How to Get More Followers on X (Twitter) in 2026: What Actually Works

I’m going to say something that most growth guides won’t.

The tactics in 90% of “how to grow on Twitter” articles were written when the algorithm was a completely different system. Manually tuned heuristics. Hashtag stacking. Follower-for-follower. Engagement pods. Posting 10 times a day.

That system is gone.

In January 2026, X’s parent company xAI published the complete recommendation algorithm on GitHub (xai-org/x-algorithm). For the first time, we can actually read how the platform decides what to show to whom. Not from reverse engineering. Not from guesswork. From the code itself.

In this post, I’m going to show how you can grow you Twitter account in 2026 without getting banned.

First…

What the Algorithm Actually Does

When you open X, a system called Home Mixer assembles your feed in real time. It runs the whole process in under 1.5 seconds and processes billions of ranking decisions every day.

The engine doing the actual scoring is Phoenix, a Grok-based transformer model. This is a major change from the previous system. The 2023 open-source release was described by Wired as “spaghetti code” full of manual rules and legacy heuristics. The 2026 release confirms all of that has been replaced by a single unified AI system.

As X’s engineering team wrote in the GitHub readme:

“We have eliminated every single hand-engineered feature and most heuristics from the system. The Grok-based transformer does all the heavy lifting.”

Phoenix doesn’t just count likes. It reads the text of your post, tracks how long someone pauses on it, and uses your entire engagement history as context. The ByteByteGo breakdown of the code is the most thorough technical read if you want to go deeper.

The feed itself pulls from two sources:

Source	What it means for you
In-network content	Posts from accounts you follow, ranked by the Real Graph model
Out-of-network content	Posts from strangers, surfaced via the Candidate Pipeline

That second row is the growth opportunity. If your post performs well with your existing followers in the first 15 minutes, Phoenix considers it for the out-of-network pool and starts pushing it to users whose engagement history matches your topic.

That’s how accounts with small followings break through.

A Note on the Engagement Weights

The 2023 Twitter algorithm release contained specific engagement weights that became widely quoted:

Engagement	2023 Weight
Retweet	20x
Reply	13.5x
Profile click	12x
Link click	11x
Bookmark	10x
Like	1x

These numbers spread everywhere and most guides still repeat them as current fact. Here’s what they don’t tell you.

The January 2026 release redacted the specific weighting constants. As VentureBeat noted in their analysis, developers who reviewed the new code confirmed the actual multipliers are hidden. The 2023 weights may still be directionally correct, but nobody outside xAI knows the exact numbers now.

What we do know from the code is the types of signals Phoenix predicts:

• P(favorite), P(reply), P(repost), P(quote), P(click), P(profile_click), P(video_view), P(dwell), P(follow_author)
• And negative signals: P(not_interested), P(block_author), P(mute_author), P(report)

The direction is clear even without exact weights. Conversations outperform passive approvals. Negative signals actively suppress your reach.

Fix Your Profile Before You Post Anything Else

Your profile is the conversion page between someone seeing your tweet and deciding to follow you.

Good content does its job and sends people to your profile. A weak profile is where you lose them.

What actually matters:

Photo. A real face outperforms a logo for personal brands. Humans follow humans at a higher rate. No blurry photos, no defaults.

Header image. Most accounts leave the default blue. That’s a wasted billboard. One message: who you help or what you do.
Bio. Write about the reader, not yourself.

“Helping [specific person] do [specific thing]” outperforms “Marketer | Coffee lover | Dog dad” consistently. You have 160 characters. Spend them on value, not identity.

Pinned post. The first thing visitors read. Put your best work there. Update it when something better exists.

On X Premium:

The algorithm gives Premium subscribers a documented reach boost and priority reply placement. X’s head of product confirmed it publicly. At $8/month it’s the lowest-cost lever for improving distribution. It doesn’t rescue bad content. It amplifies what already works.

TweepCred: The Hidden Score That Controls Your Reach

Every X account carries a hidden reputation score from 0 to 100 called TweepCred. It uses a weighted PageRank approach, so the quality of accounts that interact with you matters as much as the quantity.

The critical number is 65.

Fall below 65 and only 3 of your posts compete for algorithmic distribution at any time. The content quality is irrelevant below that threshold.

⚠️ New accounts face something worth knowing about: engagement debt. If your first posts land quietly or trigger negative signals (mutes, “not interested” taps, blocks), the algorithm suppresses your reach for months. Post your best content first. Don’t treat early days as a test period.

What moves TweepCred up or down:

Raises it	Lowers it
High followers-to-following ratio	Bot and inactive followers
Replies from high-scoring accounts	Blocks, mutes, reports from users
Consistent posting history	Aggressive follow-unfollow patterns
X Premium subscription	Consistently hostile post sentiment
Older account age	Coordinated engagement behavior

That last item in the “lowers” column is worth understanding. Phoenix includes sentiment scoring. Grok reads the tone of your posts. Accounts with consistently hostile or inflammatory patterns get a lower trust classification, which affects distribution separately from TweepCred.

The Link Penalty Nobody Warned You About

Since early 2025, posts with external links from free accounts have near-zero median organic reach.

Sprout Social’s 2026 data confirms link posts had an average 0.13% engagement rate for influencers in 2025, vs 0.48% for text posts. That gap isn’t random. Phoenix applies a reach reduction to posts that send users off-platform.

The fix is simple and it works.

The wrong way: Post your content with the link in the main tweet.
The right way:

• Post your content with no external link
• Reply to your own post immediately with the URL

The main tweet gets full algorithmic distribution. Interested readers find the link one tap away.

⚠️ If your reach dropped in early-to-mid 2025 without a clear content reason, go back and check your post history. If you were regularly putting links in main tweets, that penalty is likely a significant part of the answer.

The First 15 Minutes Decide Everything

This is the part most guides get wrong. They say “the first hour” or “the first 60 minutes.”

Based on the VentureBeat analysis of the 2026 code, the actual velocity window is tighter. If engagement signals fail to cross a dynamic threshold in the first 15 minutes, a post is unlikely to enter the general For You pool at all.

How reach decays after posting:

Time elapsed	Approximate reach remaining
0-15 minutes	Peak window, algorithm decides distribution
30 minutes	Strong but falling
3 hours	Roughly 60% of peak
6 hours	Around 50% of peak
24 hours	Near-zero for most accounts

This is why posting when your specific audience is online matters more than any generic “best time to post” advice. Your 15-minute window needs to overlap with when your followers are actually scrolling.

I use Circleboom for this specifically. The Best Time to Tweet feature analyses when your actual followers are active, not industry averages. It’s a small thing that makes a measurable difference when the window is only 15 minutes.

💡 Practical detail: Reply to every response you get inside that first 15 minutes. The algorithm tracks whether conversations continue after initial engagement. A thread that keeps going is a signal that the content is worth distributing further.

Dwell Time: The Signal Most Guides Don’t Mention

Phoenix tracks how long users look at each post. Under 3 seconds is a mild negative signal. A genuine pause before engaging is a positive signal, logged before any explicit click happens.

This changes how you should approach the first line of any post.

That first line has one job: stop the scroll. Not with bait. With something specific, counterintuitive, or immediately interesting enough that a thumb stops moving.

Formats with the best dwell time on X right now:

• Threads where each step reveals something useful
• Posts with visual structure (numbered points, deliberate line breaks, something below the fold)
• Posts that open a question in the first line and answer it below

Formats with the worst dwell time:

• Plain text announcements with all information in the first sentence
• Posts formatted identically to everything else in the feed
• Vague teaser posts that don’t deliver in the first few lines

What to Post (and What to Stop)

X is still a text-first platform. Sprout Social’s 2025 data shows text posts from influencers averaged a 0.48% engagement rate vs 0.41% for photo posts and 0.13% for link posts.

That said, format variety matters. Phoenix evaluates content originality. Posting the same format repeatedly gets scored lower over time.

What’s working in 2026:

Threads on specific niche topics. Multiple engagement points, keeps people in your profile longer, gives Phoenix more topic cluster signal. A strong niche thread is still the highest-reach format available.

Single punchy takes with a clear position. These get quoted and retweeted, which carry the highest weight in the scoring hierarchy.

Directed questions to your audience. Replies compound. A post that becomes a genuine conversation will outreach one that got a lot of silent likes.

What’s actively working against you:

Multiple hashtags. Phoenix reads your text and understands the topic without them. RivalIQ data from 2025 shows most industries getting under 0.04% engagement. Adding more hashtags does not help. 1-2 niche-specific tags maximum.

High volume with low per-post engagement. The 2026 code includes a scorer that penalizes multiple posts from the same account in a short window, as VentureBeat’s analysis notes. 3-5 quality posts per day beats 15 mediocre ones.

Engagement Is the Strategy, Not a Side Effect

The biggest gap between accounts that grow and accounts that plateau is simple.

Growing accounts spend as much time in other people’s conversations as they do posting their own content.

This isn’t reciprocity farming. It’s how Phoenix builds your topic cluster association. When you consistently reply to accounts in your niche, the algorithm maps you to that content community. Your posts start appearing in the For You feeds of people who follow those accounts.

The daily framework:

Reply to 10-15 posts in your niche before you post anything yourself. Not “great point” replies. Your actual perspective. A counterpoint. A follow-up question. Something that adds to the thread.

Then respond to every reply you get within the first 15 minutes.

The goal is to be the most valuable voice in the replies of accounts your target audience already follows. That’s organic discovery in 2026.

Clean Your Follower List (This Is a Real Growth Strategy)

Here is the counter-intuitive part.

Your follower count growing is not always good for your account performance.

When bots, fake accounts, and dormant users follow you, your engagement rate drops because they never interact with content. Phoenix sees a large following with weak engagement and interprets it as low-quality content. Distribution falls. TweepCred drops.

The follower number goes up. The reach goes down.

This is why buying followers destroys organic reach. And it’s why periodic follower audits are a legitimate growth strategy, not just housekeeping.

I use Circleboom for this. The follower audit identifies fake accounts, bots, inactive users, spam accounts, and egghead accounts (no profile photo) in bulk. You can filter by category and remove them without scrolling through thousands of profiles manually.

The Plus plan at $23.99/month covers accounts up to 100,000 followers and includes follower growth tracking, demographic breakdowns, and the Best Time to Tweet feature.

If your engagement rate has been declining with no obvious content reason, run a follower audit before changing anything else.

✅ One practical note: Don’t mass-unfollow at speed. X flags aggressive unfollow patterns as bot-like behaviour, which is a direct TweepCred negative signal. Work gradually, or use a tool that respects X’s rate limits. Circleboom handles this automatically.

Two Underused Channels Worth Your Time

X Spaces

Almost no growth guide mentions Spaces. Which is exactly why it matters.

Active Spaces surface near the top of For You feeds while they’re live. A Space with even modest activity shows up prominently for followers of all participants and for users whose interest profile matches the topic.

For accounts under 10,000 followers, hosting a regular weekly Space in your niche gives you genuine top-of-feed visibility without paying for ads.

A listener who follows you after spending 30 minutes in your Space is a higher-quality follower than someone who tapped follow after a viral post. They already know what you’re about before they hit the button.

X Communities

X Communities went fully public in February 2026. Before that, posting to a Community only reached its members. Now, strong Community posts surface in For You feeds for the general audience.

For early-stage accounts, posting into active niche Communities is one of the fastest routes to genuine early growth. You start with a warm, topically-aligned audience rather than posting into the general feed with zero existing signal.

Contribute consistently. Don’t just drop links. The algorithm distinguishes between the two.

What Doesn’t Work Anymore

Tactic	Why it fails now
Follow-unfollow	X detects the pattern. TweepCred penalty. Any followers gained have no genuine interest and hurt your engagement rate.
Buying followers	Engagement rate collapses, TweepCred drops, distribution falls. The math never works.
Links in main tweets	Confirmed reach penalty. Put links in first reply.
Engagement pods	Phoenix detects coordinated behavior from accounts that don’t naturally interact. Short-term metric bump, long-term signal damage.
Multi-hashtag posting	Engagement penalty confirmed by RivalIQ’s 2025 data. Phoenix reads your text. Hashtags are for niche discovery only.
Posting only to promote	Audiences tune out. Organic content tanks alongside the promotional posts. Presence first, promotion second.

Track What’s Actually Moving

X’s native analytics at analytics.twitter.com shows impressions, engagement rate, profile clicks, and link clicks per post. Check it every week.

Look for patterns:

• Which formats drive the most replies?
• Which topics get bookmarked?
• Which posts generate profile visits?

Profile visits are the cleanest signal of all. Someone saw your content and immediately wanted to know more about the person behind it. That’s the response every post should aim for.

Circleboom adds demographic insight on top of native analytics: follower growth over time, audience language and interest distribution, and tweet performance trends. If you’re tracking growth seriously across months, having both together gives you a clearer picture than either alone.

The Honest Timeline

Most accounts reaching 10,000 followers organically take 6-12 months of consistent effort from scratch.

3-5 posts daily. Engaging in your niche every day. Periodic follower audits. Adjusting based on what your analytics actually show.

X’s average brand engagement rate dropped from 0.029% in 2024 to 0.015% in 2025, a 48% decline according to Metricool’s annual report. The accounts that are growing are the ones beating that number, not chasing it.

The ones that grow consistently and then plateau are almost always making a technical mistake rather than a content mistake. Links in main tweets. Fake followers dragging down engagement rate. Missing the 15-minute velocity window. Fix the technical layer first. Then the content compounds on its own.

Frequently Asked Questions

Bottom Line

X in 2026 is not a platform you broadcast to.

It reads every post. It scores your account reputation. It measures how long people look at your content. It distributes based on a weighted hierarchy that most people still don’t know exists.

The accounts growing consistently treat it like a community. They reply more than they post. They protect their engagement rate. They show up in the 15-minute window that actually matters.

None of this is complicated. The execution is the hard part.

Fix the technical layer first: link penalties, follower quality, TweepCred health, posting timing. Then build the content habits on top of a clean foundation.

Do both for six months. The growth stops being mysterious.

Also on CuriousBlogger:

• Circleboom Review 2026: Is It Worth Your Time?
• How to Delete Old Tweets by Date on X/Twitter
• Twitter (X) Username Ideas That Don’t Suck