Amid Musk's Reign at Twitter, Our Digital History Is at Risk

It is a big deal if a million of active users leave.

Sampling a random selection of ~50k profiles, I found that only 9% had ever posted public content, and ~ 0.3% had been active within the past month (excluding YouTube activity exposed through G+). This itself is an interesting validation of the 90-90-1 rule of thumb for online discussion participation.

(Those results were independently validated a few months later by Stone Temple Consulting, an online marketing agency, since acquired by Perficient: <https://blogs.perficient.com/2015/04/14/real-numbers-for-the...>)

I also had a chance to look at several lists of highly-active / highly-ranked profiles over time, and found that, at least to a first approximation, more activity was a stronger predictor of later account deletion --- whether elected by the individual (e.g., disenchanted) or the platform (spam / abuse removal). This is counterintuitive at first blush but makes sense on modest reflection.

For Communities, what I found was a fantastic rate of creation and destruction. There were ~8 million communities created by the time new communities were discontinued (~January 2019), and the creation rate was on the order of thousands per day. There was also a high destruction rate, evidenced by the fact that within an hour or so, many of the communities which had been listed via a recent sitemaps query were no longer extant. I'd commented on that at the time.

For Twitter, it's highly probable that:

- Account defections are among more active members.

- Those members account for a large fraction of overall content.

- Those members account for a large fraction of significant content. E.g., not spam or bot traffic.

Of Twitter's 200m MAUs, it's likely that roughly 90% (180m) participate seldom if at all, 9% (~18m) occasionally, and roughly 2 million are the most active members. Loss of as few as a few percent of all members could see removal of virtually all substantive content.