The X algorithm, formerly known as Twitter’s algorithm, does not rank posts with a simple like counter. In the For You feed, recommendations aim to predict how interesting a piece of content will be to a particular person. Your job is therefore to be relevant to an identifiable audience, not reproduce a supposedly secret formula.
Guide updated on November 20, 2026. This version replaces advice and commercial screenshots from 2024. It distinguishes mechanisms documented by X, their limitations and editorial tests you can run. No result or ranking weight is guaranteed for your posts.
For You and Following: two ways to discover posts
According to X’s official recommendations help, For You combines content from followed accounts with recommendations from outside your network. Following displays posts from followed accounts in reverse chronological order. Appearing less often in For You does not necessarily mean your followers can no longer view your work.
As a brand, ask two different questions: do followers find useful information when they visit your account? Do your posts also give new people a reason to take an interest? A local service announcement and an analysis of your industry can meet different needs.
What the 2026 documentation actually explains
The official X recommendation system repository describes selecting posts from inside and outside the network, personalized ranking with Phoenix and visibility filters. Its notes dated October 8 and 13, 2026 report improvements in predicting time spent and finding content people read in full.
This does not mean “write longer to win.” A short explanation can be useful; an endless post can lose its reader. This is our editorial recommendation, not a bonus published by X: start with the important information, then expand only where it helps someone understand it.
The repository also documents experiments and adjustable parameters; some rules are not public. It provides insight into the system without guaranteeing the same configuration for every account and every impression. Avoid turning a code snippet into a permanent recipe.
Which signals can you examine without inventing a points system?
The ranking calculation code combines several predictions, including interactions, clicks, attention and negative feedback. Clicking a link is one of these actions; its inclusion guarantees neither a link bonus nor the absence of constraints.
A crucial point in the official parameter comments: weights apply to predicted probabilities or values, not directly to raw numbers of likes and reports. A post therefore does not have a universal score you can reconstruct from its public counters.
| What you observe | A useful editorial question | A conclusion to avoid |
|---|---|---|
| Replies and quote posts | Do readers contribute relevant questions, examples or objections? | Every reply necessarily triggers a fixed bonus. |
| Clicks to a resource | Does the post’s promise match the landing page? | A post with fewer views must be a failure. |
| Views of a video or visual | Does the media make the idea easier to understand? | Every video is favored, regardless of its content. |
| Repeatedly limited distribution | Have the subject, context, restrictions or quality changed? | One poor performance proves a penalty. |
Links, Premium and hashtags: three shortcuts to leave behind
A link in a reply is not a free pass
If your goal is to get people to read a study or request a demo, the link must remain easy to find. Compare several posts on similar subjects: clicks, qualified visits and actions after the visit. You can test its placement, but do not present a reply as a guaranteed way to bypass the algorithm.
A high-reach post without relevant visits can be less useful than a targeted message that generates concrete inquiries. Define the outcome first, then decide where to put the destination. Provide enough context for people to understand what they will find when they click.
A subscription cannot replace your value proposition
X lists reply prioritization among the benefits of its plans. This does not guarantee increased impressions for every post or make every claim from an account reliable. Check the current terms instead of relying on price screenshots from 2024.
To decide, list the features you actually need and compare their cost with your usage. Do not subscribe solely because an old tutorial portrays the badge as an automatic condition for visibility.
A trend must be relevant to your message
Use terms your customers understand and, where useful, a hashtag genuinely connected to the conversation. Stacking trends unrelated to your business mainly makes your message confusing. Do not confuse search discovery, participation in a conversation and recommendation in For You.
A realistic publishing method for brands
1. Choose a subject and supporting evidence
Define who you are addressing and what they can use: an answer to a frequent question, a real-world example or a result explained with its limitations. Prepare two or three recurring themes so you do not start from a blank page every time.
For example, a maintenance company could show how to recognize a simple problem, explain when to call a professional and specify a safety condition. This helps its audience more than a string of “happy Monday” posts or a controversy unrelated to its work.
2. Refine your opening and format
State the problem or information at the beginning. Use text alone if the message stands on its own; add an annotated photo, readable chart or demonstration if it makes the idea clearer. Review it on mobile and check the link, captions and accessible descriptions where the format supports them.
If you publish across several platforms, adapt the opening and context rather than pasting the same text everywhere. Our guide to how social media algorithms work puts these differences into an overall strategy; this article remains focused on X.
3. Leave time for genuine replies
Plan a sustainable frequency without presenting a daily minimum as an algorithmic obligation. Answer questions that move the discussion forward, acknowledge corrections and avoid empty replies intended only to bump your own post.
X’s authenticity rules prohibit inauthentic practices that manipulate the platform. Buying interactions or organizing deceptive amplification is not a community strategy. That must not be confused with legitimate content scheduling. To organize your work, see our working method for community managers.
Measuring results without confusing impressions and outcomes
Keep a tracking table with subject, format, time slot, objective and observation period. Record the data available in your X account alongside visits and conversions measured on your website. Impressions do not necessarily represent unique people, and metric definitions must remain consistent across comparisons.
If you calculate an engagement rate, specify which actions you include and the denominator. Do not compare an impressions-based ratio with one based on followers. Our social media KPI guide helps you select measurements connected to your objective.
For an initial test, choose two time slots and the same content type over several weeks. Compare several posts, not just the best performer. Record exceptional news or a paid campaign that might distort the results. This protocol is a suggested working method, not an official X threshold.
What Dailyfy lets you do for X today
With Dailyfy’s X integration, you can create content, get it approved and schedule it. Post previews help you check presentation before publication, while scheduling helps organize your calendar.
X interaction management and analytics are not currently available in Dailyfy. Reply in X and consult the data your account can access. A planning tool should not be presented as a way to receive preferential treatment from the algorithm.
Your next action plan
- Choose an audience and a question you can answer usefully.
- Prepare a few posts with evidence or a concrete example.
- Check their presentation and get approval where needed.
- Publish at a frequency you can support with real conversations.
- Compare several results using the same objective and definitions.
Keep this priority in mind: a clear promise, reliable information and consistent measurement matter more than chasing coefficients taken out of context. Reassess your method when documentation or results change, without rewriting your entire editorial approach after a single post.










