Getting likes on X might feel good. According to newly highlighted details about the platform’s recommendation system, though, likes are nowhere near the strongest signal when X decides which posts deserve more visibility.
X recently shared more information about how its feed algorithm works, offering a rare look at the signals behind what people see. The findings point toward a fairly clear shift in thinking for anyone trying to grow on the platform: creating something people actually want to pass along matters much more than collecting quick reactions.
X’s Algorithm Is Looking for More Than Likes
X’s Phoenix algorithm reportedly predicts which posts a person is most likely to engage with based partly on their previous behavior. From there, different actions are assigned different levels of importance when ranking potential posts for a user’s feed.
And the gap between those actions is surprisingly large.
A like is reportedly worth only around 0.5 points in the engagement system. A repost carries roughly 1 point, while following an account after seeing a post is weighted around 4 points. Replies, quote posts and shares through direct messages are valued at roughly 5 points.
Then there’s copying a post URL.
That action is reportedly worth around 20 points, roughly 40 times the value assigned to a like. It suggests X sees someone deliberately copying and sharing a link as a particularly strong sign that the post was worth spreading.
Posts People Want to Share Could Have the Biggest Advantage
For years, social media advice has often revolved around getting more likes and reactions. The latest X algorithm insights make that approach look increasingly incomplete.
A post that gets fewer likes but travels through DMs, quote posts or copied links could potentially send stronger quality signals than something that attracts plenty of lightweight likes and then disappears.
That changes the question creators should probably be asking.
Instead of simply wondering, “Will people like this?” it may be more useful to ask whether somebody would send it to a friend, quote it with their own opinion or drop the link into another conversation.
That usually means useful information, surprising statistics, sharp observations, genuinely funny posts, timely commentary or something that gives people a reason to keep the conversation moving.
X itself appears to be placing considerable value on that kind of behavior.
Negative Feedback Can Hurt Much More Than Weak Engagement
There’s another side to the ranking system, and the numbers become much more dramatic.
According to the reported algorithm weighting, a user reporting a post carries a penalty of around -234. Muting an account is roughly -59, selecting “not interested” is around -43 and blocking carries approximately -31.
Compare those figures with a 0.5-point like and the difference is hard to ignore.
Controversial posts can still generate replies, quote posts and attention. That doesn’t automatically make outrage-based posting a smart growth strategy. If a post pushes enough people toward reporting, muting or telling X they are not interested, those negative signals can outweigh a pile of casual engagement.
Rage Bait Still Works, but the Risk Is Bigger Than It Looks
Provoking people has always been one of the easiest ways to generate activity on social platforms. X is no exception.
The algorithm details suggest there is still room for divisive content to perform because replies and quote posts carry meaningful engagement weight. The problem comes when irritation turns into rejection.
A user arguing underneath a post is still engaging with it. A user hitting “not interested” is doing something very different. They are actively telling the recommendation system they want less of that content.
That distinction matters.
The Phoenix algorithm is personalized, so these negative signals primarily shape what that particular viewer will see rather than automatically crushing an account’s entire platform-wide reach. Reports and blocks remain negative signals, however, and repeatedly creating content audiences actively reject is hardly a promising long-term strategy.
Audience Research Suddenly Looks Even More Important
There probably isn’t a magic posting format hidden inside these numbers.
The more useful takeaway is simpler: know what your audience considers worth sharing.
Brands chasing generic engagement may end up optimizing for the wrong thing. A post designed purely to collect likes might perform differently from something built around a useful insight that people naturally want to send to coworkers, friends or followers.
That puts audience research back near the center of X strategy.
What questions does your audience keep asking? What information makes them stop scrolling? What opinions get forwarded into group chats? What kind of post makes somebody think of another person immediately?
Those behaviors are harder to manufacture than likes. They may also be much more valuable.
Social Media Today similarly notes that brands should focus on understanding what their target audiences are most likely to respond to and share, both inside and outside X.
What the X Algorithm Insights Mean for Creators and Brands
The interesting part isn’t that engagement matters. Everybody already knew that.
It’s which engagement matters.
The numbers suggest X is trying to distinguish between effortless reactions and actions that require a little more intent. Clicking like takes a fraction of a second. Copying a link, sending a post to someone or adding commentary through a quote post involves considerably more effort.
That makes shareability a much stronger target than raw engagement totals.
Creators don’t necessarily need to turn every post into a carefully engineered viral attempt, either. Sometimes the posts people share most are strangely specific, unexpectedly useful or based on an observation nobody else bothered to say.
That’s probably better news than another algorithm formula.
The Bigger Lesson: Create Posts Worth Passing Along
X’s latest algorithm disclosures don’t completely rewrite the rules of social media. They do make one thing harder to ignore.
A like is cheap.
Someone deciding your post is worth carrying into another conversation is much more meaningful.
For creators, publishers and businesses working on their X strategy in 2026, chasing increasingly large like counts may therefore be the wrong obsession. Posts that teach something, start a genuine discussion or give users a reason to share the link could have a considerably better chance of traveling.
The algorithm still considers other factors, including recency, account relationships and individual user behavior. There is no single engagement score guaranteed to make a post explode.
Still, the direction is pretty obvious.
Make something people want other people to see.