Have you ever opened Facebook and wondered why you are seeing a political post from someone you never followed, never liked, and may not even be interested in?
I have experienced this myself. Sometimes a political party’s post appears on my Facebook feed even though I have never liked its page, followed it, or interacted with its content. At the same time, there are people and pages I deliberately follow whose updates I rarely see.
That can feel strange.
But it reveals something important about how social media actually works.
Facebook is not simply showing you posts from the people and pages you have chosen to follow. Your feed is ranked by algorithms that decide which content should receive your attention.
And that gives the platform considerable power over what you see.
Meta has publicly explained that Facebook starts with a large pool of potential posts and uses machine-learning systems to predict which ones a particular user is most likely to find relevant or interact with. The company says the system considers signals including what people follow, like and interact with, as well as characteristics of the post itself and activity surrounding it.
In other words, your Facebook feed is not a complete picture of what is happening on Facebook.
It is a filtered and ranked version of it.
Why do I see posts I never liked?
This is one of the most confusing parts of modern social media. You might think that if you never liked a political party’s page, Facebook should never show you its posts.
But that isn’t how recommendation systems work.
Facebook can recommend content from accounts or pages you do not follow. The platform has publicly described ranking and recommendation systems that select content based on predicted relevance and user behaviour.
Imagine that several of your friends are discussing a political post. One person shares it. Another comments. Someone else reacts. The post starts receiving thousands of interactions. You may then become a candidate for seeing that content even though you never followed the original page.
Your previous behaviour can also matter.
If you regularly read Nepali political news, watch political videos or interact with discussions about government and elections, the system can learn that political content is relevant to you.
That doesn’t mean you specifically asked to see that particular political party.
It means the system may predict that you are likely to interact with political content.
That distinction is extremely important.
Facebook doesn’t show everyone the same internet
Two people sitting next to each other can open Facebook at the same time and see completely different feeds.
One person might see football.
Another might see politics.
Another might see entertainment.
Another might see technology.
Another might see posts from a particular political movement.
The underlying platform is the same.
The information environment is not.
Facebook has explained that its ranking system takes a large pool of potential stories and narrows it down to the posts it predicts will be most relevant to an individual user.
That means your feed is partly a reflection of what the algorithm has learned about you.
But there is another important part of the equation.
It also reflects what other people are doing.
What happens when thousands of people react to one political post?
Suppose a politician publishes a statement. Within minutes, thousands of people react to it. Some agree. Some disagree. Some share it. Others comment just to criticize it. News pages repost it. Influencers discuss it. People tag their friends.
Suddenly, the original post has generated an enormous amount of activity.
The platform can see that activity.
Facebook’s own technical explanation of Feed ranking says its systems consider information about how many people have interacted with a post and use machine-learning models to predict how valuable a story may be to a particular user.
That creates a potential feedback loop:
More engagement → more signals → greater potential distribution → more people see it → more engagement.
This is one reason certain political posts can suddenly appear everywhere.
It does not necessarily require someone at Facebook to manually decide that everyone should see the post.
The system can amplify content because the content is producing strong signals.
Nepal gives us a real example
This isn’t merely theoretical.
In February 2026, The Kathmandu Post published an analysis of 4,754 posts from 24 prominent Facebook pages in Nepal during the month before the parliamentary election.
The pages together had millions of followers.
The newspaper found that content related to the Rastriya Swatantra Party appeared in 54.1 percent of the political posts it examined — substantially more than content associated with the Nepali Congress and other major parties. The analysis also found that posts involving the party’s prominent leaders were particularly common on several large pages.
That is a significant finding.
It means the political content people encountered on those large pages was not evenly distributed among all political parties. Some political content was much more visible than other political content. But there is an important question.
Does that prove Facebook deliberately programmed its algorithm to support one party?
No.
The analysis itself says its methodology can identify patterns in published content but cannot establish intent, coordination or whether page administrators had connections with political parties.
That limitation is important.
There is evidence of unequal political visibility.
That is not automatically evidence of secret political control.
But engagement can create a very real advantage
There is another piece of evidence that helps explain what happened.
The Kathmandu Post reported that one major Nepali Facebook page’s founder said posts concerning the prominent political figure associated with the RSP consistently generated more user engagement than posts about other candidates and parties.
The explanation was essentially straightforward: the page posted political content about many candidates, but content about that particular figure attracted more reactions, comments and shares.
If that is happening across a large network of pages, the effect can become enormous.
A politician’s post attracts more attention.
Pages repost it.
People comment on it.
Others share it.
The content reaches more people.
Those people generate additional engagement.
And the cycle continues.
This can produce a powerful political visibility advantage even without evidence of a secret agreement between the platform and the politician.
But we should not pretend algorithms are neutral machines
At the same time, it would be wrong to say that algorithms are completely neutral. They aren’t. Humans design them. Companies decide what objectives their systems should pursue. They decide which signals matter. They decide how recommendation systems should respond to different forms of user behaviour.
Meta itself describes its ranking systems as machine-learning models designed to predict what users will find relevant and engaging.
Those design choices inevitably affect what people see.
If a system places significant importance on engagement, content that produces strong reactions can gain an advantage.
And political content is particularly capable of producing strong reactions.
People argue about politics.
They become angry.
They become excited.
They defend their preferred politicians.
They attack opponents.
They share controversial statements.
All of that creates engagement.
So a system optimized partly around engagement can end up amplifying political content without anyone explicitly programming it to support a particular political side.
That is an important distinction.
A post can be amplified without being “approved”
There is another misconception worth clearing up. If Facebook recommends a political post to you, that doesn’t necessarily mean Facebook agrees with it. The recommendation system isn’t necessarily making a political judgment that the post is correct. It may simply predict that the content is relevant to you or likely to generate interaction. This is why you can sometimes see content you strongly disagree with. In fact, your disagreement may itself produce engagement. You might comment:
“How can anyone believe this?”
You might share the post with friends to criticize it.
You might watch the entire video because you are angry.
From a human perspective, you are opposing the content.
From an engagement system’s perspective, you have interacted with it.
That is one of the fundamental complications of algorithmic media.
This is how political bubbles are created
Imagine you interact with a particular type of political content. The system learns from that behaviour. It recommends similar material. You interact again. The system receives another signal. More similar content appears. Eventually, your feed can become heavily concentrated around a particular political viewpoint.
At the same time, another user may behave differently. That person interacts with opposing political content.
Their recommendations develop differently.
Both people can live in the same city, use the same application and have completely different impressions of what the country is talking about.
One person thinks a political movement is everywhere.
Another barely encounters it.
Both can honestly describe what they see on Facebook.
Neither necessarily sees the whole picture.
And this is where the idea of “manipulation” becomes complicated
Does social media manipulate people?
There is certainly evidence that recommendation systems can influence what information users encounter and how much attention different pieces of content receive.
But there are different forms of manipulation.
A platform can manipulate distribution through its ranking decisions.
A political campaign can manipulate attention through coordinated messaging.
A group can artificially generate engagement.
A government can conduct an influence operation.
Users can unintentionally amplify propaganda.
And algorithms can amplify all of those activities.
These should not be treated as one identical thing.
There is documented evidence of governments using social media for influence
This isn’t just a theoretical possibility.
A Reuters investigation reported in 2024 that the CIA conducted a covert influence operation targeting Chinese public opinion, using fake online identities to spread narratives critical of the Chinese government and leak information to foreign media.
Reuters reported that the operation began during the Trump administration in 2019 and was aimed at countering China’s growing influence. The report was based on interviews with former U.S. officials familiar with the program; the CIA did not confirm the operation publicly. Reuters also reported that the program’s impact could not be independently determined.
That is an important example because it demonstrates something that should not be dismissed as a conspiracy theory:
Governments have historically used information operations and social-media activity as instruments of geopolitical influence.
But it also demonstrates why we need to be precise.
Evidence that one government conducted a particular covert influence operation does not prove that every social-media algorithm is controlled by that government.
Those are separate claims.
The problem is bigger than the United States
The discussion is often framed as:
“Are American companies controlling what the world sees?”
That question is understandable, but it is too simple.
Social media is global.
The United States has major technology companies.
China has major technology companies.
Russia has conducted documented information operations.
European governments regulate platforms.
Other governments have attempted to influence online narratives.
Political organisations everywhere compete for attention.
The information environment is therefore not controlled by one country.
It is a global competition involving governments, corporations, political movements, advertisers, media organisations and ordinary users.
TikTok makes the argument even more complicated
Consider TikTok.
TikTok is associated with ByteDance, a Chinese company.
The United States has raised national-security concerns about TikTok and has taken major legal and regulatory steps concerning its ownership and operation.
If every major social-media platform were simply controlled by the United States, this situation would make little sense.
The reality is much more complicated.
Different governments are concerned about different platforms.
Different technology companies have different relationships with governments.
And governments themselves sometimes compete over control of data, technology and information.
So why does one political party’s content sometimes seem to be everywhere?
There may be several reasons operating simultaneously.
The party may have a highly active online audience.
Its leaders may publish frequently.
Its supporters may share content extensively.
Its opponents may also generate engagement by criticizing it.
Large pages may discover that posts about the party perform better.
News organisations may cover its statements more frequently.
The algorithm may identify high engagement.
Recommendation systems may then distribute related content to more people.
The result can be a dramatic increase in visibility.
None of these explanations should automatically be dismissed.
But neither should we automatically jump to the conclusion that Facebook is secretly working for a particular political party.
The evidence needs to tell us which explanation is supported.
The Ronaldo example tells us something important
This is why the Cristiano Ronaldo example is useful. You can follow a person with hundreds of millions of followers and still miss their posts. Meanwhile, you can see repeated posts from an account with a fraction of that audience.
Why?
Because social media is not fundamentally organized around follower counts. It is organized around ranking and recommendation. The system is trying to decide what should occupy your limited attention.
And political content can become extremely powerful when it consistently generates engagement.
The most important power is not deleting content
When people think about censorship, they usually imagine something being removed. But algorithmic systems create another form of power.
Visibility.
A post can remain online and technically be available to everyone. But if almost nobody is shown it, its practical reach may be tiny. Another post can be repeatedly recommended to millions of people. Both posts may technically be allowed.
Their real-world influence can nevertheless be completely different. This is why algorithms matter so much.
The question is not only:
“What is allowed to exist?”
It is also:
“What is being given the opportunity to be seen?”
What should users do?
The safest approach is not to assume that everything on Facebook is manipulated.
It is also not to assume that everything is neutral.
Instead, understand how the system works.
If you see a political post from a page you never followed, remember that recommendation systems can show you content based on predicted relevance and engagement.
If you see the same politician repeatedly, don’t automatically assume that Facebook is secretly supporting that politician.
But don’t assume the opposite either.
Look at the evidence.
Search for the original post.
Check multiple news sources.
Look at what different political groups are saying.
Pay attention to what you are not seeing.
And if something important matters to you, don’t rely entirely on your Facebook feed to tell you about it.
If this article helped you, please consider supporting our work. Every small contribution keeps Abijita.com independent and running.
Search for it yourself.
Conclusion
Social-media algorithms have enormous influence over the modern information environment.
That influence is real.
Facebook itself acknowledges that its systems rank and personalize the content users see.
Evidence from Nepal shows that political content associated with some movements and personalities has received substantially more visibility on major Facebook pages than content associated with other political forces.
There are also documented cases of governments using covert information operations through online platforms, including the Reuters-reported CIA operation targeting Chinese public opinion.
But these facts do not justify the simplistic claim that one government controls every social-media algorithm.
The reality is more complicated.
Algorithms can amplify human behaviour.
Companies control the systems that perform that amplification.
Political groups compete to take advantage of those systems.
Governments can attempt to influence the information environment.
And ordinary users can unintentionally become part of the distribution mechanism simply by liking, commenting, sharing or arguing.
That is perhaps the most important lesson.
You are not seeing the entire internet when you open Facebook. You are seeing a version of the internet that a recommendation system has selected for you.
And once we understand that, a much bigger question appears:
If an algorithm can decide what millions of people see repeatedly — and what millions of people barely see at all — how much power should that algorithm have over our information environment?
Sources
- Meta: How Facebook’s News Feed ranking works
- Meta Engineering: How machine learning powers Facebook News Feed ranking
- The Kathmandu Post: How Facebook’s algorithm is amplifying one party over all others
- The Kathmandu Post: Can Nepal’s Viral Mayor Run the Country?
- Reuters: Trump launched CIA covert influence operation against China
Disclaimer: This article is intended for general informational and educational purposes. It explains how social media recommendation and ranking systems can influence what users see across platforms such as Facebook, Instagram, TikTok, YouTube, and X. References to political content, government activity, or platform practices are based on publicly available information and reported sources. The article does not claim that any particular government secretly controls social media algorithms, nor does it endorse or oppose any political party, politician, or ideology. Social media systems change over time, and individual users may experience different recommendations based on their activity, interests, connections, and other signals.
