When the algorithm steps back: ocean communications on Bluesky

Bluesky has been attracting quite a bit of attention from the research community of late.

Since Elon Musk purchased what was then called Twitter and the growth of an apparently hostile environment to science and scientists, many researchers sought out alternatives - like Bluesky. Recent research (that hasn’t been peer-reviewed) suggests that Bluesky posts featuring “scholarly articles” garner higher levels of interaction than they used to on X.

“Our study shows that scholarly posts on Bluesky are not only more original but also receive stronger engagement than on X, already making it a credible platform for science communication,” Er-Te Zheng, one of the study’s co-authors, said.

Zheng went on to say something else that gave me pause…

“What strikes me personally is how much ‘purer’ the scientific discussion feels on Bluesky: my feed is almost entirely filled with researchers and journals, unlike the more mixed content on X.”

Honestly, this has not entirely been my experience.

I think part of the reason comes down to how content is surfaced on Bluesky. And how content is surfaced matters for communication efforts.

The highs and lows of an algorithm

For those not familiar with Bluesky, it’s a microblogging platform that looks and feels a lot like X. It has a few different feed types - ways to see posts. I’m going to focus on the “following” feed as that is the default - and most commonly used.

As the name suggests, the following feed shares posts from people you follow in chronological order. That’s right - no algorithm stepping in to decide what you might find interesting, relevant, or indeed react to. Just posts from your followers, ordered from newest to oldest.

At first, I loved the idea of an algorithm-free feed. I saw many of my colleagues talk about how their feeds on X were flooded with what they felt were irrelevant posts, hate speech, and more. Even though I never experienced that, it was not something that I wanted. Chronological timelines seemed a win!

Except it hasn’t entirely been.

The thing about chronological order is that you get everything from your followers. And if you, say, follow a lot of marine scientists who happen to be based in the USA and want to share their thoughts about the current political situation there, then your follower feed can become overwhelmed with US politics.

Then say you follow other ocean scientists in countries like Australia, New Zealand, or Japan, but you’re based in Europe. Want to see their posts? Well, you’d best be checking the feed when those people are most likely to be posting, otherwise you’ll be doing a lot of scrolling down to reach them.

For all its faults, when I used Twitter (I stopped using it a long time ago), its algorithm usually showed me posts I was interested in, regardless of when they were posted. Ocean science, policy updates, conference threads, and research papers would often appear in my feed even if I didn’t follow the author, simply because they were gaining traction or intersecting with topics I engaged with. LinkedIn is similar - and it even shows me posts from a few weeks back that I missed (I particularly like this feature). The algorithms don’t always get it right, but the feed is usually relevant.

There are many reasons why each of our experiences with algorithms might be different. Leaving aside Musk’s manipulation of the algorithm to push through certain types of posts and send others to the abyss, algorithms shape our social media experiences based on our individual behaviours and the feedback loops we create through our interactions. Every like, share, pause, or scroll trains the algorithm on what we want to see - or at least what captures our attention (hence click bait, rage bait…). On Twitter, I probably didn’t engage with news from the US, so the algorithm didn’t show these types of posts to me. I’d say that my location outside the US also helped here, but for some reason, X thinks my dormant account is based in the US.

This isn’t to say that Twitter/X’s algorithms are fantastic. All algorithms are created by people who can unwittingly (or sometimes wittingly) bake their own biases into their design. LinkedIn, for example, has recently received some criticism for their algorithm, which some say is biased against women. And with X, we can see how an algorithm can be manipulated to push whatever agenda those in control of these platforms want.

And there is an upside to the chronological feed. Sometimes, I’ve discovered cool things I perhaps wouldn’t have with an algorithm curating my feed for me. That includes accounts with small followings. Their posts don’t lose out to accounts with huge followings or huge likes. Everything is fair game in a chronological feed.

What does a chronological feed mean for ocean communications?

Without an algorithm working behind the scenes to push your ocean posts, perhaps we need to approach how we post on platforms with chronological feeds differently.

Timing becomes critical.

The time a post goes out has always mattered to some extent, but timing is especially important for chronological feeds. If the people you want to target are clustered in the same or similar time zones, your job is much easier than if you want your post to reach people spread over very different time zones. A single post about a new research paper or policy development could easily get buried within hours. More than other platforms, if you want to maximise your reach (especially across time zones), you might want to post about your research paper multiple times - and at very different times.

The nature of "successful" posts changes

On algorithm-driven platforms, posts that spark strong reactions - controversy, outrage, excitement - often get pushed to more people. On Bluesky's chronological feed, that pressure is off. Your post doesn't need to be engineered for virality; it just needs to resonate with the people who follow you. This might actually be liberating for science communication, where nuance can be sacrificed for engagement metrics.

Building and maintaining networks matters more

On algorithm-driven platforms, you might reach people who don't follow you through shares and recommendations. On Bluesky, your discoverability is reduced.

Bluesky does have a Discover feed, which is powered by an algorithm that shows what it considers to be posts you might like to see from accounts you’re not following - and accounts you are. But it is a separate feed - and not the default. This means that forming genuine connections with others on the platform may be particularly important for helping others find you. If you comment on someone’s post, their followers may see your comment. If someone comments on your post, their followers may see it.

This means investing time in following the right people, engaging authentically, and curating your connections becomes part of the communication strategy itself. Yes, it’s a bit more work, but it does get back to something that’s arguably being lost from social media in general - being social.

Chronological feeds don’t make ocean science communication harder, but they do change how it is seen. Discovery becomes more dependent on existing networks, timing, and direct interaction, rather than on systems designed to surface content beyond immediate circles. Bluesky’s chronological feed isn’t better or worse than any platform that uses an algorithm to curate users’ feeds. It simply requires a more intentional approach.