Twitter experiments with new ways to enable custom algorithm controls

July 12, 2022

Jayde Robertson

Ah - the algorithm. Love it - or hate it, it exists for a reason. We’re all susceptible to it, and as it evolves, its effects are felt from users through to companies.

When Elon Musk announced his $44 billion purchase of Twitter, he pledged to improve the social network by "making the algorithms open-source to promote trust," among other things. A TED talk earlier this year, Musk suggested that the algorithm that determines how tweets are promoted and demoted could be uploaded to the software hosting platform GitHub, making it available to people outside of the company. “People can look through it and say, ‘Oh, I see a problem here, I don’t agree with this,’” Musk said. “They can highlight issues and suggest changes, in the same way, that you update Linux or Signal.”

In actuality, cracking open Twitter to learn how it works would need much more than simply posting some code to GitHub. And demonstrating the existence—or lack thereof—of biases that may be subtle in nature and are dependent on a plethora of constantly changing factors may prove far more difficult than Musk suggests.

That said - of course - there has been a development in the Twitterverse.

What if you could choose which tweets you see when you open the app, with configurable choices to modify how the algorithm works and the order in which appropriate tweets are delivered based on your preferences?

Well - this may be on the cards, according to a new discovery by reverse engineering expert Jane Manchun Wong.

Custom timelines enabled by third-party tools? Sounds complicated - but interesting too. Non-Twitter platforms would appear to be able to construct their own algorithms, in contrast to Twitter's, allowing for new ways to explore the latest tweets.

Wong goes on to say that this might essentially allow users to "choose their own "algorithm," rather than being limited to Twitter's Home Timeline algorithm.

More algorithmic transparency was also a significant feature of Elon Musk's Twitter takeover push, which has been proposed in various versions in the past. Musk has recently challenged Twitter's current algorithm sorting, which he thinks is tampering with the user experience. According to Musk, the solution is to consider opening the algorithm, which would allow users to gain a greater understanding of the platform's systems, allowing them to make more educated decisions about their in-app interactions. Twitter has been experimenting with something similar through its 'Bluesky' initiative for some time, though the core idea of regular users having a better grasp of such systems may not be as simple in practice, as a level of code knowledge is likely considered necessary to fully grasp the complexities of such when laid bare.

An alternative would be to make Twitter's algorithmic parameters open-source, allowing developers to create new, bespoke algorithms that users may then pick from to personalize their tweet experience. As explained by Nathan Baschez: “For example, I’d want to try an algorithm that attempts to prioritize nuanced conversations about important topics. Maybe someone else would want algorithms to find mind-expanding threads, savage dunks, or thirst traps of hot new Snax.”

That appears to be what Twitter is aiming for with this latest tweak in the code, which, as Wong points out, alludes to other providers creating a marketplace for different algorithms within the app.

It has the potential to alter the whole Twitter experience as well as marketing strategies.

This appears to be the case, albeit it's too early to guess what 'content management tools' may entail for the overall tweet experience.

But that may be a significant shift — and with Elon on board, it'd be more likely to survive the (possible) takeover, making this a priority project.

There's a lot to think about here. For everyone, a completely separate Twitter feed. A marketplace where you may edit and update your chronology. A novel approach to learning about algorithmic amplification.

Twitter's next stage could look entirely different than it does now.

So…. what now?

With the introduction of customization algorithms, the world of marketing has altered forever. Traditional techniques took a broad view of the situation. Businesses would promote their greatest products and services. They would then hope. It's possible that it reached its intended audience, but it's also possible that it didn't. All of that has changed, and there is no turning back now. Internet consumers increasingly want to see content that is relevant to them and suited to their needs. The latest technique to sell to potential clients is through personalization. This effectively all relies on and comes down to  Data scientists and artificial intelligence.

Data about client behavior is collected by the algorithm. Then it offers customers the experience based on those data sets. The learning system then generates user-specific messages, promotions, and adverts. This translates to higher conversion rates and a better user experience. Most institutions today employ these marketing algorithms in their advertising campaigns.

Explaining the algorithm.

An algorithm is a set of mathematical rules that describe how a set of data behaves. Algorithms help preserve order in social media and aid in the ranking of search results and adverts.

There were roughly 4.5 billion internet users as of October 2021, with about 57% of the global population using social media. That's a huge task to keep track of and manage. This is why algorithms are so important when it comes to establishing the legitimacy and positioning of social media accounts and information.

Here’s what we think.

At this stage - it’s almost impossible to tell what a personalized algorithm would mean for marketers, but to speculate - it simply means that we need to again learn how to focus on personalization over personification. Simply put - it's crucial to remember that as Ed marketers - we are speaking to people and not organizations. People come with their own set of rules, preferences, and objections. In order to feature in their personalized algorithms, we need to fully understand what they want to see - and why they want to see it. This is where accurately identifying our target persona comes in handy! Once we know what our audience wants to see - what is relevant to them, then we will be better positioned to deliver that content - and appear on their timeline!

Are you a marketer curious to learn about the aspects that affect our industry?

Or do you want to tap into our expertise to sharpen your tactics? Look no further than our free downloadable resources. We work hard to research our content - and offer it to you - free of charge.

The power of social media is here to stay.

Master the power of the algorithm - download our free Ebook: understanding the algorithm, as we break down the digital jigsaw puzzle for you.