Zepla: Endwalker is the Worst FFXIV Expansion I have played

[Absurdity]

Aye, the more I dig the more those figures I used look bad.

SE announced a 27 million characters created milestone last year with estimates that it'll hit 30 mil at some point this year (Likely with some fanfare when it happens). There's no way 20 mil of those are actively subbing still IMO.

Whoops!

Even Lucky Bancho has some funky numbers. They reported 1,130,000 "active" characters as of the 22nd of August.

Now the question is at which point a character is considered no longer active. Since the previous census was on the 20th of June there has been quite some time between the two, so we actually have no idea if this is monthly active players or daily or simply characters active between those two data points and then you also have to take into account that "active characters" doesn't actually mean "subscribed accounts" since it includes alts as long as they are actively played.

[Lyth]

Now the question is at which point a character is considered no longer active.

To be considered active, your HP, XP, mount count, or minion count has to change between two surveys. So this value will be an underestimate, but that's the point.

[Jeeqbit]

Now the question is at which point a character is considered no longer active.

If there is no observable change on their lodestone profile. I don't think they check gear or title changes, but it does check for a lot of other changes such as a different job, changes in level, exp, mounts or minions.

we actually have no idea if this is monthly active players or daily or simply characters active between those two data points

They are very clear it is since the last survey, which at the moment seem to happenen every 2-3 months. Obviously, this doesn't necessarily reflect the number of people subscribed right now, but the data is still relevant because it allows you to compare survey to survey or the difference between a survey in one expansion compared to another, or between the same month on two given years.

If a survey now is saying "more than 1 million" and surveys always used to be under a million, it still tells us that the game is in a better place statistically than it used to be and therefore to see a decline we'd have to start seeing numbers below 1 million again.

and then you also have to take into account that "active characters" doesn't actually mean "subscribed accounts" since it includes alts

"Comparison of the number of people who acquired bonus mounts from the 7-Eleven collaboration campaign held in 2016 and the number of active characters who owned the same mounts at the time. The other is an article from June 13, 2019, which compares the number of paid members at the general meeting of shareholders at the same time and the number of active characters at the time.

In the 2017 survey, the alt character rate was 14%. A 2019 survey estimated 20%.

The reason for the increase in the latter is that it was during the pre-expansion period, when there were few updates, and also because level boosts that were not available during the former period made it possible to exceed the active threshold. Still 20%. If there are 5 characters, one of them will be an alt character for one of the remaining 4 characters."

(Translated from their blog)

[Lauront]

Even Lucky Bancho has some funky numbers. They reported 1,130,000 "active" characters as of the 22nd of August.

Now the question is at which point a character is considered no longer active. Since the previous census was on the 20th of June there has been quite some time between the two, so we actually have no idea if this is monthly active players or daily or simply characters active between those two data points and then you also have to take into account that "active characters" doesn't actually mean "subscribed accounts" since it includes alts as long as they are actively played.

Provided you meet the criteria it has to count a character as active (he lists them here - you can run it through deepl if you don't speak Japanese) and log in at least once during the survey period and perform activity meeting that threshold, for the purposes of that survey, you are considered active within the period it ran. This is why it's neither usable as an active monthly player metric nor as one of paid subs. It also can't ferret out alts or free login campaign activity, where applicable. The fact that he says he miscalculated the cutoff date for the most recent one, meaning it captured some of 6.45, also means it would've captured some people who re-subbed for the mini patch, which was in his notes on it.