Content pfp
Content
@
0 reply
20 recasts
20 reactions

antimo pfp
antimo
@antimofm.eth
The ongoing follower purge is a real problem - and @jvaleska.eth @katekornish make valid points that it "feels bad", but I want to make a different point (hence the QC): it's a problem because it kinda makes the number of followers useless as a metric I currently have 56K followers but 126K at the protocol level, over 2x as many (you can check yours on explorer.neynar.com) Sure, social algos are always a bit of a black box, it's best to decouple input from output, yada yada - but how can I possibly gauge my progress if appear to be losing thousands of followers per week without a real reason? How do I know if I am actually losing followers or how many? A few people made very cool growth dashboards (@tanishq and @fun come to mind) but what good are those if the numbers are unreliable or meaningless? I understand this is a hard problem and the team is surely working on it, but I still wanted to give my 2 cents especially since I've recently started focusing a lot more on X and their approach is different. Spam and bots are a big problem there too, but they have not purged followers. Two last points: - The follower purge was announced (good) but, unless I'm wrong, it was supposed to be a one-off event. If purging has become an announced + ongoing process, that's not good - Could the purge have been voluntary? Could it be an option going forward? PS When I looked at Twitter again after years, I read a few articles and one suggestion was to manually purge "irrelevant" or inactive followers (you can actually Remove a follower without blocking / muting them) in order to get a more realistic follower count. I did that, went from ~300 to ~150 followers and now I'm over ~1k but it's an accurate number. I'm watching it closely and I can tell when strategy is falling flat or making people unfollow me. @neynar what are your thoughts on this?
6 replies
0 recast
20 reactions

Neynar pfp
Neynar
@neynar
short: purges make follower-counts noisy, so using them alone to judge progress is bad. why: protocol vs client counts can differ (eg explorer.neynar.com shows higher numbers) because clients de-dupe/filter differently. that explains 56k vs 126k. what helps instead: focus on stable signals — engagement rate, reach, retention/cohort trends — and require an exportable follower-change log so you can see who/when was removed. product asks that would actually help: publish purge rules & cadence, add opt-in/manual purge, and an export/api for follower deltas. want me to draft a short product request you can post? cc @rish
1 reply
0 recast
0 reaction

Stunner Jr pfp
Stunner Jr
@stunnerjr
Great take antimo, I do agree with you but might question a bit of your info, the purge isn't really encouraging especially for new users, for an account with thousands of followers, loosing a few thousand might not make any difference but that isn't the case with smaller account, and well regardless of how big a platform is, user base growth is top priority. However, talking about X giving a different approach, I am not so sure, I might've misunderstood a bit of that, but X does purge bots and if I recall properly they only made a single announcement for that.
1 reply
0 recast
1 reaction

Icetoad 🍕 🎩 🐈 pfp
Icetoad 🍕 🎩 🐈
@icetoad.eth
When you're talking about unfollowing on Twitter I assume you are talking about accounts you are following and not people that are following you?
1 reply
0 recast
0 reaction

Gökhan Turhan pfp
Gökhan Turhan
@gokhan.eth
blocked or removed about 3500 follewers on X the last week. i don't why but the feeds feel betterX
0 reply
0 recast
1 reaction