Quote Originally Posted by dimab View Post
Williams and Holland's Law:
If enough data is collected, anything may be proven by statistical methods.

Below is some statistics I could gather by somewhat limited "Advanced search" of our members list: I looked for semiannual number of active (10+ posts in my definition) users that have posted their first or their last post during each period. The 10+ limit I took to drop all the one-time posters (valuation requests etc.), and also I didn't include H2 2018 - as we still couldn't say with all certainty that whoever hasn't posted since then has quitted (I hope not ).
Name: SRP Forums Semiannual changes.png Views: 190 Size: 12.3 KB
I'm sure the moderators, probably having more direct access to forums' database, could have made more detailed analysis - based on total post counts, posts per user etc. etc., but judging from the graph above I wouldn't say there was a significant change in the last year or two - just a continuation of previously established tendency. On the other hand, remembering the law I cited in the beginning, your conclusions may vary
That's pretty cool!!!

I've looked at different measures in the past and the overall picture is what I mentioned earlier. My best measure ended up the visitors which is different than the forum activity though they're strongly correlated. The former is larger than the later so less noise and also I think it is more representative of the forum impact which to me is more important than the size or activity.
As far as cut-out time frame, some time ago when I analyzed the data I found that 3 months of inactivity is where people basically drop off long term. Of course that's statistical and there are plenty of exceptions plus personally I look at every individual as unique and interesting completely unrelated to anything statistic.