Experiment Finds That Questions Generate More Reach on Threads

Looking to get a better understanding of how the Threads algorithm works?

This might help.

Business Insider reporter Katie Notopoulos recently conducted an experiment in “rage baiting”, in which she posted a range of controversial queries on Threads designed to spark user responses.

The results, as you can see in this example, were significant.

Notopoulos’ posts sparked strong emotional reactions, which, as history shows us, drives more engagement. But the main impetus of the experiment was to see whether rage bait works as well on Threads, and what other elements the Threads algorithm may be focused on.

The key note from Notopoulos’ findings? Posts with a lot of comments look to be more heavily weighted when it comes to what shows up in your Threads feed. That’s over re-shares and Likes, with the experiment seemingly suggesting that the Threads algorithm is geared towards incentivizing discussion over everything else.

Which makes sense. Meta has repeatedly noted that it wants Threads to be a more positive version of what Twitter had become in the end, and what X now is, where Likes and re-posts still drive significant reach benefits.

By focusing on posts that prompt actual discussion, that, at least in theory, could get more people talking, and sharing their own thoughts, but then again, the rage bait cycle is also a significant part of the reason why fewer people are now sharing their thoughts in public, for fear of being shot down and virtually trampled by online mobs.

But there is also another reason why Meta may be focused on discussion on Threads over other forms of engagement.

Meta, like various other companies, is currently hard at work developing its own generative AI models, with its AI chatbot now a key focus of its broader push to bring AI to the masses.

In order to fuel AI bots, developers need human inputs, and ideally, human-generated conversations that include direct questions and answers.

That’s why Reddit is considered such a valuable data source for large language models, because much of Reddit’s content is a question posts, which then spark a range of actual, human answers, replies written by real humans that can train LLMs in how people engage.

Meta needs that too, and as such, it would make sense that the company would be looking to angle its algorithms towards amplifying question posts, in order to spark more user responses, that it can then build into its language models.

As such, I’m not sure that it’s rage bait that’s the main impetus for engagement here, so much as questions themselves, and prompting responses, ideally to commonly asked questions.

So, if you’re looking to boost your Threads presence, asking questions is likely key, as the algorithm does seem to favor this type of engagement.

You can make them divisive queries if you really want to trigger responses, but I’d hazard a guess that all questions will get a bit of a boost in the system.