Should I number the tweets in a thread?

Generally no, in 2026. Modern X threads chain naturally because the UI handles the flow. Numbering ('1/', '2/') was useful when threads were visually fragile but now signals 'I'm trying to be a thread account'. Skip it unless your audience expects it from your prior style.

How long should a thread be?

3–8 tweets. Anything longer reads as 'I should have written a blog post.' If your newsletter has 12 takeaways, that's two threads on different days, not one mega-thread.

Should the first tweet have a hook or a hashtag?

Hook, never a hashtag. Hashtags in tweet 1 dilute attention. If you must use one, put it at the end of the last tweet.

Can I link to my Substack from inside the thread?

Yes, in the LAST tweet only. Putting a Substack link in tweet 1 actively kills reach — X's algorithm down-ranks tweets with external links. Hold the link until you've earned the reader's attention.

What's a 'standalone hook'?

A first tweet that delivers value or intrigue even if no one reads tweet 2. If a stranger could screenshot tweet 1 and share it as its own thing, you've nailed it. If tweet 1 only makes sense as setup for tweet 2, you've buried the lede.

Turning a Substack newsletter into an X (Twitter) thread that doesn't suck

If you write on Substack and you've tried to turn an issue into an X thread, you already know the problem. The newsletter version flows beautifully — paragraphs, callbacks, a slow build to a payoff. The thread version reads like the newsletter chopped into 280-character chunks: each tweet a fragment that doesn't quite stand on its own.

This guide explains the structural differences and shows what a Substack-to-X conversion looks like when it's done well.

What's different between Substack and X threads

Substack rewards depth. Subscribers opted in. They'll read 2000 words. They want context, narrative, and craft. The opening can be a slow-burn anecdote that sets up a bigger argument.

X rewards punch. Most readers see your thread because tweet 1 hit their feed via a follower's repost or the algorithm. They have no commitment to keep reading. Tweet 1 has to deliver value or intrigue before tweet 2 appears.

This means the conversion isn't a translation — it's a restructure. The thread isn't your newsletter compressed; it's a different argument, drawn from the same underlying material, optimized for a different attention pattern.

The 3 thread shapes that work

Shape 1: The number-led list

Claude Code went from zero to the #1 AI coding tool in 8 months. We surveyed 900+ engineers to find out what's actually happening with AI tooling in 2026.

95% of respondents use AI tools at least weekly. 75% use AI for half or more of their engineering work. The question is no longer whether to use AI — it's which tools.

Claude Code leads tool usage, followed by chatbots, GitHub Copilot, and Cursor. Most engineers juggle 2–4 tools at once.

Each tweet has a self-contained insight. The thread can be skimmed; reader can drop off at any tweet without missing the thesis.

Shape 2: The contrarian arc

Last quarter we deleted our pricing page. Conversions doubled.

The thinking: pricing pages are objection generators. Prospects see the number before they see the value, and they anchor on cost.

What replaced the pricing page: a 15-minute discovery call. Sales cycle stayed the same length. Average deal size went up 30%.

Counter-intuitive lesson: removing friction at the top of the funnel often means adding it lower down. The right question isn't "fewer steps." It's "fewer wrong steps for unqualified leads."

Open with the punchy claim. Each subsequent tweet earns the reader's trust by adding a piece of evidence. End with the takeaway.

Shape 3: The personal-stake unfolding

I almost shut down our content team last quarter. Here's what changed instead.

We were spending $40K/quarter on freelance writers and seeing flat traffic. The metric I trusted (organic sessions) had been gaming me — bots and brand searches inflated it.

...

Personal stakes are rare in B2B X. They cut through because most threads are pattern-matched as "thought leadership". A real story isn't.

Tweet-level rules

Each tweet ≤ 280 characters strictly. URLs count as 23 characters even if they look longer. Quote tweets count differently. Test before posting.
Vary length intentionally. A 100-character tweet followed by a 250-character tweet creates rhythm. If every tweet is 280 characters, the thread reads exhausting.
No "1/", "2/", "3/" labels unless your audience already expects them from your style. They're vestigial.
No hashtags inside the thread. One at the end of the last tweet, max — and only if it's a real topic tag.
The last tweet should be either a punchy summary or a soft CTA. "Read the full piece on Substack: [link]" works fine if it's the last tweet, not the first.

Hook patterns that perform on X

Empirically, threads led by these openings outperform:

Specific numbers: "Claude Code went from zero to #1 in 8 months. Here's the data."
Counter-intuitive claims: "We deleted our pricing page. Conversions doubled."
Industry observations: "Every B2B SaaS founder under $1M ARR I've talked to in the last 12 months makes the same content mistake."
Personal confession: "I almost killed our content team last quarter. Here's what changed."

Hook patterns that flop:

"Today I want to talk about..."
"I've been thinking a lot about..."
"Here's a thread on..."
Anything that sounds like a podcast intro.

What to automate

The mechanical work is heavy on a thread:

Splitting prose into 280-char chunks at sentence boundaries (not mid-word).
Rewriting transitional language so each tweet stands alone.
Pacing length variation across tweets.
Removing exclamation marks and AI-tells (em-dashes used as ellipses, "delve into", "unpack", etc.).

A tool like Letterfork extracts your voice from past Substack issues and applies it to every thread, so the output sounds like your real writing rather than generic AI rewrites. The key is the voice extraction — without it, the threads read polished but anonymous.

What to keep manual

The thesis. Tools can structure a thread but they can't decide which argument from your newsletter is worth threading this week.
Real-time tweaks. If a competitor tweeted something related two hours ago, your thread should subtly reference it. That's judgment, not template.
The decision to post a thread vs. a single tweet. Some takeaways thrive as one-liners. Don't pad them into threads just because you have a tool.

A worked example

Substack excerpt (from a B2B sales newsletter):

Last quarter we ran an unusual experiment. We removed the pricing page from our trial-to-paid funnel. Conversions doubled, from 11% to 23%. The hypothesis going in was that pricing pages are objection generators — prospects see a number before they see value, and they anchor on cost rather than outcome. What we replaced the pricing page with was a 15-minute discovery call. The friction is now lower at the top of the funnel and intentionally higher in the middle, where unqualified leads filter themselves out. Sales cycles got shorter, not longer.

X thread (converted):

Last quarter we deleted our pricing page. Conversions doubled.

11% → 23% trial-to-paid.

The thinking: pricing pages are objection generators. Prospects see the number before the value, and they anchor on cost.

What replaced it: a 15-minute discovery call. We added friction in the middle of the funnel, not the top.

Sales cycles got shorter, not longer. Unqualified leads filter themselves out at the call stage.

Counter-intuitive: the right question isn't "fewer steps". It's "fewer wrong steps for unqualified leads."

Same insight, different shape. The newsletter version is dense and analytical; the thread is sharper, with each tweet a complete idea. A tool that handles the structural work lets you spend your time choosing which insights become threads, not chunking sentences.