That chart is based on Google Trends, which is a terrible way to measure social network activity/popularity.

A lot of people don't use bookmarks and even when they know very well it is facebook.com they still hit search "facebook" on Google.

That doesn't matter much. The largest difference is that the majority of people access social networks through mobile apps now, and Google Trends can't reflect that at all.

For example, Facebook shows a 71% drop worldwide in Google Trends over the last 5 years. Their Q2 2019 earnings showed a 20% worldwide growth in monthly users over the last two years alone.

Google Trends can't tell you anything useful about a site's growth or activity.

Do you have anything better?

The assumption that Tumblr's loss of users is entirely or even mostly due to suppressing adult content is probably unwarranted, but I'd be interested in data either way.

The chart shows that there was already a downward trend, but it really fell off a cliff after the announcement.

Also, for platforms with network effects, a drop-off (or even belief that there will be a drop-off) can do major damage to even unrelated communities on the platform. For example, Tumblr used to be a huge referrer for my startup, thanks to Tumblr's robust ADHD, dyslexia, and CFS communities who spread the word about our tools. We still get occasional traffic from posts, but I get the sense that those communities have been hollowed out by the perception that Tumblr is dead. I don't know if they've moved elsewhere (perhaps Twitter), but they sure aren't on Tumblr much anymore.

I certainly hope Automattic can revive Tumblr one way or another, but it seems like the smart play here would have been to announce that they'll be looking to find a way to restore the prior functionality, and then work out the details later. That way they can get a bump in traffic from the expectation that there will be a bigger community, which helps them regardless of the extent to which the ban is reversed.

On July 15, 2017, ~36 million posts were made on Tumblr [1]. A year later, July 14, 2018 [4], about ~31 million posts were made [2]. July 13, 2019 saw only ~21 million [3].

The Wayback Machine doesn't have fine-grained enough details to get a good estimate of how much the loss of adult content itself was responsible, but this does reinforce the impression given from the Google Trends link cited several times that there has been a steady decline in Tumblr punctuated with a massive hit on or around the removal of adult content.

[1] https://web.archive.org/web/20170711122839/https://www.tumbl...

[2] https://web.archive.org/web/20180714223645/http://tumblr.com...

[3] https://web.archive.org/web/20190713180936/http://tumblr.com...

[4] Each of these dates is a Saturday, so not quite a year later, but matching time of year and day-of-week is better anyways.

People who had nudes on their tumblr weren't on tumblr just for nudes. It was just great to have a place you could stash everything you liked, kinks included.

So, network effects work both ways, people leave because their kink got banned and people who followed that person now have one less reason to log on.

Yes and no. As someone commented above, when people lost one of their interests, the dropped it completely and took all their presence away.

One of the reasons that I left is the need for Yahoo to share all my actions with some of the shittiest trackers, advertisers, and privacy violators. Not fun fact: you can't shake/AdBlock/NoScript/PrivacyBadger them if Tumblr itself is doing the sale.