Twitter Thread

Twitter/X Thread Writing

When to Use

Use this skill when:

• The user explicitly requests a Twitter thread, X thread, or tweetstorm on any topic
• The user wants to break down a complex idea, argument, or story into sequenced tweet-sized segments that build on each other
• The user wants to establish authority or grow an audience by sharing expertise in a structured, readable format
• The user needs to document a narrative -- a case study, personal journey, business lesson, or real-time event -- in a format native to the platform
• The user wants to drive engagement (replies, retweets, follows) through educational or persuasive content structured as a thread
• The user has a long-form insight that belongs on Twitter/X but would be ignored as a single tweet
• The user wants to repurpose a blog post, newsletter section, or talk outline into a Twitter-native format

Do NOT use this skill when:

• The user wants a LinkedIn post -- use linkedin-post instead (LinkedIn rewards different structure, longer paragraphs, and a professional framing that differs fundamentally from Twitter's voice)
• The user wants a LinkedIn article -- use linkedin-article instead (3,000+ words, SEO formatting, and academic sourcing conventions are incompatible with this format)
• The user wants a blog post -- use blog-post-writing instead (threads are not blog posts compressed into tweets; they serve a different structural and reader-attention model)
• The user wants Instagram captions -- use instagram-caption instead (Instagram captions are image-dependent and use hashtag strategies that are actively counterproductive on Twitter)
• The user wants a single standalone tweet -- a single tweet requires a different optimization strategy than a thread and does not need this skill
• The user wants a Twitter bio, pinned tweet, or profile copy -- those are distinct formats with their own rules
• The topic is inherently visual and cannot be understood without images as the primary medium -- Twitter threads work when text carries the argument

Process

Step 1: Gather Thread Context Before Writing Anything

Never begin drafting without establishing these parameters. Ask the user directly if they have not provided them:

• Core topic and central argument: What is the single sentence this entire thread is proving or demonstrating? If the user cannot articulate it, help them narrow it down -- a thread with two central arguments is a thread that fails.
• Target audience: Who on Twitter/X will read this? "Founders," "developers," "parents of teenagers," and "marketing professionals" require entirely different vocabulary, cultural references, and assumed knowledge levels.
• Thread goal: Educate, persuade, entertain, inspire, or promote? Each goal changes the architecture. Educational threads use the "insight ladder" structure. Persuasive threads use "claim-evidence-implication" per tweet. Story threads use narrative arc with tension and resolution.
• Thread length preference: Default to 8-12 tweets if unspecified. Under 5 is not a thread -- it is a tweet with subtweets. Over 15 loses reader momentum.
• Voice: Does the user have existing tweets to reference? If yes, match their cadence, vocabulary, and punctuation habits. If no, ask whether they want expert authority, peer conversation, or provocateur framing.
• Existing material: Does the user have a blog post, talk notes, newsletter draft, or research to work from? Converting existing material is faster and produces more accurate threads than inventing claims.

Step 2: Diagnose the Thread Archetype

Before writing a single word, classify the thread into one of five proven archetypes. Each has a different skeleton:

Archetype 1 -- The Insight Bomb Structure: Bold counterintuitive claim → evidence chain → practical application → CTA Best for: Sharing expertise, building authority, reaching new audiences Example signals: "I've been doing X for 10 years and here's what I learned," "Everyone believes X but the data says Y"

Archetype 2 -- The Story Arc Structure: Inciting incident → conflict/struggle → turning point → resolution → lesson extracted Best for: Personal narrative, case studies, company stories, before/after transformations Example signals: "How I built," "What happened when I," "We made a mistake and here's what we learned"

Archetype 3 -- The Listicle Thread Structure: Hook promising N items → one tweet per item with a single concrete insight → closing synthesis Best for: Tactical advice, tools, frameworks, resources, rules Example signals: "X things I wish I knew about Y," "The tools I use every day," "N rules for doing X"

Archetype 4 -- The Argument Thread Structure: Thesis tweet → 2-3 tweets building the case → addressing the strongest counterargument → reaffirming thesis with nuance → CTA for debate Best for: Contrarian takes, thought leadership, generating replies and discussion Example signals: "The conventional wisdom on X is wrong," "Here's why everyone is missing the point about Y"

Archetype 5 -- The How-To Thread Structure: Problem statement → why existing solutions fail → step-by-step process → expected results → where to go deeper Best for: Tutorials, frameworks, tactical guides, process documentation Example signals: "How to do X in Y steps," "The exact process I use to," "A step-by-step guide to"

Selecting the wrong archetype is the most common thread failure. A story written in listicle format loses its emotional arc. An argument written as a how-to loses its persuasive force.

Step 3: Write the Hook Tweet -- The Most Important Sentence You Will Write

Tweet 1 is the only tweet that appears in the Twitter timeline before the reader decides to click. Every other tweet is invisible until the hook earns the read. Treat tweet 1 as a standalone advertisement for the thread.

Proven hook formulas (use one per thread):

• The specific surprising number: "I reviewed 400 pitch decks last year. 87% failed for the same reason." -- works because precision creates credibility and the specificity triggers curiosity.
• The counterintuitive claim: "Reading more books is making you worse at thinking." -- works because it challenges an assumption the reader holds.
• The transformative promise: "This one question will change how you run every meeting you hold." -- works when the promise is specific enough to be believable.
• The earned confession: "I lost $340,000 on a product launch. Here is exactly what happened." -- works because stakes are real and the humility signals honesty.
• The provocative question: "Why do the most productive people you know keep terrible to-do lists?" -- works when the question is specific enough that the reader does not already know the answer.

Hook anti-patterns to eliminate:

• "Thread:" or "A thread on X:" -- wastes characters and signals that the writer has nothing urgent to say
• "I've been thinking about X lately and wanted to share some thoughts" -- zero tension, zero reason to read
• Starting with a definition ("Cognitive load is the mental effort...") -- sounds like a textbook, not a person
• Opening with a humble disclaimer ("This might not apply to everyone...") -- destroys authority before building it
• Generic wisdom ("Hard work pays off") -- no surprise, no reason to read further

Hook construction rule: Write 5 different hook options for every thread. The first hook you write is almost never the best one. The fifth usually is.

Step 4: Plan the Full Thread Architecture on Paper Before Drafting

Write one sentence per tweet describing what that tweet will accomplish -- not what it will say. This is the thread outline. Each sentence should be a function, not content:

Example outline for a 10-tweet thread:

• Tweet 1: Hook -- create tension with a counterintuitive claim about hiring
• Tweet 2: Validate the reader's experience -- they have felt this problem
• Tweet 3: Name the root cause that most people misdiagnose
• Tweet 4: Evidence from data (specific study or personal data)
• Tweet 5: Real example illustrating the root cause in action
• Tweet 6: The turn -- what actually works instead
• Tweet 7: Step 1 of the framework
• Tweet 8: Step 2 of the framework with common mistake to avoid
• Tweet 9: Step 3 with expected outcome
• Tweet 10: Close -- callback to hook, one-line synthesis, CTA

This architecture step prevents the two most common structural failures: threads that meander without progression, and threads that run out of runway before delivering on the hook's promise.

Step 5: Write the Body Tweets -- One Idea, One Concept, One Constraint Per Tweet

Each body tweet must follow these construction rules:

• One idea only. If a tweet contains "and" connecting two distinct concepts, split it into two tweets.
• Never continue a sentence from the previous tweet. Each tweet is a complete grammatical unit. Readers share individual tweets. A tweet that begins mid-sentence cannot be shared alone.
• Use the three-line format for complex ideas: State the claim on line 1. Add the evidence or example on line 2. Deliver the implication on line 3. Three short lines outperform one long paragraph in readability on mobile.
• Anchor every claim with a specific number or date. "Most companies struggle with this" is weak. "In a 2023 survey of 1,200 SaaS companies, 68% reported this as their #1 growth blocker" is strong. Use real data the user provides or from established sources -- never fabricate numbers.
• Use white space deliberately. Line breaks within a tweet cost characters but dramatically improve readability on mobile. A 250-character tweet with 3 line breaks reads better than a 200-character tweet with none.
• Vary tweet length. After a long, dense tweet (220+ characters), follow with a short punchy one (80-100 characters). Rhythm keeps readers moving. Uniform length creates monotony.
• Build in tension and release. Not every tweet should deliver a complete resolution. Occasionally end a tweet with a question or an incomplete setup that the next tweet answers. This is the serialization hook that keeps readers clicking through.

Step 6: Write the Closing Tweet -- Convert Attention into Action

The final tweet is the most economically valuable tweet in the thread. By the time a reader reaches it, they have already invested 2-5 minutes. They are maximally primed to act. Wasting this position on a weak close is the most common thread mistake after a weak hook.

Closing tweet construction:

• Choose exactly one CTA. Multiple CTAs produce zero action. Options: follow the account, reply with an answer to a question, retweet, reply with a specific emoji to receive something, visit a link, sign up for a newsletter. Pick one.
• Make the CTA specific to the thread's content. "Follow me for more productivity content" is generic and gets ignored. "If you want to try this framework, reply with 'YES' and I'll send you the 1-page template I use" is specific, low-friction, and creates engagement.
• Consider the callback close. Referencing the hook tweet in the closing tweet creates narrative closure that readers find satisfying. "I started this thread saying most productivity advice makes you busier. Here's the one thing that makes you effective instead: subtraction." -- then the CTA.
• Optional: include the thread summary tweet. A numbered summary in the last tweet ("TL;DR: 1. Do X. 2. Stop doing Y. 3. Test Z.") makes the thread reshareable and adds bookmark value.

Step 7: Audit Every Tweet for Character Count, Voice, and Engagement

Before finalizing the output, conduct a systematic review pass:

• Hard limit check: Count every tweet. Any tweet over 280 characters must be cut -- not continued in the next tweet, but shortened by removing words.
• Trim strategy: Cut adverbs first ("very," "really," "actually," "just"). Cut transitional throat-clearing ("So, what I'm saying is..."). Cut hedging language ("might," "could potentially," "in some cases"). Replace full words with numerals (three → 3) to save characters. Replace "and" with an em dash or line break where appropriate.
• Voice consistency check: Read the thread aloud. Every tweet should sound like the same person. Inconsistency in formality (switching from "you should" to "one ought to") breaks the reader's trust in the voice.
• Engagement pattern check: Does every tweet earn a reason to continue reading? Read the thread and ask after each tweet: "If I stopped here, did I get enough value?" If yes for every tweet, the thread has self-contained value density. If no for some tweets, those are filler and should be cut or merged.
• Hashtag audit: Remove hashtags from body tweets unless the user specifically requests them or the content is genuinely tied to a live trending topic. Hashtags in the body of educational threads reduce the intellectual authority of the content. If hashtags are needed, place 1-2 maximum in the final tweet only.

Output Format

Deliver the thread in this exact structure:

**THREAD: [Topic -- one line summary]**
**Archetype:** [Name of archetype selected]
**Goal:** [Educate / Persuade / Entertain / Inspire / Promote]
**Target Audience:** [Specific audience]
**Tweet Count:** [N tweets]

---

1/ [Hook tweet.

Use line breaks for readability.
Bold counterintuitive claim, surprising statistic, or specific question.
Never mention "thread" in this tweet.]

[Character count: XXX/280]

---

2/ [Context or setup tweet.

One idea. Complete sentence. No sentence continues from tweet 1.
Validates reader experience or builds the foundation for the argument.]

[Character count: XXX/280]

---

3/ [First body tweet.

Named root cause, evidence, or first point.
Specific data or example anchors the claim.]

[Character count: XXX/280]

---

4/ [Second body tweet.

Advances the argument or story.
New information -- not a restatement of tweet 3.]

[Character count: XXX/280]

---

5/ [Third body tweet.

Evidence, counter-example, or story beat.
If applicable: the tension or complication before the resolution.]

[Character count: XXX/280]

---

6/ [Turn or pivot tweet.

The "but here's what actually works" or "here's the lesson" moment.
The highest-value insight in the thread belongs here, not in the closing tweet.]

[Character count: XXX/280]

---

7/ [Application tweet.

How the reader uses this insight immediately.
Specific, actionable, low-friction.]

[Character count: XXX/280]

---

[Continue numbering for additional body tweets as needed]

---

N/ [Closing tweet.

One CTA only.
Optional callback to tweet 1 for narrative closure.
Optional TL;DR summary if thread exceeds 10 tweets.]

[Character count: XXX/280]

---

**CHARACTER COUNT SUMMARY:**
| Tweet | Count | Status |
|-------|-------|--------|
| 1/ | XXX | ✅ Under 280 |
| 2/ | XXX | ✅ Under 280 |
| ... | ... | ... |
| N/ | XXX | ✅ Under 280 |

**ENGAGEMENT NOTES:**
- Hook formula used: [name the formula]
- Suggested best time to post: [weekday, 8-10am or 12-1pm or 6-8pm in user's timezone -- ask if unknown]
- First reply suggestion: [a short 1-2 sentence reply the user should post as a reply to their own tweet 1 immediately after publishing, to boost engagement signal]
- Reshare summary tweet: [a standalone tweet summarizing the thread in 2-3 lines, formatted for quote-tweeting tweet 1 the next day]

Rules

1. Never exceed 280 characters per tweet -- zero exceptions. The 280-character limit is a hard platform constraint. If a tweet runs 281 characters, it cannot be posted. When a tweet is over limit, cut words -- do not move the overflow to the next tweet, which would break that tweet's structural independence.

2. Never start tweet 1 with meta-announcements. "Thread:", "A thread on:", "Some thoughts on:", "THREAD 🧵:" -- all of these signal to the Twitter algorithm and to readers that the writer has nothing urgent enough to lead with. The hook is the content, not the announcement that content is coming.

3. Never continue a sentence from one tweet to the next. A reader who sees any single tweet retweeted out of context must be able to understand it. A tweet that begins mid-sentence ("...which is why most companies fail at this.") provides zero standalone value and looks broken when shared.

4. Never use more than 2 hashtags per tweet, and prefer zero in body tweets. Twitter's own internal data has shown diminishing returns on hashtag use in feed content after 2 tags. More than 2 hashtags in an educational or opinion thread makes the content look spammy and reduces click-through from readers who came via algorithmic recommendations rather than tag searches.

5. Never pad a single-tweet idea into a thread. A thread must contain enough material to justify 5+ distinct tweets. If the core insight can be expressed in one strong tweet and two supporting points, it is a single tweet with a two-tweet reply -- not a thread. Ask the user whether the topic genuinely requires thread format before starting.

6. Always number every tweet in sequence (1/, 2/, 3/, etc.) at the opening of the tweet. This serves two functions: it tells readers where they are in the thread, and it signals to new readers seeing a mid-thread tweet that more context exists.

7. Always make tweet 1 powerful enough to succeed as a standalone tweet. The majority of users who see the thread will see only tweet 1 in their timeline. The thread must earn the click. Tweet 1 should be testable as a standalone tweet -- if it would get engagement as a single post, it will earn clicks into the thread.

8. Always place links strategically. Links in tweet 1 are penalized by the Twitter algorithm in organic reach because the platform deprioritizes content that sends users away immediately. Place links in the final tweet or in a dedicated "resources" tweet near the end. If a link is essential in an early tweet, note this trade-off to the user.

9. Always include the character count for every tweet in the output. Transparency on character counts allows the user to verify compliance and edit with awareness of how much space remains. Tweets that are well under limit (under 150 characters) should be flagged -- very short tweets in the middle of a thread can feel abrupt and may benefit from expansion.

10. Always provide a reshare summary tweet and a first-reply suggestion. The thread-posting workflow for high-performing creators includes three actions: posting the thread, immediately replying to tweet 1 with a personal note or question (this boosts engagement signal in the first 30 minutes), and posting a summary quote-tweet the following day to extend the thread's reach into a new news cycle. These are not optional enhancements -- they are part of the complete deliverable.

11. Always match the thread length to the content, not to the user's preference. If a user asks for a 15-tweet thread but the topic has 8 distinct points, write 8 strong tweets rather than 15 padded ones. Explain the reasoning. Threads lose 40-60% of readers by tweet 7 on average, so every tweet must justify its existence in terms of the reader who is one tweet away from stopping.

12. Never fabricate data or statistics. If the user provides specific numbers, use them. If the user does not provide data, use frameworks and patterns ("In most cases," "Research consistently shows," "The data from X study") or ask the user to provide the supporting data. A thread citing a made-up study is worse than a thread that relies on strong reasoning alone -- it destroys the user's credibility when the fabrication is discovered.

Edge Cases

The User Wants a Thread Longer Than 15 Tweets

Threads that exceed 15 tweets suffer a documented drop-off in completion rates that makes tweets 12-15 essentially invisible to most readers. If the user's topic genuinely requires 20+ tweets, recommend splitting into two separate threads connected by a link in the final tweet of Part 1 and a reference in the hook of Part 2. Frame Part 2 as a sequel ("Yesterday I shared why X fails. Today: the exact framework to fix it."). Each part must have its own strong hook -- "Part 2" is not a hook.

If the user insists on a single long thread, deliver it but annotate each tweet above tweet 12 with a warning flag indicating the average reader drop-off risk at that point. This allows the user to make an informed choice about whether the later tweets deserve to be posts in their own right rather than buried in a long thread.

The User Wants to Include External Links

Link placement is a strategic decision, not just a formatting choice. Twitter's algorithm measures "dwell time" on the platform. A link in tweet 1 sends readers away immediately, which signals to the algorithm that the content did not hold attention. For organic reach, place all links in the final tweet or in a dedicated tweet near the end labeled clearly ("Resources and further reading:").

Exception: if the thread is explicitly a promotional thread (the user is promoting their product, newsletter, or event), the link trade-off is acceptable because the goal is conversions, not algorithmic reach. In that case, place the link in both a mid-thread tweet and the final tweet for maximum exposure, but note to the user that this choice optimizes for clicks, not for reach.

The User Provides an Existing Blog Post or Essay to Convert

Converting long-form content into a thread is not compression -- it is restructuring. The blog post likely has an academic introduction, supporting sections, and a conclusion that are built for reading duration, not scanning speed. Do not summarize the blog post into tweets. Instead:

1. Identify the single most counterintuitive or valuable insight in the piece -- this becomes the hook.
2. Find the 6-10 discrete claims or findings that can each stand alone -- these become body tweets.
3. Discard the transitional text, introductory framing, and conclusion qualifications -- they do not translate to Twitter's format.
4. Ask the user whether the thread should link back to the original post (if so, place the link in the final tweet as a "read the full version" CTA).

The Topic Is Highly Technical or Requires Domain-Specific Terminology

Technical threads for a technical audience (developers reading about system architecture, physicians reading about clinical trials) can and should use precise terminology. Do not over-simplify for a specialist audience -- condescension destroys authority faster than complexity does.

However, technical threads for a general audience require a different approach. For every technical term, allocate one full tweet to explain it with an analogy before using it in the argument. The rule: never use a term in an analytical tweet before you have grounded it in a concrete example in a prior tweet. If the topic requires more than 2 rounds of concept-grounding, the topic needs a blog post, not a thread.

The User Wants to Write the Thread in a Voice Different From Their Own

This occurs when a ghostwriter or content strategist is creating threads for a client. The skill still applies, but two additional inputs are required: 3-5 sample tweets from the target account that represent their best-performing content, and a vocabulary check (do they use em dashes or hyphens? Do they say "folks" or "people"? Do they use exclamation marks?). Match these micro-patterns exactly. The most detectable ghost-writing tells are punctuation habits, sentence-opener vocabulary, and how the person closes emotional statements. Surface these from the samples and mirror them throughout.

The Thread Contains a Controversial or Divisive Claim

Threads that challenge consensus ("The 4-day workweek is a trap," "Most therapy is ineffective for anxiety") generate high engagement through disagreement, which the Twitter algorithm amplifies because reply velocity is a reach signal. However, this is a double-edged mechanism: low-quality controversy without evidence gets ratioed (more replies than likes), which signals the algorithm that the content is divisive rather than popular.

The mitigation structure: make the controversial claim in tweet 1, dedicate tweets 2-3 to acknowledging the strongest version of the opposing view with genuine respect, then deliver the evidence for the contrarian position in tweets 4 onwards. This structure -- "steelmanning before attacking" -- produces threads that generate debate without triggering defensive pile-ons, because readers who disagree feel heard before they see the argument.

The User Wants to Promote a Product, Service, or Event

Promotional threads follow a different architecture from educational threads. The failure mode of promotional threads is leading with the product -- the reader does not care about your product until they care about the problem it solves. The structure that works: lead with the problem (tweet 1 hook about the pain, not the solution), build the case for why existing solutions fail (tweets 2-4), introduce the product as the answer to a question the reader is now asking (tweet 5 or 6), provide social proof or specific results (tweets 7-8), and close with the offer and link (final tweet). The CTA must include friction-reduction language ("free," "takes 2 minutes," "no credit card required") to lower the activation energy for the click.

Example

Input: "Write a Twitter thread about why most startup pitch decks fail, specifically because founders explain what their product does instead of why it matters. I'm a VC analyst who has reviewed 500+ decks. Target audience is early-stage founders. 9-10 tweets."

Output:

THREAD: Why Startup Pitch Decks Fail -- What vs. Why Archetype: Insight Bomb Goal: Educate + Establish Authority Target Audience: Early-stage founders seeking seed/Series A funding Tweet Count: 10 tweets

1/ I've reviewed 500+ pitch decks in the last 3 years.

The same mistake kills 80% of them -- and founders have no idea they're making it.

You're explaining what your product does.

You should be explaining why it matters.

[Character count: 196/280]

2/ Here's what the typical deck looks like:

Slide 1: "We're building an AI tool that automates invoice processing."

Slide 2: Features. Slide 3: More features. Slide 4: Pricing.

The product is everywhere. The customer's pain is nowhere.

[Character count: 229/280]

3/ The founder who built the product knows every detail of how it works.

That knowledge becomes a trap.

You pitch the solution because you've lived inside it for 18 months.

The investor has been inside this meeting for 4 minutes.

[Character count: 210/280]

4/ Investors are not buying your product.

They're buying a bet on a problem worth solving.

Before I care how your software works, I need to believe the problem is real, painful, and large enough to build a company on.

Most decks never make that case.

[Character count: 247/280]

5/ The test I use on every deck:

Can I explain the founder's business to a smart friend in 1 sentence -- using only the problem?

"Companies lose $40K a year per employee to manual invoice errors that existing tools don't catch."

That's a problem. I'm listening.

[Character count: 256/280]

6/ Compare that to the typical version:

"Our platform uses ML to streamline accounts payable workflows."

One of these sentences makes me want to ask a question.

The other makes me want to check my phone.

Investors fund curiosity, not comprehension.

[Character count: 243/280]

7/ The fix is not complicated.

Lead every slide with the customer's problem, not your solution.

Problem slide before product slide. Always.

The product is evidence that you can solve the problem. It is not the argument.

[Character count: 218/280]

8/ Rewrite your pitch by removing every mention of your product for 10 minutes.

What's left?

If the remaining slides don't tell a compelling story about a painful, widespread, unsolved problem -- your deck isn't ready.

The product fills a hole. Make sure the hole exists first.

[Character count: 271/280]

9/ The decks I've passed to partners all share one thing:

By slide 3, I'm personally feeling the problem.

I've either experienced it myself, or the data makes me feel the weight of it for someone else.

That emotional and logical alignment is what moves a deck from "interesting" to "send the term sheet."

[Character count: 278/280]

10/ Your product is not your pitch.

The problem your customer can't stop losing money, time, or sleep over -- that's your pitch.

Lead with the wound. Then show you have the cure.

If you're revising your deck this week, start with slide 2. What is the problem in one sentence?

[Character count: 261/280]

CHARACTER COUNT SUMMARY:

ENGAGEMENT NOTES:

• Hook formula used: Specific surprising number ("500+ pitch decks") + counterintuitive gap claim ("you're doing X when you should be doing Y")
• Suggested best time to post: Tuesday or Wednesday, 8:00-10:00am in your local timezone -- peak VC and founder active windows on Twitter/X align with early weekday mornings
• First reply suggestion: Post this as a reply to your own tweet 1 within the first 5 minutes of publishing: "I'm happy to look at anyone's one-liner problem statement in the replies. Drop it below." -- this creates immediate reply activity, which boosts the thread in the algorithm's early signals window
• Reshare summary tweet: The following day, quote-tweet tweet 1 with this summary: "Yesterday's thread on pitch decks blew up. The short version: investors fund problems, not products. Lead with the wound. Show you have the cure. Your product is not your pitch." -- this gives the thread a second algorithmic push and reaches people who missed the original posting