Key takeaways
To write a hook, lead with your most specific detail, cut the warm-up, and make a promise the body keeps.
Nine hook types cover most good openers: observation, bold claim, number, question, story cold open, confession, contrarian line, framework promise and aphorism.
Draft five versions of every hook and keep the most specific one; the first draft is usually the warm-up.
Retire the openers that now read as AI defaults, such as the two-part "most people think" opener and the credentials opener.
Naval compresses, Paul Graham claims then sharpens, Sahil Bloom announces a framework; each works only because it fits the writer's whole voice.
Updated 9 November 2026 with a link to our guide on post structure. On 22 October 2026: this guide now opens with how to write hooks step by step, nine hook types with examples, and notes for tweets, LinkedIn posts and essays. The Naval, Paul Graham and Sahil Bloom breakdown is still here, tightened, as the second half.
To write a hook, put the most specific, surprising or useful thing you have in the first line, cut every warm-up word, and make a promise the rest of the post keeps. A good hook is one or two sentences long, names something concrete (a number you can stand behind, a person, a moment), and opens a question the reader wants answered. Write five versions, keep the most specific one, then check that the body pays it off.
That is the whole craft in three sentences. The rest of this guide is the detail: a five-step method, nine hook types with examples, the hooks that now read as AI-made, and a close read of how three writers with very different hook habits, Naval Ravikant, Paul Graham and Sahil Bloom, open their posts on Twitter. All example hooks below are written for this article to show the pattern. None are quotes.
What is a hook in writing?
A hook is the opening line or two that decides whether anyone reads the rest. On Twitter it is the start of the first tweet. On LinkedIn it is the couple of lines that show above the “see more” fold. In an essay it is the first sentence or two of the introduction, before the thesis. The format changes, the job does not: earn the second sentence.
A hook is not the topic and not the wording. It is the structural move the writer makes in that first line: a claim, a question, a scene, a number, a confession. Two posts on completely different subjects can share a hook move, which is why hooks are so easy to borrow and so easy to spot when borrowed. If you want ready-made prompts to generate options, our guide to AI hook prompts for Twitter and LinkedIn has them, along with why shared prompts go stale.
How to write a hook in 5 steps
Step 1: Find the one specific thing. Before you write a word, name the most concrete detail in your idea: the number, the scene, the mistake, the result. A hook built on “consistency matters” has nothing to grab. A hook built on “the week I posted nothing and gained more followers than the month before” does.
Step 2: Pick a hook type. Choose one of the nine types below that fits the detail. A surprising result suits a claim or a number. A turning point suits a story cold open. A disagreement suits a contrarian line in your own words.
Step 3: Draft five versions. Write the same hook five ways, ideally using two or three different types. The first version is usually the warm-up; the third or fourth is usually the one.
Step 4: Cut the warm-up. Delete openers like “In today’s world”, “Let’s talk about” and “Here’s the thing”. Start on the first word that carries meaning. Read it aloud; if you would not say it to a friend, rewrite it.
Step 5: Check the promise. Read the hook, then the body. Does the body deliver exactly what the hook set up? A strong hook on a weak body does more damage than a plain hook, because it raises an expectation the post then breaks.
9 types of hooks, with examples
These are the hook types that show up again and again in good posts, essays and threads. Each example is written for this article to show the move, not taken from a real account.
Nine hook types, what each does, and an example of each (examples written for this article).
Hook type | What it does | Example |
|---|---|---|
Specific observation | Shares something you noticed that others missed | “Every candidate who asked about our on-call rota in the first interview ended up staying more than two years.” |
Bold claim | States a position plainly so the reader wants to test it | “Most onboarding emails should be one sentence long.” |
Number or data | Leads with a figure you can stand behind | “We cut our pricing page from 1,400 words to 300. Sign-ups held.” |
Question | Asks the exact thing the reader is stuck on (never a yes or no question) | “What would you stop doing if nobody could see your follower count?” |
Story cold open | Starts in the middle of a moment | “The client hung up eleven minutes into the call. I still think about the eighth minute.” |
Confession | Admits a mistake or a changed mind | “I told our team to post daily for a year. I was wrong about why it worked.” |
Contrarian, in your words | Disagrees with a common view, with your reason attached | “I don’t batch my posts anymore. The batched ones all read like the same afternoon.” |
List or framework promise | Tells the reader what the post will deliver | “Three questions I ask before I publish anything, and the one that kills most drafts.” |
Aphorism | Compresses an idea into one line | “Taste is knowing what to delete.” |
Notice what the strong examples share: a detail that could only come from one writer. “Most people struggle with consistency” could be anyone. “The week I posted nothing” is someone.
How long should a hook be?
Short enough to read in one glance. On Twitter, that usually means one sentence, or two short ones. On LinkedIn, it means the first line has to work on its own, because the fold hides most of the post on a phone. In an essay, the hook should lead into your thesis within about three sentences. If the hook needs a setup sentence to make sense, the setup is the hook; cut whatever came before it.
How to write hooks for tweets, LinkedIn posts and essays
Tweets and threads. The first tweet does all the work. For a thread, the hook also has to promise a payoff big enough to justify the scroll. Our guide to writing a Twitter thread without the tired formulas covers thread hooks in detail, and long-form posts on Twitter covers the first line of a longer post.
LinkedIn posts. Write only the lines above “see more” first, then the rest. One idea per line, a line break after the first sentence, no throat-clearing.
Essays and college essays. A hook sentence for an essay can be a surprising fact, a short anecdote, a pointed question or a vivid description, but it has to connect to the thesis quickly. For an argumentative essay, a strong move is to open with the strongest point against you and then answer it. For a personal or college essay, open on a specific moment rather than a quote or a dictionary definition.
Hooks to avoid in 2026
Some hook shapes were fresh a few years ago and now read as AI-made or template-made, because general AI models reach for them by default. Use them sparingly, or not at all:
The symmetric two-part opener: “Most people think it’s about talent. The reality is it’s about reps.” Or its sibling, “It’s not a tool, it’s a mindset.”
“Unpopular opinion:” followed by a popular opinion.
“Nobody is talking about this” about something everybody is talking about.
The credentials opener: “After 10 years and millions of views, here’s what I learned.”
The thread emoji and counter in the first tweet.
A list promise with generic items inside (“5 lessons that changed my life”).
The full list of patterns that make text read as machine-written is in 9 signs of AI writing, and the words AI overuses covers the vocabulary side. If you want to check a draft hook quickly, the free AI detector highlights the patterns it finds and suggests a human fix for each.
How Naval, Paul Graham and Sahil Bloom open their posts
The fastest way to get better at hooks is to study writers whose opening move is stable enough to describe. These three are useful because their moves are so different. The descriptions below stay at the level of structure, observable from their public posts, not at the level of their actual sentences.
Naval Ravikant: aphoristic compression
Naval’s typical move is that the whole post is the hook. One sentence, no setup, no qualifier. When there is a second sentence, it usually restates the first more tightly rather than extending it. The verbs are usually present tense and active, and qualifications, when they appear, come as a fresh short sentence instead of a “however” clause.
Why it works: in a fast feed, a one-line aphorism breaks the rhythm before the reader has processed what it says. Why imitations fail: they add a setup paragraph to explain the aphorism, which is exactly the move the pattern refuses. If your natural register is compressed, start at the claim and stop. If it is not, this pattern will read as borrowed.
Paul Graham: claim, then qualification
Paul Graham’s posts tend to open with a blunt claim, then add a qualification, often in a parenthetical or a follow-up sentence, that narrows the claim to the version he can defend. Sentence length varies a lot inside the opening: a short claim, a longer qualifying sentence, sometimes a short sharpening line after it.
Why it works: the qualification is the credibility signal. It tells the reader the writer has already thought about the edge case. Why imitations fail: they copy the shape but write a hedge (“sometimes”, “in most cases”) that softens the claim instead of sharpening it. The test: after your qualification, can you state the claim more precisely than before? If not, cut it.
Sahil Bloom: the framework announcement
Sahil Bloom’s common opening announces a count and a structural noun (a number of lessons, principles or mental models), and the body then delivers exactly that number. The reader knows the shape of the post before reading it, which makes a long post feel safe to start.
Why it works: a clear structural promise earns the reader’s attention budget. Why it is the riskiest of the three to copy: framework-first hooks are now one of the most common AI defaults, so readers meet them with suspicion. What separates the working version is the items inside. If each item comes from your own experience (a named situation, a real number, a when and a where), the pattern holds. If the items could appear in anyone’s list, it collapses.
The three hook patterns side by side.
Writer | Structural move | Why it works | How imitations fail |
|---|---|---|---|
Naval Ravikant | Start at the claim, stop | Compression breaks the feed rhythm | Adding the setup the pattern refuses |
Paul Graham | Blunt claim, then a sharpening qualification | The qualification signals the claim was tested | A hedge that softens instead of narrows |
Sahil Bloom | Announce a count, deliver it | A clear promise earns attention | Generic items inside the framework |
Why you cannot just copy a hook pattern
A visible hook pattern sits on top of a whole voice. Naval’s compression is the same compression that runs through his replies and bio. Paul Graham’s qualifications are a habit carried over from long essays. Sahil Bloom’s frameworks are filled from a body of specific examples. Copy the surface without what sits underneath, and the hook opens well and the body collapses. The reader who clicked through notices the gap, and that costs more trust than a plain opener would have.
This is also why general AI tools struggle with hooks: they can reproduce the surface pattern but not the body that earns it. The longer version of that argument is in why all AI-written tweets sound the same.
Want more than the opening line? Here is how to learn a creator's post structure without copying them.
Audit your own hooks in three steps
Step 1: Find your default. Pull your last 30 posts and read only the first line or two of each. Label the move (observation, claim, number, question, story, confession, contrarian, framework, aphorism). The label that shows up most is your current hook signature.
Step 2: Check it against your body voice. Does the hook fit your sentence length, vocabulary and usual post length? If the hook and the body do not match, one of them has drifted, often because you imported a hook from someone you read recently.
Step 3: Write your hook taboos. List the openers you refuse because they read as template: the two-part opener, the credentials opener, the framework count with generic items, the thread emoji. The vocabulary version of this list is in the words AI overuses.
Where hooks fit in the rest of your voice
Hook patterns are one of the elements of writing style, and the most visible one, but not the one that carries a writer over hundreds of posts. Tone, vocabulary, rhythm and the things you refuse to write do that work in the body. For the cross-platform system that turns these elements into a written brief, see personal brand voice: a framework for creators. For the post types that hold up best over time, see 9 tweet types that compound, and for what makes a single tweet land beyond its first line, viral tweet anatomy.
If you want help drafting hooks in your own patterns rather than another creator’s or a template’s, audenAI, the brain inside VoiceMoat, learns from your own posts, avoids the usual AI-template openers such as the “it’s not this, it’s that” hook, and checks your drafts for Style and Stance. VoiceMoat is a Personal Brand OS for LinkedIn and Twitter; plans start at $35 a month, with a 3-day trial for $1 (prices checked 22 October 2026). audenAI suggests. You decide. For a founder-specific take on studying other writers’ patterns without copying their topics, see AI Twitter for SaaS founders.
Frequently asked questions
How do you write a good hook?
Start with the most specific detail in your idea, pick a hook type that fits it (an observation, a claim, a number, a question, a story or a confession), write five versions, cut every warm-up word, and check that the body delivers what the hook promises. A good hook is one or two sentences and names something concrete.
What are the main types of hooks?
Nine types cover most good openers: the specific observation, the bold claim, the number or data point, the question, the story cold open, the confession, the contrarian line in your own words, the list or framework promise, and the aphorism. The strongest versions of each contain a detail only one writer could have written.
How long should a hook be?
Short enough to read in one glance: usually one sentence or two short ones on Twitter, a first line that works on its own above the LinkedIn fold, and no more than about three sentences before the thesis in an essay. If the hook needs a setup sentence, the setup is the hook.
How do you write a hook for an essay?
Open with a surprising fact, a short anecdote, a pointed question or a vivid description, then connect it to your thesis within a few sentences. For an argumentative essay, opening with the strongest point against you and then answering it works well. For a college essay, start on a specific moment rather than a quote or a definition.
Which hooks should you avoid in 2026?
Avoid the openers that now read as AI or template defaults: the symmetric two-part opener ("most people think one thing, the reality is another"), "Unpopular opinion:" before a popular opinion, the credentials opener, the thread emoji and counter, and a framework count filled with generic items.
How do Naval, Paul Graham and Sahil Bloom open their posts?
Naval tends to start at the claim and stop, with no setup. Paul Graham tends to state a blunt claim and then qualify it in a way that sharpens it. Sahil Bloom often announces a count of lessons or models and then delivers exactly that number. Each pattern works because it fits the rest of the writer's voice.
Keep reading
Craft
Elements of writing style: the 10 elements that make a writer recognizable
The elements of writing style are the choices that make one writer different from another: diction, syntax, tone, voice, rhythm, punctuation, figurative language, point of view, structure and register. This guide explains each with examples, shows how to describe a style, turns the elements into the 10 signals of voice, and covers how to find your writing voice, build a personal brand voice and catch voice drift.
16 min read
·
Craft
Drafting on Twitter across devices: where voice comes through and where context bleeds in
Standard cross-device guides treat iPhone, Mac, Android, and Windows as interchangeable platforms with minor UX differences. The voice-first reading is that each device produces a different draft because the writer's context, attention, and editing rhythm change with the device. Here's how to use each one deliberately.
7 min read
·
From the Twitter algorithm series
The Twitter/X algorithm in May 2026: why voice wins under the new transformer
Read ›
