Within Ranking
When People Like You Shape Your Feed
Recommendation systems can use patterns from people like you, making a narrow cluster of behaviour look like a wider reality.
On this page
- How lookalike behaviour enters recommendations
- Why personal feeds can feel like consensus
- How to test whether a claim travels beyond your cluster
Page outline Jump by section
Introduction
A personalised feed does not only learn from you. It often learns from people who behave like you. Recommendation systems on major platforms use patterns from groups of users with similar viewing, clicking, liking or sharing habits to predict what you may want next. As a result, a feed can reflect the behaviour of a particular cluster of users rather than the full range of what the wider public is discussing. [The Washington Post]washingtonpost.comThe Washington Post Are you in Tik Tok's cat niche?What 121,000 videos reveal.December 10, 2025 — The Washington Post's investigation into TikTok’s recommendation algorithm provides an in…
This matters for critical thinking because repeated exposure creates a sense of normality. If a recommendation system keeps showing claims, opinions, jokes, products or concerns that are popular within a lookalike group, those topics can start to feel common, mainstream or widely accepted even when they are concentrated within a relatively narrow audience. Understanding this mechanism helps explain why different people can open the same app and come away with very different impressions of what “everyone is talking about”. [The Washington Post]washingtonpost.comThe Washington Post Are you in Tik Tok's cat niche?What 121,000 videos reveal.December 10, 2025 — The Washington Post's investigation into TikTok’s recommendation algorithm provides an in…
How Lookalike Behaviour Enters Recommendations
Recommendation systems often rely on a technique known as collaborative filtering. Instead of analysing only the content itself, the system studies patterns of behaviour across many users. If people who watched, liked or shared one item also tended to engage with another, the platform learns that those items are connected.
TikTok openly states that recommendations can be influenced by users with similar interests. Investigations into TikTok’s recommendation patterns have found that content is organised into behavioural clusters, where videos viewed by similar audiences end up connected even when they are not obviously related by keywords or hashtags. The result is a recommendation map built partly from what people like you tend to consume. [The Washington Post]washingtonpost.comThe Washington Post Are you in Tik Tok's cat niche?What 121,000 videos reveal.December 10, 2025 — The Washington Post's investigation into TikTok’s recommendation algorithm provides an in…
A similar principle appears across recommendation systems more broadly. YouTube has explained that recommendations are designed to match viewers with content they are likely to watch and enjoy, while analyses of the platform describe recommendation models that consider what similar viewers watch alongside a user’s own history. [blog.youtube]blog.youtubeon youtubes recommendation systemOn YouTube's recommendation system15 Sept 2021 — Our recommendation system is built on the simple principle of helping people find the vi…
The key point is that your feed is not merely a reflection of your preferences. It is also influenced by the behaviour of a statistical neighbourhood: other users whose actions resemble yours.
Why Personal Feeds Can Feel Like Consensus
Humans naturally estimate what is normal from what they encounter repeatedly. A recommendation system does not need to tell users that an idea is popular. It only needs to show that idea often.
Imagine a user who watches several videos about a niche health trend. The platform may identify thousands of other users with similar behaviour and begin recommending additional content that circulated successfully within that group. Over time, the user sees more creators discussing the same trend, more comments referring to it and more examples that appear to confirm its importance.
The impression created is not necessarily false. The trend may genuinely be popular among that cluster. The problem is scale. A user can easily mistake popularity within a recommendation cluster for popularity across society as a whole.
Researchers and commentators often describe related phenomena using terms such as filter bubbles or echo chambers, although scholars continue to debate how strong these effects are in practice. Evidence suggests that recommendation systems can narrow exposure in some contexts, but the size and consequences of those effects vary across platforms and situations. Some studies find meaningful clustering and reinforcement, while others find weaker impacts than public discussions sometimes assume. [arXiv+2arXiv]arxiv.orgOpen source on arxiv.org.
For critical thinkers, the important lesson is not that every personalised feed traps users in isolation. It is that a feed should not automatically be treated as a representative sample of public opinion.
A Concrete Example: The Many Sides of One Platform
Large recommendation systems often divide users into countless overlapping interest communities.
A large-scale analysis of TikTok viewing patterns found distinct clusters around topics such as relationships, mental health, literature, gaming, engineering, fandoms and humour. Users who spend time in one cluster tend to encounter more material connected to that cluster, while vast areas of the platform remain largely invisible to them. The investigation also found notable differences in the types of content shown to different groups of users. [The Washington Post]washingtonpost.comThe Washington Post Are you in Tik Tok's cat niche?What 121,000 videos reveal.December 10, 2025 — The Washington Post's investigation into TikTok’s recommendation algorithm provides an in…
This creates a common misunderstanding. Two users may both spend hours on the same platform and honestly believe they are seeing what the platform is “mostly about”. Yet each is observing a different slice shaped by the behaviour of similar users.
When people compare experiences, they are sometimes surprised that others rarely encounter topics that seem omnipresent in their own feeds. The surprise itself is evidence of how powerful behavioural clustering can be.
Why Similar-User Recommendations Can Be Useful
The mechanism is not inherently harmful.
Recommendations based on similar users help people discover niche interests, specialist communities, educational content and creators they would probably never find through simple chronological feeds. Many users value these systems precisely because they surface material that feels personally relevant. Platforms rely on this ability to reduce information overload and help users find content they enjoy. [blog.youtube]blog.youtubeon youtubes recommendation systemOn YouTube's recommendation system15 Sept 2021 — Our recommendation system is built on the simple principle of helping people find the vi…
There is also evidence that recommendation systems do not always increase fragmentation. Some research suggests that recommendations can create shared patterns of consumption across different users or broaden individual exposure in certain contexts. The relationship between personalisation and social division is therefore more complicated than a simple story of isolation. [arXiv]arxiv.orgDeconstructing the Filter Bubble: User Decision-Making and Recommender SystemsApril 23, 2019…
The critical-thinking challenge is not to reject recommendations altogether. It is to recognise their limits when judging how widespread an idea, belief or trend really is.
How to Test Whether a Claim Travels Beyond Your Cluster
When a topic seems unavoidable in your feed, it is worth asking whether the visibility comes from broad public attention or from recommendation dynamics within a lookalike group.
Several practical checks can help:
- Search outside your feed. Look for independent reporting, search results and discussions that are not generated solely by your recommendation stream.
- Compare across platforms. A topic that appears everywhere on one platform but nowhere else may be benefiting from platform-specific recommendation patterns.
- Check audience diversity. Are different communities discussing the issue, or only a narrow set of creators and followers?
- Ask who is absent. Which perspectives, demographics or regions are not represented in the content you are seeing?
- Use recommendation controls deliberately. Following new topics, searching for unfamiliar viewpoints and providing feedback signals can sometimes broaden what appears in future recommendations. [The Washington Post+2blog.youtube]washingtonpost.comThe Washington Post Are you in Tik Tok's cat niche?What 121,000 videos reveal.December 10, 2025 — The Washington Post's investigation into TikTok’s recommendation algorithm provides an in…
These checks do not prove whether a claim is true. They answer a different question: whether the claim has escaped the boundaries of the cluster that helped deliver it to you.
The Critical-Thinking Takeaway
Recommendation systems increasingly act as social mirrors built from groups of similar users. The reflection they provide can be useful, relevant and engaging, but it is not necessarily a picture of the wider public conversation. A feed may tell you a great deal about what people like you are watching, discussing or believing. It may tell you much less about everyone else.
The habit that matters is simple: when something feels normal because it appears repeatedly in your feed, pause and ask whether you are seeing society at large or the preferences of a lookalike crowd. That distinction is often invisible inside the feed itself, yet it can change how confidently you interpret what seems popular, accepted or true.
Amazon book picks
Further Reading
Books and field guides related to When People Like You Shape Your Feed. Use these as the next step if you want deeper reading beyond the article.
The Righteous Mind
First published 2012. Subjects: Political psychology, Social psychology, Ethics, Religious Psychology, nyt:combined-print-and-e-book-nonf...
The Filter Bubble
First published 2011. Subjects: World Wide Web, Invisible Web, Information organization, Social aspects, Subject access.
eBay marketplace picks
Marketplace Samples
Example marketplace items related to this page. Use the search link to explore similar finds on eBay.
Endnotes
-
Source: blog.youtube
Title: on youtubes recommendation system
Link: https://blog.youtube/inside-youtube/on-youtubes-recommendation-system/Source snippet
On YouTube's recommendation system15 Sept 2021 — Our recommendation system is built on the simple principle of helping people find the vi...
-
Source: arxiv.org
Link: https://arxiv.org/abs/2307.01221 -
Source: arxiv.org
Title: arXiv Understanding Filter Bubbles and Polarization in Social Networks
Link: https://arxiv.org/abs/1906.08772 -
Source: arxiv.org
Link: https://arxiv.org/abs/1904.10527Source snippet
Deconstructing the Filter Bubble: User Decision-Making and Recommender SystemsApril 23, 2019...
Published: April 23, 2019
-
Source: youtube.com
Link: https://www.youtube.com/watch?v=uCspNBRLvw4Source snippet
If YouTube is not recommending your videos, try these...These features that I'm going to show you how you can use on your videos will he...
-
Source: youtube.com
Link: https://www.youtube.com/watch?v=P0SMq2DypU0Source snippet
"Obliteration (Official Video)From the album The Algorithm and The Algorithm: Ultra Edition. Animation by Atanas Shopski. [https://shopskia..."](https://shopskia...")...
-
Source: youtube.com
Title: Down the You Tube Rabbit Hole
Link: https://www.youtube.com/watch?v=bl1zM7c0ZqQSource snippet
How Recommendation Algorithms Actually Work...
-
Source: youtube.com
Title: How Recommendation Algorithms Actually Work
Link: https://www.youtube.com/watch?v=iWwQGYkZTAwSource snippet
Collaborative Filtering Explained | Recommender Systems Tutorial for Beginners...
-
Source: youtube.com
Title: Collaborative Filtering Explained | Recommender Systems Tutorial for Beginners
Link: https://www.youtube.com/watch?v=rizwooaTNEASource snippet
Recommendation Systems - A Deep Dive into Collaborative Filtering...
-
Source: youtube.com
Title: Recommendation Systems
Link: https://www.youtube.com/watch?v=EdDj06vxjnw -
Source: washingtonpost.com
Title: The Washington Post Are you in Tik Tok’s cat niche?
Link: https://www.washingtonpost.com/technology/interactive/2025/tiktok-algorithm-video-map-interests/Source snippet
What 121,000 videos reveal.December 10, 2025 — The Washington Post's investigation into TikTok’s recommendation algorithm provides an in...
Published: December 10, 2025
-
Source: facebook.com
Link: https://www.facebook.com/Filter/Source snippet
"238776 likes · 1184 talking about this. New Album THE ALGORITHM: Ultra Edition is out now! [https://linktr.ee/officialfilter..."](https://linktr.ee/officialfilter...")...
Additional References
-
Source: instagram.com
Link: https://www.instagram.com/officialfilter/?hl=enSource snippet
FILTER (@officialfilter) • Instagram photos and videosThe Official Instagram of FILTER The Algorithm: Ultra Edition AND Short Bus 30th An...
-
Source: merriam-webster.com
Link: https://www.merriam-webster.com/dictionary/filterSource snippet
FILTER Definition & Meaning4 days ago — The meaning of FILTER is a porous article or mass (as of paper or sand) through which a gas or li...
-
Source: medium.com
Link: https://medium.com/%40zentrinozen/how-youtubes-recommendation-algorithm-actually-works-the-complete-breakdown-for-creators-6b395be74db2Source snippet
How YouTube's Recommendation Algorithm Actually WorksYouTube's AI analyzes your watch history, what you've liked, what similar viewers wa...
-
Source: officialfilter.com
Link: https://officialfilter.com/Source snippet
Official FilterThe project fuses industrial trap metal with raw hostility, delivering an unapologetically aggressive modern sound that re...
-
Source: reddit.com
Link: https://www.reddit.com/r/NewTubers/comments/ghwioq/does_my_personal_viewing_history_affect_how_my/Source snippet
Should I keep a separate account for my watching? I mostly watch league of legends vods on...
-
Source: support.google.com
Link: https://support.google.com/youtube/answer/16559651?hl=enSource snippet
Google HelpGood to know about recommendations for YouTube's...What matters is how viewers respond to each video when it's recommended to...
-
Source: filterking.com
Link: https://filterking.com/?srsltid=AfmBOoqIki0RHeTjVxon_os7klGY9Pcfpc955PsJLFITHOK8eFvDajPVSource snippet
ee shipping, bulk orders, subscribe & save...
-
Source: filtersfast.com
Title: Find the right filter for your home today!
Link: https://www.filtersfast.com/?srsltid=AfmBOopXUauHS988V173DqZRhXa5TaoofY2AY7wIyMrXua8Imk_Qxwk7Source snippet
Water & Air Filters, Home Filtration & Replacement Parts...Shop Filters Fast for a wide selection of water & air filters, home filtratio...
-
Source: news.ycombinator.com
Link: https://news.ycombinator.com/item?id=35445992Source snippet
with direct knowledge of YouTube Algorithm-why...4 Apr 2023 — Whenever a 3blue1brown video pops up in my feed again, it's like seeing an...
-
Source: Wikipedia
Title: Filter (band)
Link: https://en.wikipedia.org/wiki/Filter_%28band%29Source snippet
Filter (band)Filter is an American rock band formed in 1993 in Cleveland, Ohio, by singer Richard Patrick, along with guitarist and pr...
Topic Tree