How the Twitter X algorithm works

Twitter X made an unusual decision in 2023 that no other major social platform has replicated: it published the source code for its recommendation algorithm. That transparency gave researchers, developers, and brands more direct insight into how content is ranked and distributed than any platform has offered before. The result is that the Twitter X algorithm is less of a mystery than it used to be. The key signals are documented, the weighting logic is known, and the behaviors that earn wider distribution are confirmed rather than speculated about. What that means for brands is that optimizing for the Twitter X algorithm is less about guessing and more about understanding a documented system and creating content that fits it.

This article covers how the Twitter X algorithm works in 2026, which signals it uses to rank and distribute content, and what brands can do to improve their organic reach based on what the algorithm actually rewards.

How does the Twitter X For You feed work?

Two feeds, two distribution systems

Twitter X shows users two feeds. The Following feed shows posts from accounts the user directly follows, ordered by recency. The For You feed is the default and shows algorithmically selected content from both followed and unfollowed accounts. The For You feed is where the majority of user attention goes, which means it is the primary distribution channel for brand content that wants to reach audiences beyond its existing follower base. Understanding the For You feed algorithm is what determines whether a brand's content stays within its follower bubble or earns wider distribution.

Candidate generation: how posts enter the selection pool

The algorithm begins by pulling a candidate pool of posts from two sources: posts from accounts the user follows (in-network) and posts from accounts the user does not follow (out-of-network). Out-of-network posts enter the candidate pool based on engagement signals from the user's social graph: if people the user follows have engaged with a post, that post becomes a candidate for the user's feed. This is how strong content from a small account can earn significant reach: if a few key followers engage with it, the algorithm surfaces it to their networks.

Ranking: how posts are scored and ordered

Each candidate post is assigned a relevance score based on a combination of factors: the user's past engagement history with similar content, the post's engagement velocity (how quickly it is earning replies, reposts, and likes relative to how recently it was published), the relationship between the user and the poster, and the content type. Posts that are earning strong engagement quickly after publication score higher and receive wider distribution. Posts that are old or earning weak engagement are ranked lower regardless of follower count.

Filtering: what gets removed before the feed is shown

After ranking, the algorithm applies filters that remove content the user has signaled they do not want: posts from blocked or muted accounts, posts that contain content the user has previously dismissed, and content that falls below a quality threshold based on the account's engagement history. The filtering step also applies diversity rules that prevent the feed from being dominated by a single account or topic, which means even a high-performing account will not fill a user's entire For You feed regardless of how much content it publishes.

Boosting factors that increase distribution

X's published algorithm documentation confirms several factors that increase a post's distribution score. Replies earn more weight than reposts, which earn more weight than likes. Media posts (images and video) earn higher scores than text-only posts in many contexts. X Premium subscribers receive a distribution boost over equivalent posts from non-Premium accounts. Posts from accounts with higher follower counts and stronger historical engagement receive a baseline boost. Recency is weighted heavily: a post published an hour ago scores higher than an equivalent post from yesterday.

What signals does the algorithm use to rank content?

Engagement velocity in the first hour

The speed at which a post earns engagement after publication is the strongest early signal the algorithm uses to determine distribution. A post that earns 20 replies in the first 30 minutes will receive significantly more distribution than a post that earns 20 replies over 24 hours, even if the total engagement count is the same. The first hour after publishing is the highest-leverage window for organic reach on Twitter X, and brands that treat publishing as the end of the activity rather than the start of an engagement period consistently underperform against those that actively participate in the reply thread immediately after posting.

Reply-to-repost ratio as a quality signal

The algorithm treats a high ratio of replies relative to reposts as a signal that a post is generating genuine conversation rather than passive amplification. Content that earns substantive replies from engaged users is weighted more favorably than content that earns a high volume of reposts but few replies. For brands, this means posts that ask specific questions, take positions that invite response, or participate in ongoing conversations will earn more algorithmic benefit than posts that are share-worthy but not conversation-starting.

Dwell time and content completion

The algorithm measures how long a user spends viewing a post (dwell time) and, for video and thread content, whether they watch or read to completion. A post that users pause on, expand, or read in full earns a higher quality signal than one they scroll past after a fraction of a second. For brands, this means posts that deliver their value in the first line (causing users to expand and read more) and threads that keep readers engaged to the final post earn stronger distribution than posts that front-load nothing and reward only those who already intended to engage.

Account-level credibility signals

The algorithm evaluates not just individual posts but the account that publishes them. An account with a strong history of earning engagement, a consistent posting cadence, and a low ratio of posts that earned zero engagement will distribute new posts more widely than an account starting from scratch. Account credibility builds over time: each well-performing post improves the baseline distribution the algorithm assigns to subsequent posts from the same account, which is why consistency matters as much as individual post quality.

Social graph engagement as a distribution multiplier

When users within a target user's social graph (accounts they follow, accounts they interact with) engage with a post, that post becomes a candidate for the target user's For You feed. This means the initial audience for a post is a seeding mechanism for broader distribution. A post that earns strong early engagement from accounts with large, engaged followings earns access to those accounts' audiences through the social graph signal. For brands, this explains why a single reply or repost from a well-connected account can produce a disproportionate spike in reach.

What does the algorithm mean for brand strategy?

Publish when the target audience is most active

Because engagement velocity in the first hour drives distribution, publishing when the target audience is most likely to be online and engaging directly impacts reach. Platform-wide peak times are typically weekdays between 8 a.m. and 10 a.m. and between 12 p.m. and 1 p.m. in the target audience's time zone. However, the brand's own analytics data on when its specific audience is most active will always produce more accurate timing decisions than general benchmarks. Publishing into an active audience produces faster early engagement, which triggers the distribution cascade the algorithm is designed to reward.

Engage in the reply thread immediately after publishing

Replying to the first comments on a post within minutes of publishing extends the engagement signal, notifies commenters and keeps the thread active, and generates additional reply activity that feeds back into the algorithm's quality scoring. Brands that publish and immediately log off miss the engagement window that determines whether the post earns extended distribution. Treating the 30 to 60 minutes after publishing as an active engagement period rather than idle time is one of the highest-leverage algorithm optimization habits a brand can develop on Twitter X.

Write posts that invite specific responses

The reply-to-repost weighting means posts that are designed to generate conversation will outperform algorithmically over posts that are designed to be shared. A question with a specific, answerable framing earns more replies than a question that is too broad to prompt a direct response. A position that takes a clear, debatable stance earns more replies than a neutral observation. Writing with the reply signal in mind means ending posts with something the audience can respond to specifically, not generically.

Maintain a consistent daily posting cadence

Account-level credibility builds through consistent posting, and accounts that go silent for extended periods lose the baseline distribution advantage that consistency produces. Publishing at least once per day, even during periods when there is no major news to react to or campaign to run, maintains the algorithm's model of the account as an active, credible source. Consistency does not mean posting regardless of quality; it means maintaining a regular enough cadence that the algorithm does not treat the account as dormant between bursts of activity.

Use media posts strategically, not habitually

Media posts (images and video) earn a distribution boost from the algorithm in many contexts, but this boost only applies when the media adds genuine value to the content. An image added to a post purely to trigger the media boost, without contributing to the post's message, earns the initial distribution but produces weak engagement signals that undermine the account's credibility score over time. The algorithm rewards media that keeps users engaged longer; decorative or irrelevant media produces a short dwell time that signals low content quality to the ranking system.

For the content strategy that is built around these algorithm signals, see Twitter X content strategy. For the organic growth tactics that compound the algorithm's distribution over time, see Twitter X organic growth strategy. For visual formats that earn the media distribution boost, see Twitter X visual strategy. For measuring how algorithm performance translates into reach and engagement, see Twitter X analytics and insights.

How does your website connect to the Twitter X algorithm?

A post that earns strong algorithmic distribution and sends a surge of visitors to the brand's website has done its job. What happens next depends entirely on the website. Traffic that arrives from a trending conversation, a viral reply, or a well-timed post is high-intent but impatient: visitors who came from a fast-moving feed expect the page they land on to be immediately relevant, fast-loading, and easy to navigate. A slow or generic website landing page converts almost none of that algorithmically earned traffic into commercial outcomes.

WEMASY's website builder and Analytics and Insights tools give brands the fast, conversion-ready pages and the attribution data to make Twitter X algorithm-driven traffic count. See what is included at /pricing.

Frequently asked questions