Twitter Threads: How to Write One That Gets Read in 2026

The short version: most Twitter threads die after tweet two because writers optimise the hook and forget the reader has to keep tapping "show more" on purpose. Write for the scroll, not the essay, keep every tweet a complete thought that earns the next click, and stop pretending length equals value because the data says otherwise.

The uncomfortable bit nobody tells you about threads in 2026

I have written maybe 400 threads on X since 2016. I still write them for clients. And here is the thing almost every "how to write a viral thread" guide skips: the algorithm does not particularly want you to write a thread anymore.

X's own ranking signals reward dwell time and reply engagement, not "number of tweets in a chain." A single sharp tweet that gets 40 replies will out-perform a 12-tweet thread that gets 3 replies and a handful of silent scrolls, every single time. I have tested this on my own account back to back in the same week, same topic, one as a single tweet and one as a thread. The single tweet did better on impressions per hour. The thread did better on saves and profile clicks. Different jobs, not a hierarchy.

So the first honest thing to say is this: a thread is not a "better tweet." It is a different tool, for a different goal, and if you use it because you think X favours threads by default, you are working from a rumour that stopped being true around 2022.

What "gets read" means on X

Twitter's own analytics used to show completion data more openly. What I have seen consistently across accounts I manage: roughly 60 to 70 percent of people who open tweet one will read tweet two. By tweet five, that number is closer to 20 to 30 percent. By tweet ten, you are often down to single digits unless the thread is gripping or the topic is niche enough that the remaining readers are self-selected and invested.

That drop-off is not a failure of your writing skill. It is how scrolling works. People tap "show this thread" out of curiosity, not commitment. Your job in every single tweet is to re-win that commitment, not assume it.

A thread that should have worked and didn't

Here is the honest story. Two years ago I wrote a nine-tweet thread on AI tools for small business owners. Strong hook, tested three variations before posting. Numbered list format, which usually performs well for me. Screenshot in tweet four. A "save this for later" line in tweet seven. Everything the playbook says to do.

It got 340 impressions in the first hour and basically flatlined. Meanwhile a throwaway single tweet I posted the same afternoon, just a blunt opinion about ChatGPT replacing junior copywriters, pulled 22,000 impressions and 180 replies.

What went wrong with the thread wasn't the hook. It was tweet three. I'd written "Let's break this down" as a transition line with zero new information in it, just a placeholder while I set up tweet four. That single dead tweet was where the drop-off spiked hardest when I checked the numbers later. Readers do not forgive a wasted tap. They will punish one weak link by leaving the whole chain.

The fix I now use on every thread: read every single tweet in isolation and ask "if someone only saw this one tweet, would it be worth their time?" If the answer is no, that tweet gets cut or rewritten, no exceptions, even if it breaks my numbering.

The actual step-by-step structure

This is the process I use now, for myself and for clients, after years of testing:

• Step 1: Write the last tweet first. Know your ending before you write your hook. Threads without a destination wander, and wandering is where readers drop off.
• Step 2: Write five hook variations for tweet one. Not one. Five. Pick the one that makes a specific promise ("I lost £4,000 on a Facebook ad in one weekend, here's exactly what I got wrong") over a vague one ("Some thoughts on ad spend").
• Step 3: Keep it under 7 tweets unless the topic demands more. My own data shows completion rate drops sharply after tweet seven. If you need more than that, consider whether it should be a blog post with a thread that links to it instead.
• Step 4: Front-load the specific detail. Numbers, names, dates, screenshots, by tweet three at the latest. Generic advice in tweets two and three is the single most common thing that kills a thread before it gets going.
• Step 5: End with an action, not a summary. "So that's my thread" is a wasted final tweet. Ask a question, invite a reply, or point to one clear next step.
• Step 6: Post it, then reply to your own first comment within 5 minutes with a related point. This is an old trick and it still works because it gives the thread its first bit of reply activity before real replies arrive, which nudges the algorithm.

Why the numbered-list format still works (and where it breaks)

Numbered threads ("7 mistakes I made building a six-figure consultancy") remain reliable because they set a contract with the reader: you know exactly how much is left. That predictability reduces the drop-off I mentioned earlier. But they break the moment one entry in the list is weaker than the others, because readers notice inconsistency in a numbered format faster than in a narrative one. If you cannot make all seven points equally strong, cut to five. A tight five beats a padded nine, every time I have tested it.

The profile matters more than people admit

A thread's real job, most of the time, is not the impressions on the thread itself. It is what happens after someone finishes reading and clicks through to your profile. I have seen threads with modest reach produce disproportionate follower growth simply because the bio and pinned tweet did the closing work the thread couldn't. If your Twitter bio isn't written to convert that specific click, you are leaving the actual value of the thread on the table. People read the thread, then decide whether you're worth following in about four seconds on your profile page, so that page has to carry weight too.

Distribution is not optional

Writing the thread well gets you maybe 30 percent of the way to it being "read" by more than your existing followers. The rest is distribution mechanics: posting time, whether you engage in the replies within the first 20 minutes, whether anyone with reach quotes it. If you want more consistent reach, it is worth understanding what drives retweets on X in 2026 rather than assuming quality alone will carry a thread. Quality gets you noticed by the algorithm briefly. Retweets and quote tweets are what carry it past that first window.

Should you even bother chasing threads for money

People sometimes ask me if writing threads is worth it financially, usually because they've seen headlines about creator payouts. Two things worth knowing. First, X's payout structure has always been thin for most accounts and heavily weighted toward huge follower counts and premium subscribers engaging, so before you build a content strategy around it, look honestly at how much X pays for views and whether that math works for your account size. Second, people often confuse X with Meta's separate Threads app when talking about creator pay, and the models are different animals, worth reading the real numbers on what Threads creators get paid before assuming either platform will fund your content habit. For most small business owners and consultants, threads are worth writing for authority and lead generation, not direct platform revenue.

When not to write a thread at all

Some ideas are single tweets pretending they need ten. If your point can be made in 280 characters with punch, splitting it into a thread dilutes it and gives readers eight extra chances to leave before they get your point. I make this mistake myself when I'm attached to a story I want to tell in full. Test: if you removed tweets three through six, would the thread still make its point? If yes, you didn't need a thread. You needed an edit.

Other ideas need the room. Case studies, step-by-step breakdowns, anything with sequential logic where step four only makes sense after step three, those earn the thread format because a single tweet would strip out the reasoning that makes the advice trustworthy.

The line I keep coming back to

After nearly a decade of writing these things, my rule is blunt: every tweet in a thread has to survive on its own, because most readers will only ever see one of them out of context, quoted, screenshotted, or shared without the rest attached. Write for that stranger seeing tweet four with no idea what tweet one said. If it still makes sense and still says something worth their four seconds, you've written a thread that gets read, not just posted.

This guide is part of my Twitter (X) Help: 21 Guides to Features, Verification, Video Fixes and Tools.

Frequently asked questions

How long should a Twitter thread be in 2026?

Five to seven tweets is the sweet spot for most topics because completion rates drop off sharply after tweet seven; only go longer if each additional tweet adds new information rather than padding out the same point.

Does X's algorithm still favour threads over single tweets?

No, not by default; X ranks on dwell time and reply engagement, so a strong single tweet with high reply activity can outperform a long thread with weak engagement on the same day, same topic.

What is the biggest reason threads stop getting read halfway through?

One weak transition tweet with no new information is usually the exact spot readers drop off, because a single wasted tap teaches them the rest probably isn't worth their time either.

Is it worth writing threads for money on X instead of Meta's Threads app?

Only if your follower count and premium subscriber engagement already support meaningful payouts; for most accounts, threads on either platform are better used for authority building and lead generation than direct ad revenue.