This is not as fresh as I’d like it to be (I used data provided by @Piggy at the time, which has not been refreshed for some time), but it gives us a general idea of whether profiles created on a given year are still active years later, or whether they drop-out along the way. The data is from November 2018:
<…>
Interpretation:
- For example, in 2012 (read at row level) -> 27.846 accounts were registered. Out of those, 10.897 (39,13%) dropped out in 2012, 2.994 (10,75%) in 2013, 1.194 (4,29%) in 2014, 700 (2,51%) in 2015, 400 (1,44%) in 2016, 2.832 (10,17%) in 2017, and only 8.829 (31,71%) of the account have logged-in during 2018 so far.
- Now if we focus specifically on the 2018 column, we see that out of the accounts created during 2009, 54,55% logged-in during 2018, only 20,66% of those created in 2010 did so, 46,88% of those created in 2011, 31,71% of those in 2012, 15,93% of those created in 2013, 15,95% of those in 2014, only 4,80% of those created in 2015, 9,21% in 2016 and 33,52% of those created in 2017.
- 2015 and 2016 follow a weird pattern, since the large majority of the created accounts were only active during the year of creation (in 2015 this happened with 89,50% of the created accounts, and 84,95% in 2016). These values are much higher than other years by comparison.
The hacking that took place in 2015 may certainly have had to do with it, scaring people off, but accounts created before didn’t dropout in masses during 2015/2016.
- Accounts created in 2011 are, on the other hand, the longest survivors, with 46,88% of the accounts having logged-in during 2018 (2009 created accounts have a larger ratio (54,55%), but are a small number). <…>
The interesting thing here is that, on November 2018:
54,55% of the accounts created in 2009 were still active.
20,66% of the accounts created in 2010 were still active.
46,88% of the accounts created in 2011 were still active.
31,71% of the accounts created in 2012 were still active.
15,93% of the accounts created in 2013 were still active.
15,95% of the accounts created in 2014 were still active.
4,80% of the accounts created in 2015 were still active.
9,21% of the accounts created in 2016 were still active.
33,52% of the accounts created in 2017 were still active.
To take into consideration: existence of massive Newbie account creation on certain years (for bot spam I figure), forum hacks, etc. These facts affect wildly the percentages. The thread where the quote comes from details the information by Ranks.