--- title: "Run a launch thread on X that converts" excerpt: "9-step playbook for writing the launch thread, scheduling the engagement window, and converting impressions into signups, not just likes." category: "Template" --- # Run a launch thread on X that converts A 9-step playbook. Open in Dock and you'll get four surfaces: - **Steps** (table) — 9 gates from drafting the hook to closing the reply window - **Drafts** (doc) — every candidate hook, body tweet, and CTA tweet so you can A/B before you post - **Engagement log** (table) — impressions, likes, replies, click-throughs per tweet, refreshed every 30 min - **Reply queue** (table) — every reply you get, the draft response, and the priority Read `Steps` top-to-bottom. The hook tweet alone deserves 60 minutes of work; the rest of the thread is downstream. ## Outcome A launch thread that drives 100k+ impressions, 500-3000 click-throughs to your landing page, and 50-300 signups in the first 24 hours. Plus a reply queue you can keep mining for a week. **Estimated time:** 1-2 days drafting, 24 hours live, 1 week of reply mining **Difficulty:** intermediate **For:** Founders + indie hackers running a public launch on X. ## What you'll need Pre-register or install before you start. - **[X (Twitter)](https://x.com/)** _($8/mo for X Premium (verified, longer threads))_ — The platform itself. A verified account dramatically helps reach. - **[Hypefury](https://hypefury.com/)** _($19/mo for the Standard plan)_ — Schedule + auto-retweet thread tweets, draft in a workspace. - **[Typefully](https://typefully.com/)** _($12.50/mo for Pro)_ — Native thread editor with collaboration and preview. - **[Buffer](https://buffer.com/)** _($15/mo for Essentials)_ — Cross-post the launch thread to Bluesky and LinkedIn. - **[Plausible](https://plausible.io/)** _($9/mo starter)_ — Real-time click attribution with UTM-tagged links. --- # The template · 9 steps ## Step 1: Decide your single conversion goal _Estimated time: 30 min_ A launch thread without one explicit conversion goal turns into a vanity-metrics run. Pick one of: signups, paid conversions, waitlist additions, demo bookings, or top-of-funnel landing page visits. Optimise the entire thread for that one metric. Multiple-goal threads dilute the CTA and convert poorly. ### Tasks - [ ] Pick the single conversion goal: signup / paid / waitlist / demo / visit - [ ] Pick the proxy metric to track: 'signups in 24 hours from utm_source=x' - [ ] Set a target: 'I want 200 signups in 24 hours' - [ ] Build a UTM-tagged version of every link you'll share so attribution is clean ### Pointers - **[Tool]** [UTM builder](https://ga-dev-tools.google/campaign-url-builder/) > [!CAUTION] > **Gotchas** > > - If you have multiple goals, you're optimising for none. Pick one and live with the tradeoff. > - Likes are not a goal. Likes correlate weakly with conversion; reply count and CTR correlate strongly. > - Vanity goals like '10k impressions' aren't reachable without paid amplification. Set a goal for what your audience can give you organically. ## Step 2: Write 5-10 candidate hook tweets, then pick one _Estimated time: 1-2 hours_ The hook tweet is everything. It's the only tweet 90% of viewers will see, and it's the gate that decides whether they read tweet 2. Hooks should pose a tension, share a counter-intuitive number, or open a curiosity gap, not announce. 'I built X' is dead. 'For 2 years, I struggled with X. Then I built this.' is alive. ### Tasks - [ ] Draft 5-10 candidate hooks across 3 hook types: tension hook, number hook, curiosity hook - [ ] Each hook < 200 chars (the truncation point in feeds before 'show more') - [ ] Read each aloud — does it pose a question the reader wants to answer? - [ ] Pair each hook with a candidate first-image (screenshot, GIF, or before-after split) - [ ] Pick the one a stranger would screenshot to send to a friend > [!CAUTION] > **Gotchas** > > - Hooks that lead with the product name read as ads. Lead with the user pain or the result, name the product in tweet 2. > - Emoji-only opens get reduced reach. The X algorithm reads emoji as 'low-information content'. > - Truncation at ~200 chars matters. Anything past that is hidden behind 'show more'. Front-load the hook. ### Agent prompt for this step ```text Draft 10 candidate hook tweets for our X launch thread. Read the Drafts doc for product context + Brief. Generate hooks across these archetypes: 1. Tension hook: 'For X years I struggled with Y. Then I built this.' 2. Number hook: 'I shipped X in Y days for Z dollars. Here's the playbook.' 3. Curiosity hook: 'X is the most underrated way to do Y. We built the [tool] for it.' 4. Counter-intuitive hook: 'Everyone says X. We did the opposite, and Y happened.' 5. Concrete-result hook: 'In 30 days, this saved me X hours. Here's how it works.' Constraints: each hook 200 chars max, no exclamation marks, no 'thrilled to announce', no emoji-only first line, written in your founder voice not corporate voice. Output as a 10-row table: hook, archetype, why-it-works, char-count. Pair each with a suggested first-tweet image type (screenshot / before-after / GIF / metric). ``` ## Step 3: Draft the body of the thread (4-7 tweets, one screenshot each) _Estimated time: 3-4 hours_ The body tweets carry the value. Each tweet should be one self-contained idea + one supporting visual. Screenshots dramatically outperform plain text. The reader is scrolling, not reading; the visual carries the message and the text is the caption. Aim for 4-7 body tweets — fewer reads thin, more reads exhausting on a phone. ### Tasks - [ ] Outline the 4-7 ideas you want to land in order: problem, why-now, what-it-does, what's-different, proof, who-it's-for - [ ] Write each tweet < 240 chars (leaves room for hashtags or quoted-text) - [ ] Take a screenshot or GIF for every body tweet (the most-magical-moment shot, not a settings page) - [ ] Crop screenshots to 1920x1080 or 1200x630; X compresses heavily so make text bold and high-contrast - [ ] Avoid stock images at all costs; product screenshots and real-data charts win > [!CAUTION] > **Gotchas** > > - Tweets without images get 30-60% less engagement on X. Every body tweet needs a visual. > - Walls of text in a single tweet (400+ chars now possible with Premium) tank engagement. Stick to 240 chars and let the image do the work. > - Don't put the link in body tweets. Reserve the link for the CTA tweet (last in the thread). Mid-thread links cap reach. ### Agent prompt for this step ```text Draft the 4-7 body tweets for our X launch thread. Read the Drafts doc for the chosen hook + product brief. For each body tweet, output: 1. Tweet text (240 chars max) 2. Suggested image type: screenshot of [specific UI surface] / before-after split / chart of [specific metric] / GIF of [specific flow] 3. Why this tweet earns its slot The body tweets should land in this order, but adapt: - Body 1: the problem (in user terms, not product terms) - Body 2: why-now / what changed - Body 3: the actual solution / what the product does - Body 4: what's different from prior attempts - Body 5: proof — a metric, a quote, a result - Body 6: who it's for / who it's not for - Body 7 (optional): a personal note from the founder Constraints: every tweet stands alone (someone reading only that tweet should get one piece of value), no marketing voice, no exclamation marks, no '🚀'. ``` ## Step 4: Write the CTA tweet (the only one with a link) _Estimated time: 30 min_ X's algorithm caps reach on tweets with external links. The trick is to put the link in only ONE tweet — the CTA — and structure the thread so readers arrive at the CTA with intent. The CTA tweet should restate the value, name the action, and put the link on its own line so it's clearly clickable. ### Tasks - [ ] Write the CTA tweet: 1 sentence value-restate + 1 sentence action + the link on its own line - [ ] Add UTM tags to the link: ?utm_source=x&utm_medium=launchthread&utm_campaign=[productname] - [ ] Optionally: a P.S. tweet 1 hour later quoting your own CTA tweet with one more screenshot > [!CAUTION] > **Gotchas** > > - Link-bearing tweets get less reach. Don't sprinkle the link in 3 tweets; pick one. > - Link-shorteners (bit.ly, t.co) tank trust. Use the canonical URL with UTM params. > - If you're verified, a clickable link in the BIO is a backup if the algorithm soft-blocks the in-thread link. ## Step 5: Pick the optimal launch hour _Estimated time: 30 min_ X engagement curves vary by audience. Tech founders post Tuesday-Thursday between 9am and 1pm Eastern. Indie-hacker audiences skew later, 11am-3pm Eastern. Consumer audiences, 7-10pm Eastern. Look at when YOUR engaged followers post and tweet — that's the window that compounds. ### Tasks - [ ] Check your X analytics for the hours your audience posts most (proxy for awake-hours) - [ ] Pick a Tuesday-Thursday slot in that window - [ ] Confirm there's no major industry news / Apple keynote / OpenAI launch on the same day - [ ] Schedule the thread in Hypefury or Typefully for the chosen hour - [ ] Block 4 hours in your calendar starting at the post time for live reply work ### Pointers - **[Official]** [X analytics](https://analytics.x.com/) > [!CAUTION] > **Gotchas** > > - Posting at midnight your time means you're asleep when the engagement window opens. You miss the 30-min reply window that the algorithm rewards. > - Friday afternoons through Sunday afternoon are dead air for B2B audiences. Save them for consumer launches. > - If your audience is global, pick the timezone of your highest-engagement segment, not the average. ## Step 6: Pre-warm the launch and queue the soft-supporter list _Estimated time: 2-3 hours_ X's algorithm rewards engagement velocity in the first 30 minutes. A thread that gets 20 replies in the first 30 min outranks one that gets 100 replies spread over 24 hours. Pre-warming means quietly notifying 10-20 people who actually engage with you that the launch is coming, so they're ready to reply substantively at hour 0. ### Tasks - [ ] Build a list of 10-20 people in your network who reply substantively (not just like) - [ ] Send each a personal note 24 hours before: 'launching tomorrow at [time], would love your eyes on the thread' - [ ] Pin a 'launching tomorrow' tweet on your profile the night before - [ ] Optionally: cross-post the launch as a soft preview in 1-2 niche communities (HN, IndieHackers) the day before > [!CAUTION] > **Gotchas** > > - Mass-DMing to ask for engagement is detectable and gets you reach-throttled. Personal notes only, max 20 people. > - Pre-warming with 'please engage' phrasing reads as begging. Use 'would love your eyes' or 'curious what you think'. > - X penalises threads that get a burst of identical-pattern replies (vote brigade pattern). Real, varied replies beat coordinated identical ones. ## Step 7: Post the thread and run the live engagement window _Estimated time: 4-6 hours of active engagement_ The first 60 minutes after posting decide reach. Reply to every reply (yes, every) within 5 minutes. Engagement velocity in this window pushes the thread up the For You algorithm. Quote-tweet the most insightful reply at 30 min in, drives a second-wave bump. ### Tasks - [ ] Post the thread at the scheduled hour - [ ] Within 60 seconds: like the hook tweet from your second account or notification it to your team - [ ] Reply to every reply within 5 minutes for the first 60 min, 15 min for the next 3 hours - [ ] At 30 min in: quote-tweet the most insightful reply with your response - [ ] At 60 min in: post a follow-up tweet 'wow, this is taking off, here's a deeper dive on X' - [ ] Track impressions, replies, clicks, signups in the Engagement log every 30 min > [!CAUTION] > **Gotchas** > > - Going dark for 2+ hours during the live window kills the thread. Block the time on your calendar. > - Replying with one-word emojis ('🙏', '❤️') signals you don't care. Reply with sentences. > - Quote-tweeting your own thread later in the day to 'boost' it is detected and downranked. Quote-tweet other people's reactions instead. ### Agent prompt for this step ```text Track the live engagement window for our launch thread. Every 30 minutes during the first 6 hours, fetch from X analytics: - Impressions per tweet - Likes per tweet - Replies per tweet - Profile visits - Click-throughs (URL tweet) Append a row to Engagement log with timestamp + each metric. For each new reply, append a row to Reply queue with: - Replier handle - Content - Sentiment (positive / curious / critical / spam) - Priority (quote-worthy / normal / ignore) - Suggested response draft (the human will edit + post) Highlight any reply > 50 likes — it's a candidate to quote-tweet for a second wave. ``` ## Step 8: Mine the reply queue for 7 days _Estimated time: 30 min/day for 7 days_ Most launches die at hour 24 because the founder stops engaging. A great launch thread keeps mining the reply queue for a full week: turning insightful replies into testimonials, follow-up content, and product feedback. The compound effect is that the thread keeps surfacing in 'For You' for days. ### Tasks - [ ] Each day for 7 days: scan the Reply queue for new top-level replies - [ ] Reply to anything substantive within 24 hours of post - [ ] Save the 5 best replies to use as testimonials on the landing page - [ ] Save the 3 most-cited objections for the FAQ / improvement queue - [ ] On day 3 + day 7: quote-tweet your original thread with an update ('here's what changed in the last 3 days based on your feedback') > [!CAUTION] > **Gotchas** > > - Direct quote-tweets of your own original thread > 7 days later read as desperate. Cap at 2-3 follow-ups per launch. > - If a critical reply has > 100 likes, address it directly with substance. Ignoring it lets the criticism become the thread's narrative. > - Save EVERY reply with substance to the Reply queue surface; you'll mine it for landing page copy, FAQ, and the next launch. ## Step 9: Write the post-mortem and reuse the assets _Estimated time: 1-2 hours after day 7_ Every launch generates 50+ pieces of reusable content: screenshots, captions, replies, charts, founder voice. After 7 days, mine the launch into a launch-residue archive: the screenshots become landing page copy, the replies become testimonials, the engagement log becomes the data point for the next pitch deck. ### Tasks - [ ] Export the engagement log to find the highest-CTR tweet, the highest-reply tweet, the highest-quote-tweet - [ ] Save the 5 best screenshots to the asset library for landing page + future ads - [ ] Save the 5 best replies as testimonial candidates (DM each replier asking permission) - [ ] Note the conversion funnel: impressions → clicks → signups → paid - [ ] Draft the post-mortem in the Drafts doc: what worked, what to repeat, what to drop > [!CAUTION] > **Gotchas** > > - Likes are the easiest metric to game and the worst to optimise. Skip them in the post-mortem. > - If your CTR was below 0.5%, the thread didn't earn the link click — the hook or the body needs work for next time. > - Save screenshots BEFORE deleting any tweets in the thread. X loses the original image if you re-upload from a deleted tweet. ### Agent prompt for this step ```text Draft the launch-thread post-mortem. Read the Engagement log + the Reply queue + the original Drafts. Output a markdown doc: 1. By the numbers: impressions, total replies, total likes, total quote-tweets, total clicks, signups (or whatever the conversion goal was) 2. The best-performing tweet (CTR-wise, not likes) — and why 3. The worst-performing tweet — and why 4. Top 5 replies worth saving as testimonials (with handle + quote) 5. Top 3 critical replies worth addressing in next iteration of product or copy 6. What to repeat for the next launch 7. What to drop Constraints: data-led. Don't celebrate vanity metrics; analyse what drove the conversion goal. ``` --- ## Hand the template to your agent Paste the prompt below into your agent's permanent system prompt so the agent reads, writes, and maintains this workspace as you work through the steps. ```text You are an agent on the "Run a launch thread on X" playbook workspace at your-org/run-a-launch-thread-on-x. Your role: keep the four surfaces (Steps, Drafts, Engagement log, Reply queue) updated as the user runs the launch. Cadence: - Draft thread variants in the Drafts doc; the user picks the final. - During the live window, every 30 min: refresh impressions/likes/replies/clicks per tweet from the X analytics export and append to Engagement log. - For each new reply, append a row to Reply queue with: replier handle, content, suggested response, priority (quote-worthy, normal, ignore). - After 48 hours, draft a one-page summary in the Drafts doc: which tweet had highest CTR, which had highest reply-to-impression, what to repeat. First MCP tool calls: 1. list_surfaces(workspace_slug="run-a-launch-thread-on-x") 2. list_rows(workspace_slug="run-a-launch-thread-on-x", surface_slug="steps") 3. get_doc(workspace_slug="run-a-launch-thread-on-x", surface_slug="drafts") Do NOT post to X. Drafts go in the Drafts doc; the human reviews and posts. ``` --- ## FAQ ### How long should a launch thread be? 5-9 tweets total: hook + 4-7 body tweets + 1 CTA. Shorter than 5 reads thin; longer than 9 loses readers on mobile. Each tweet should carry one self-contained idea so a reader who quits halfway still got value. ### Should I use hashtags in my launch thread? No. Hashtags on X used to drive discovery, but post-2023 they're treated as low-signal spam by the algorithm and reduce reach. The exception is one or two highly-relevant community tags (#buildinpublic, #shipped) on the CTA tweet only. ### Do I need verified status (X Premium) to launch? Not strictly, but verified accounts get prioritised in the For You algorithm and can post longer threads (4000-char tweets). $8/mo is worth it for a launch. The bigger win is the longer-tweet ability for the CTA, not the blue check itself. ### What time is the best time to post a launch thread? For tech / B2B audiences: Tuesday-Thursday, 9am-1pm Eastern. For indie-hacker audiences: 11am-3pm Eastern. For consumer audiences: 7-10pm Eastern. The single best signal is when YOUR engaged followers are most active, visible in your X analytics dashboard. Post in their awake hours. ### Should I cross-post the same thread to LinkedIn and Bluesky? Yes, but adapt the tone. LinkedIn rewards founder-story narrative (1200-2000 chars, one image, one link). Bluesky is structurally similar to X but the audience reads more carefully — body tweets can be slightly longer. Don't copy-paste; rewrite for each platform's voice. ### Can my AI agents help write the launch thread? Yes, especially for: drafting 10 candidate hooks across archetypes (so you can pick the strongest), drafting the body tweets with screenshot suggestions, drafting CTA variants, tracking the live engagement log every 30 min, and queuing reply drafts for human review. The thread itself is posted by the human; agents draft and track.