this is the case for
Victorian bun pennies, where the 1860 (which I have) has the lowest recorded mintage, but the 1864 (which I also have) is actually much scarcer, and the actual rarest date appears to be 1869 (which I don't have).
Anyway, all of this means that one of the very few ways for me to actually have the key (aside from accidents, which I actually mostly don't recall) for a series that I could be said to collect by date is to get (nearly) the entire series in one shot - usually from a bargain bin.
Of course, your typical bargain bin isn't all that likely to have scarce dates, so more often than not, I'm still missing a few, and the key is usually among those few.
Which brings me to what might well be the only obvious exception in my collection - that I could recall, anyway...

Luxembourg, 25 centimes, 1968.
With a mintage of 600,000, easily the key date of the
1954-72 aluminium 25 centimes type; this is confirmed by the Numista exchange lists (a lot shorter for this date than for any other date of the type).
And yes, I ended up getting all eight dates of this type in one shot from a single bargain bin, at the 2018 COINS show in Moscow, on November 2, 2018.
(I had seven different dates when I decided to check on Numista whether any other dates exist; it pointed me to the missing 1968.
I had to check most of the bin for that, but did find an 1968, which is the coin I'm trying to show you today.)