What Twitter Wants

Four weeks ago, based on things I had observed previously and some of the conversation here I realized how I was using Twitter was all wrong. It became clear that what I was doing was resulting in them thinking of my accounts as “low quality” and because of that I wasn’t getting the visibility that I was after.

To put it simply, they don’t want activity, they want interaction. I’d put up a few of my own posts, but then I’d retweet posts from “community members” that matched the theme of the account (I’ve got quite a few accounts). Because retweets with comments often get mangled when the account your retweeting is marked “sensitive content” I wasn’t adding any comment. Thanks to the QAnon/Trump folks, over the past 6+ months Twitter has taken a real disliking to retweets - even (kinda) turning off the feature late in the election process. Because I didn’t “add anything to the conversation” what I was doing was problematic / low quality in their eyes.

So 4 weeks ago I changed strategy. I had been “thanking” people for mentions by liking the post. I stopped that completely. And I stopped almost all retweets. Instead, now when I see something I would have retweeted in the past I do one of two things - I either like it and then reply with a short, positive comment like “Great cock!”. I might include some hashtags, but often I wouldn’t even do that. OR I’d tweet the picture of video using the method I described in a thread on XBiz.net.

Anyway, the like + comment looks like interaction to Twitter, and tweeting the pic/video with a meaningful comment looks like quality original content to them. The results were quite a bit better than how things were going before. I looked at the analytics about halfway through their 28 day reporting period and then at the end. Here’s the results…

An account with 35K followers

2 weeks, 2 days later
Tweets up 333%
Tweet impressions up 206%
Profile visits up 3.8%
Mentions up 8.3%
Follower graph shows noticeable uptick in slope of line

4 weeks later
Tweets up 150%
Tweet impressions up 340%
Profile visits up 14%
Mentions up 6.4%

An account with 12K followers

2 weeks, 2 days later
Tweets up 3,567%
Tweet impressions up 438%
Profile visits up 50%
Mentions up 142%
Follower graph shows slight uptick in slope of line

4 weeks later
Tweets up 1,857%
Tweet impressions up 606%
Profile Visits up 83%
Mentions up 159%

An account with 3,400 followers

2 weeks, 2 days later
Tweets up 160%
Tweet impressions up 11.4%
Profile visits up 30%
Mentions up 75%
Follower graph shows noticeable uptick in slope of line

4 weeks later
Tweets up 360%
Tweet impressions up 127%
Profile visits up 47%
Mentions up 70%

What’s important to note is that even though I was technically tweeting more, my total time spent doing stuff on Twitter was about the same.

When people talk about “shadow banning” I honestly think this is part of what’s going on - you’ll get shadow banned if Twitter considers your use of their platform to be “low quality” or (worse) spammy.

Bottom line - stop retweeting and you’ll see better results.

1 Like

That’s a very detailed test you did, well done. I don’t have any stats to compare with but I have noticed since we started to interact with users and other Twitter accounts the amount of followers have increased more and faster. Now, my complaint with Twitter is that it still does not drive much traffic to my site, never has. It feels like a closed system in a way, just like Instagram.

Bjorn - do you use Twitter Cards? I find that’s the best way to drive traffic to a site. When the user clicks on the image to make it bigger they’re taken to your site. Here’s an example (you sorta have to see it on Twitter, the preview doesn’t really get the full effect)…

Yes, we did start to use them and could see it improved engagement (i think its called)

Good job Jay! I haven’ done a study but have noticed that the best results on twitter were always from actual conversations. It does take a bit more time to do it meaningfully, but the idea is to read through what relevant accounts are saying and reply, or ask questions. Twitter is certainly using algorhythms to determine whether it’s a bot running the account or an actual person, and neither posts not retweets guarantee a real person like actual conversations do. When I ran ModelCentro twitter I’d engage in all sorts of talks - from recommending a breakfast recipe to a model who asked about it to suggesting good books (and adult ones of course but it always had to be a conversation) - and that account grew like bamboo.
On @Mojohost, I’m doing it a bit differently. I aim to crete content that will in itself result in conversation. It’s difficult being an infrastructure provider, but still possible. I utilize everything from ticket reviews that are humorous to memes to get that interaction, and it’s working quite well.

This is great information. I was wondering what was happening there. Thanks Jay !