Twitter Unfollow Practices: Spammers Are The New Inactives

When ManageFlitter and Crowdfire sat at the epicentre of Twitter follow management tools, usage of those apps helped spread a notion that one of the best targets for unfollowing was the inactive user. And the Twitter ‘manage’ apps made the process of following so-called “inactives” very, very simple.

But in January 2019, a range of Twitter unfollow apps, including both of the aforementioned, had their API access disabled by Twitter. This rendered the best known unfollow tools inoperable. And because these and other apps had such an enormous presence among Twitter users, the effective shutdown of their core functionality actually changed the Twitter landscape – even for those who never used them.

For example, you may recall that this time last year, if you went more than 30 days without updating your timeline, your mutuals would start to unfollow you. That was because ManageFlitter and its derivatives had an “Inactives” unfollow category, which used a 30-day filter to determine who was, or was not, still using Twitter. The implication was that app users should unfollow accounts in the “Inactives” category, and many did.

Since ManageFlitter and similar apps have lost their core Twitter functions, that post-30-day drop in followers no longer happens to anything like the same degree.

But supposed inactivity was never a good way to determine who could safely be unfollowed, as I reported in the 2015 post Why You Should Not Unfollow Inactive Twitter Users. The truth was that many of these apparently inactive users were not inactive at all. A lot were still logging in, and I found that a high proportion would unfollow back – even if it took them a number of weeks to get round to it.


Another big change since Twitter’s blitz on big-name unfollow apps, has been the decline in power among spammers. These are people who’ve always used automation or so-called “managed services” to follow a vast volume of accounts, and unfollow the ones who either didn’t follow back, or did, but later unfollowed. Auto-spammers’ friend lists are often huge. Frequently five digits, sometimes even six. And many of these spammers were using the same tools that allowed them to follow and unfollow at scale, to also smokescreen their spamming.

For instance, rather than simply posting link after link, they’d pad out their timelines with retweets, attributed quotes, and other stuff a bot could collect and post for them. So when they followed you, you might take a quick look at their profile, see that their feed looked legit, and follow them back. They’d never, ever look at your tweets – ever. And you kind of knew that. But they put an extra one on your followers count, and you could easily mute them. So it went on…


The good news is that in today’s post-ManageFlitter/Crowdfire environment, many of those spammers no longer have the capability to unfollow you. There’s no app managing their friend list anymore. So if you’re one of, say, 50,000 people they’re following, and you unfollow them, then unless they’ve switched to another unfollow app and are personally taking on the work of monitoring what’s going on, they probably won’t know what you’ve done. Thus, spammers with significantly large friend lists are now, I’ve determined, a much safer unfollow than so-called inactives.

The bad news? I’ve yet to find an app that reliably identifies spammers, so if you want to reliably ID them, you’re probably going to need to look at their profiles. Manually.


Please let me stress that I’m not saying you can unfollow any person with a five digit friend list, without any danger of them unfollowing back. That’s not the case. There are still unfollow apps available, so if users with big lists are prepared to take the time to unfollow manually using those apps, they will still see you’ve unfollowed them, and unfollow you back.

What I am saying, is that spammers are typically by nature lazy. So if spammers have five digit friend lists (and sometimes even if they don’t), the theory is that they’ll be too idle to bother managing their lists manually, now that the auto and managed services have been disabled. That means in many cases you can unfollow them, without being unfollowed back. But only the lazies. Not literally everyone with a five digit list.


Going through my Twitter friend list and manually checking out the profile page of each account I was following, was something I never did until after the great app purge of 2019.

If this sounds like a terrible chore, you already know you’re following an absolute mass of useless accounts, and that should be all the more reason to do it. Obviously, if your friend list is huge, looking through each profile page, one by one, will be impractical. In that situation I’d suggest being realistic about what your account is achieving. If your tweets get good engagement, you obviously won’t want to change anything, and you’ll probably just have to accept that your friend list is what it is.

But if you’re following, say, 35,000 accounts, and your tweets get one or two Likes each, seriously consider starting a new account and building a much smaller and more meaningful friend list from scratch. Keep the old account, but direct it, and any followers who are actually listening, to the new account.

The best thing about having a relatively small friend list is that you can see what’s going on. You know who you’re following, and it’s much easier to interact with people from the main Twitter homepage. You can’t do that if you’re following 10K, and your timeline is overwhelmed with spam.

Another great advantage with a small friend list is that if Twitter completely pulls the plug on follow/unfollow apps at some stage, you don’t face suddenly losing an absolute mass of mutuals, upon whom it has dawned that you can no longer reciprocate their unfollow via an app.


If you’ve ever played the follow-4-follow game, you may need to brace yourself for a shock when starting the task of visiting those profile pages. The first thing I noticed was that I’d followed not only some truly appalling spammers, but also some pretty toxic people. And I hadn’t followed anyone automatically. When following new accounts I’d checked the bio and most recent tweet at the very least, but clearly, that was not enough.

What I hadn’t bargained for, was people who were using retweets and aggregate as a shield for their spam, or who were simply stealing tweets from other people. These were things that had increasingly come to concern me over time – especially considering Twitter’s own rising intolerance of spam. I didn’t want to be damned by association with spam accounts. Hence my decision to laboriously look at every profile page I was following, and dump all the crap.


My initial intention was not to improve my follow ratio. I just wanted rid of all the rubbish, and I was prepared for a big drop in followers. But I did figure, in the light of the manage app purge, that some of the spammers with high friend counts would fail to spot that I’d unfollowed them. I didn’t know how many. I expected a lot to unfollow. I was wrong.

On one of my accounts I unfollowed about 240 active spam profiles with the loss of about 30 followers – some of whom were just suspensions or deactivations, and so would have been lost anyway. At a guess, I’d say about 20 of the spam accounts I unfollowed, actually unfollowed me back. That’s about 8%. Much, much lower than expected, and much, much lower than the percentage of inactives that would typically unfollow back.


So how did I go about indentifying spam accounts? I won’t deny there was an element of intuition at the end of the day, but here’s the kind of thing I looked for when deciding which accounts to unfollow…

  • High friend count – over 10,000. But I would consider this alongside other factors. A five digit friend count is often a sign of spam-following, but it doesn’t necessarily indicate a low quality account or an inattentive user, so I’d want to cross reference it.
  • Use of custom link shorteners. That’s,,, etc. These appear instead of the actual website links, both in tweets and in the bio area. They’re a very good indication that the person running the account is primarily focused on driving traffic somewhere else. People like this are generally not there to listen to you. They’re there to get you to listen to them and ultimately make money off you.
  • Repetitive timeline formula. Retweet, attributed quote, link… retweet, attributed quote, link… retweet, attributed quote, link… That type of consistent, repetitive pattern is almost certain to mean the account is not being managed by a human. Again, no point in trying to keep hold of their follow if they’re not even there.
  • Overwhelming dominance of Facebook/Instagram or other shares. These are normally Facebook or Instagram users who are simply auto-sharing everything to Twitter. Once more, the likelihood of them seeing your tweets is close to nil.
  • Multiple auto tweets in a row. Stuff like horoscope readings, number of followers gained/lost, etc. Even one of these is not a great sign, but if you’re seeing a whole line of them in a row, it means the user doesn’t even give a stuff about their own timeline. So the chance of them giving a stuff about yours is literally zero.
  • High followers count but low engagement totals. If someone has 100K followers, but no one Likes their tweets, you can probably put them down as a spammer without even looking any further. Check that you’re looking at their general tweets and not their specific replies though. Replies normally do only get one or two Likes, even on large accounts.
  • Replies to multiple accounts with the same message. Especially if the message is self-promotional, just bin accounts like this. They’re more likely to do you harm than good.
  • “Blue” tweets. “Blue” tweets are tweets in which everything is a link of some sort (based on Twitter’s default link colour being blue). These tweets are typically loaded up with a barrage of @usernames and hashtags. No plain text. If all or most tweets on the user’s timeline are collections of @usernames and hashtags, it’s just a spam page.
  • The news. Relentless un-themed links to news articles, with no replies in between, are a sign of a spammer. Stuff like this is easy for bots to harvest. As with most spam practices, the repetition gives it away.
  • – via @username. “Via” means it’s come from someone else. Typically, accounts tweeting a high quota of “- via @username” will be automated. It’s another of many techniques spammers use to pad out their feeds. Take all of this zero-effort, secondhand blab away, and all you’d see is self-promo. That’s why the secondhand blab is there.
  • Commercial bio. The bio is either full of business buzzwords like “solutions” and “strategy”, or it specifically refers to goods/services for sale. Business users hardly ever read their timelines, because that’s not why they’re on Twitter. I decided to unfollow people with a business bio by default, and everyone who identified themselves by name as a business. Only if their account had some other mitigating feature would I keep my finger off the unfollow button.
  • Too much great content. 20 brilliant jokes a day – every day? 15 brilliant landscape pictures a day – every day? That’s almost inevitably stolen content, because no one can produce brilliance in that kind of volume, with that kind of consistency. It takes a second to select the text in a tweet and Google it. And that will immediately tell you if the work is stolen. People who tweet secondhand content without context are spammers, and should not be supported. But beyond that, if they’re mutuals and they ever do see one of your wonderfully creative tweets, what do you think they’ll do? Like it, retweet it, or steal it? You know the answer to that.


The fundamental question I asked myself when re-evaluating each account was: “How much does it matter if this person unfollows me?” If the answer was that it wouldn’t matter at all beyond the numerical value, I could unfollow that user. I didn’t really want my followers totals to drop, but in real terms it makes no difference whether people like that are following me or not. You have to go into it with that attitude.

Unfollowing spam accounts didn’t turn out to be quite the chore I imagined it to be. I learned a lot about the kind of tricks spammers were playing, and about my own stupidity in following users without sufficient investigation. I haven’t yet finished the task on all of my accounts, but I’ve improved things a lot, and most of all, what I’ve seen has taught me to be much, much more careful who I follow in future.