Enter a search term such as “mobile analytics” or browse our content using the filters above.
That’s not only a poor Scrabble score but we also couldn’t find any results matching
Check your spelling or try broadening your search.
Sorry about this, there is a problem with our search at the moment.
Please try again later.
Not all Twitter followers are created equal.
Chasing large numbers for their own sake is a very dangerous game that a lot of brands still fall into, but unless you’re doing customer service there’s often no real point in following back everyone who follows you.
That isn’t to say that there’s no value in a large audience. Having a thousand people who really do like your brand and who talk about you has all sorts of benefits, from increasing lead potential to good old fashioned brand lift. Buzz really does count in many cases.
With this in mind, how do you sort the wheat from the chaff, and how do you make sure your important followers stick with you? I’ve been looking into the levels of churn on our own accounts recently, and thought it would be worth sharing my findings and start thinking about ways to address this...
How do we identify churn?
If you have access to Twitter ads, then you’ll find a section marked ‘TimeLine activity’ in your analytics. Here’s a recent snapshot of ours:
As you can see, the peaks and troughs of unfollows correlate very closely to the amount of new followers we receive.
This leads me to believe that people aren’t unfollowing because we’re spamming them into submission. We do send a fair amount of Tweets, but generally speaking it isn’t more than one or two an hour, so unless you only follow about ten people, we shouldn’t be flooding your stream.
This means that it’s probably a case of people unfollowing based on the content of a single tweet, which seems slightly odd to me, but is understandable as it matches follows. It feels as though:
- New users tend to follow us because of the content of a single tweet – probably because they’ve seen it retweeted.
- They unfollow for the same reason (?)
It follows that our followers are, generally speaking, not bots. Unfollowing based on specific content suggests an intelligent action.
We often publish niche content, occasionally in batches if it is in support of one of our reports. So perhaps if these people have an interest in Responsive Design, or Content Marketing, then they follow us on a day when we’ve mentioned these subjects more than once on the blog.
Once they realise we actually talk about a wide range of marketing subjects they decide to unfollow, which is a shame really, because we’ve probably got some content that might be useful to them.
I’m not arrogant enough to suggest we’re immune from bots entirely though, so let’s take a closer look at those figures.
For this I’m using the pro version of Followerwonk, which will provide you with a churn graph showing activity up to three months old:
Again, we see that fairly close correlation with follows and unfollows. So, our most popular content is... also our most unpopular.
Or at least polarising. This is probably a good thing – at least people aren’t bored – but can we find out who the individuals are who are unfollowing?
Who's leaving you?
There are two ways to discover this.
As mentioned I’m using followerwonk so I can easily pull out csv lists of follows and unfollows and compare them week on week.. Hooray for being lazy.
There a few services like Just Unfollow, which will send you updates as well, but often the information they send is incomplete. If you’re on a tight budget then use your emails from Twitter.
Head to your Twitter settings and make sure you’re receiving emails for new followers:
In the past, this point would see you fiddling about setting up RSS feeds from email for an hour. Now, you can just use If This Then That. Set your Twitter emails to forward to IFTTT, and have the service send an RSS update every time you get a new email. Feed the RSS into a Google doc.
To do this just add: =ImportFeed("https://econsultancy.com/uk/blog.atom",,True) to cell A1 in your sheet. In this example I've used the Econsultancy blog rss feed, you'll need to change to the RSS you've set up.
The sheet should now update each time a new email is sent to your rss. If you want more information or need to split the information out, Google has more info on this here.
Now you’ll just need to download and concatenate that sheet each week to see which @names have vanished. As we’re dealing in numbers, I’d order them according to their number of followers and do a bit of manual searching to see if they are quality followers or complete spammers:
Names have been obscured in this image. Except for Richard Branson, because come back Sir Rich we love you (and this has nothing to do with me wanting cheap flights).
And that’s it – you now know who’s leaving you, and (roughly) when. If your loss numbers exceed your gains or are comprised entirely of important customers then you’ve got some work to do.
Better content, better communication
Minimising this is largely a case of optimising your content.
Look at your follows and unfollows daily and identify when somebody followed and left. Was there a correlation related to content?
Check their profile; Are they still active on Twitter? Have they changed career, or moved to a new location? Look for reasons they may have left you – and at this point you might just need to resign yourself to the fact that some people like to mix it up and keep it fresh on Twitter.
Looking at our lists I can see that we’ve got a number of niche followers who may be quite valuable, so we’ll need to run some deeper investigations here.
If they were subscribers we can see if they made full use of our products, and if not then it might be possible to contact them. Ask people why they left, but do it in a sensitive manner, remember that there’s no tone on Twitter, so it might be time to turn to email so that you can conduct better analysis.
Are these the sort of churn levels you’ve seen? Given Econsultancy's position as a publisher I’d assume that ours were more volatile than many other accounts, so would love to hear any example data you might have, and do please share your tips on limiting churn in the comments.