If you haven’t noticed, there are a heap more tweets being shoved into your Twitter timeline these days, either from people you don’t follow or from people you do follow, but content you wouldn’t normally see (like their replies to others). And there’s good reason for that – as per Twitter’s latest earnings results, putting heavier emphasis on showing people more relevant tweets is helping to boost engagement, despite overall audience numbers stalling.
Highlighting tweets based on your noted interests is making Twitter smarter, more attuned to each user – and therefore, more able to keep you on Twitter for longer.
The latest evolution of this is the introduction of algorithm-defined tweets, based on your noted interest, to the Explore tab in the mobile app. Now, when you scroll through Explore, you’ll see not only Trends and Moments, but a new listing of popular tweets under topic banners, which are being shown to you based on your activity.
Right now, the topics seem fairly broad – these are the matches I’m seeing, and they kind of relate to my interests. But more importantly, they also provide additional perspective on how Twitter’s algorithm works, and how sensitive it can be. Which some marketers are now looking to use to advantage.
For example, as you can see from the topics above, I’m being shown listing for ‘Science and Technology’, ‘NBA’ and ‘Music’. And for each of these, I can pretty confidently say why, what’s lead Twitter to showing me these topics.
In my case, my advantage is that I haven’t been highly active on my personal Twitter of late, so I haven’t given Twitter a heap to go on. The first listing is fairly obvious – I read a lot of science and tech related content, so it could infer this from the tweets I’ve clicked on. The other two are not things I’m highly engaged with, or actively participating in – but I have interacted with related content lately.
For the NBA, a couple of times in the last week I’ve searched for the latest news on the potential trade of Kyrie Irving. I follow, maybe, two NBA related accounts, and I check ESPN for the latest, maybe once a day. It’s not something that I’m highly engaged with – but you can see how Twitter is trying to work with my more recent interests to show me more content I might like.
For music, again, I follow hardly any music-related accounts, but I recently started following Bon Iver and Royal Blood – and they’re the last two accounts I’ve followed. So, again, Twitter’s trying to infer interest based on my most recent, significant indicators, being who I’ve chosen to follow.
What this shows, at least to some degree, is that Twitter’s algorithm:
- Focuses on recency – As opposed to the accounts you’re following or what, historically, you tweet about – which makes sense, if they want to get new users more active, they need to show them more relevant content based on their latest interests
- Can be heavily influenced by user actions – For example, you’d probably have difficulty knowing exactly why Facebook’s algorithm is showing you this or that, but as I recently found, you can change your Twitter timeline by simply selecting ‘I don’t like this’ on, maybe, two or three tweets
These aren’t definitive, of course, these are just my experiences, but based on this, and other research I’ve seen and conducted, it shows that Twitter’s algorithm is not as complex or robust as Facebook’s News Feed. And that, to some degree, leaves it open to manipulation.
Another thing we’ve noticed, through the Social Media Today handle, is that we’re getting tagged more in random conversations, which have nothing to do with us. Why? Because we have a large audience – and because Twitter’s algorithm is highlighting tweets to other users based on replies and mentions, it’s to the advantage of marketers to tag larger accounts to boost their exposure.
And while that’s generally spammy behavior, it does underline how the algorithm-defined feed is changing the Twitter marketing process – or, at least, adding new considerations into how to generate more exposure for your business on the platform.
So how can you take advantage?
1. Generate engagement with your tweets
Obviously, this is a fairly safe bet at any time, but the algorithm highlights tweets that are generating conversation and interaction – if there’s a lot of people in your audience talking about a specific subject, Twitter’s algorithm will assume you don’t want to be left out of the loop. The more replies and engagement you can inspire, the better.
2. Tag where relevant
Don’t, as I’ve noted here, just tag random large accounts. While I do suspect this is working to a degree, the main hope in doing so would be that these large accounts might give your tweet a Like, which would then ramp up its exposure. Tag at random and this won’t happen (and you could get reported for spam), but tag relevant people in your posts and you stand a better chance of boosting your tweet exposure by having them, ideally, like and reply, helping you show up higher in the feeds of their followers
3. Reply to and Like mentions
Another engagement factor, but as noted, you’ll have seen a lot more of those ‘Replying to…’ tweets in your feed or @user liked’
These will show up even if you’re not following the person replying – for example, I follow Mathew Ingram but not Jillian Jorgensen, yet I still saw this tweet. Given this, even basic interactions and response can help boost your tweet exposure.
As with any ‘tactics’, these measures can also be misused, and you’ll often see people pushing a little too hard to engage, like asking random questions to force interaction. It’s important to consider the balance of increased exposure with genuine engagement – is it better to have more people see your content or to have a genuine relationship with your audience?
There are specific weigh-ups you’ll need to make in each case, but as Twitter puts more emphasis on algorithm-identified tweets, it’s worth considering how that applies to your tactics, and how it can be used to best optimize your performance.