How to launch on Hacker News with Show HN

Someone analysed all 188,085 Show HN posts made between 2012 and April 2026. The median score was 2 points.

Not 20. Two. Half of every Show HN ever posted got the submitter's own vote and one stranger who was awake. The mean is 13.5, which only looks better because a handful of monsters drag it up, and 50 points already puts you in the top 6 percent.

That is the number to hold in your head before you launch, because most guides on how to launch on Hacker News are written from the survivor's seat. They describe what a front page post looked like and quietly skip the part where the same post, submitted on a different Thursday, would have died in /new with three points and no comments.

This is the other version. What a Show HN actually is, what kills one before anybody reads it, how the ranking math works, when to post, and what a front page is genuinely worth compared to the other places you could spend that day.

What the Show HN numbers actually look like

Here is the distribution, from that same 14-year dataset:

The competition has also gotten much worse. Show HN volume roughly tripled between 2019 and 2025, from about ten thousand posts a year to over twenty-eight thousand. That is something like seventy-five submissions a day fighting for a front page that still holds thirty slots and still turns over in hours.

So the realistic goal is not "hit the front page." The realistic goal is to not be one of the 50 percent that score 2, because almost every post in that bucket got there for a fixable reason: it was flagged, it was off-topic for Show HN, the title read like an ad, or nobody could try the thing without signing up.

What counts as a Show HN (and what gets it killed on sight)

Hacker News publishes the Show HN rules and almost nobody reads them. They are short and unusually literal.

Show HN is for something you made that other people can play with. That is the whole test. Things people can run on their computer or hold in their hands. Hardware counts if you post a video or a detailed writeup. A book counts if you post a sample chapter.

The off-topic list is where founders die:

• Blog posts
• Sign-up pages
• Newsletters
• Lists
• Landing pages
• Fundraisers
• "Other reading material"

Those are not banned from Hacker News. They are banned from Show HN. Post them as a normal submission and they are fine. Prefix them with Show HN and you are asking to be flagged by people who read the guidelines and take them seriously, which on that site is most of them.

Two more that catch people:

New features and upgrades generally are not substantive enough. You get roughly one Show HN per product, not one per release. Shipping a dark mode is not a launch.

If it is not ready for users to try, do not post it. A waitlist is the single most reliable way to get a Show HN flagged. So is a demo behind a signup wall. The guidelines say explicitly to make it easy to try "without barriers such as signups," and the community enforces that harder than the mods do.

The flip side is generous, and founders underrate it: the guidelines say a Show HN "needn't be complicated or look slick" and that the community is comfortable with early-stage work. Rough and working beats polished and gated, every time. This is the exact opposite of the instinct that serves you on a directory, where a tight product description gets you approved and a rough one stalls in review.

Why did my Show HN get no upvotes or disappear?

Same problem as Reddit: several different systems can bury a post and they all look identical from your side of the screen. The difference is that Hacker News tells you slightly more, if you know where to look.

Turn on showdead first. In your profile settings, set showdead to yes. Now you can see posts that have been killed, including your own. Without it, a dead post looks to you like a live post that nobody voted on, and you will spend a week diagnosing the wrong problem.

That last row is the honest one. Most Show HNs are not suppressed. They are simply the median.

A flagged post can be rescued with the vouch feature, but flagging takes 31 karma and vouching takes the same, and in practice flagged stories are almost never vouched back. Do not plan around it.

Vote rings are the mistake with the worst payoff. Dropping your HN link in a Discord and asking for upvotes does not just fail to help, it trips the vote-ring detector, which quietly stops the post ranking at all. No warning, no appeal, and you get to watch 40 points sit at position 90 forever. A burst of votes from accounts with no other activity looks exactly like what it is.

How Hacker News ranking actually works

The core formula has been public for years and it is simpler than people assume:

score = (points - 1) / (hours + 2) ^ 1.8

Your own vote is subtracted. Time is in the denominator with an exponent of 1.8, which is the "gravity" that pulls everything down the page whether it deserves it or not.

Three consequences worth internalising:

The first two hours decide everything. With gravity at 1.8, a post that gets ten points in its first hour outranks a post that gets thirty over eight hours. Early velocity is not a nice-to-have, it is the entire mechanism.

Text posts with no URL are penalised. Submissions without a link get their score multiplied by roughly 0.4. If your Show HN is a text post because you have nothing to link to, you have a problem that is not about writing.

Overheated threads get downweighted. When comment count runs well ahead of score, the software reads it as a flamewar and quietly detunes the post. A Show HN that turns into an argument about your pricing or your license can lose the front page precisely because it is generating discussion. Answering calmly is not just good manners, it is ranking.

There is also the second-chance pool, running since late 2014, which most founders have never heard of. Moderators pull older submissions that look like a good fit, reset their timestamps, and drop them onto the lower front page to see if the community bites. You can email hn@ycombinator.com to suggest one. They say it is fine to suggest your own and that they prefer suggestions about other people's work. One polite email, not a strategy.

Why boring Show HN titles win

The title field is where most launches are lost, and the fix is counterintuitive if you have spent any time writing marketing copy.

The template that works:

Show HN: [Name] – [literally what it does]

That is it. No adjectives, no superlatives, no benefit framing, no "the AI-powered platform for." Hacker News readers have a finely tuned allergy to ad copy and they express it with the flag button.

The good ones are longer and less exciting. That is correct. On Hacker News the title's job is qualification, not persuasion: you want the people who care about SQL-queryable Postgres analytics to click and everyone else to scroll past.

This is the exact inverse of a launch tweet, where curiosity and compression carry the post. Do not reuse that copy here. Also worth knowing: you get two hours to edit a submission, after which the title is frozen.

The first comment is your real launch post

The submission is a link and a title. The thing people actually read is the comment you leave on your own post within a minute of submitting it.

It should cover five things, in this order, in about two hundred words:

1. Who you are and why you built this. One or two sentences. The specific annoyance, not the market opportunity.
2. How it works. The technical shape. What stack, what the hard part was, what you tried first that did not work.
3. What is genuinely interesting about it. The bit an engineer would find non-obvious.
4. What does not work yet. Name the limitations before someone else does. This is not humility theatre, it is the highest-trust move available to you on that site.
5. One real question. Something you actually want an answer to, not "thoughts?"

The limitations point deserves emphasis because it feels wrong. Every marketing instinct says do not lead with what is broken. On Hacker News, "the Windows build is unsigned and I know it" earns more goodwill than any feature list, and it pre-empts the top comment that would otherwise be someone discovering it for you in public. It is the same instinct that makes an honest, problem-first post survive on Reddit, turned up a notch.

When is the best time to post a Show HN?

There is timing data, it is real, and it matters less than you want it to.

Across that 188,000-post dataset, the odds of scoring 50 or more points swing by roughly 4x depending on the slot:

The pattern is not "post during US work hours." It is the opposite: the best slots are the quiet ones, late Sunday and the edges of the weekend, when submission volume drops but enough readers are still around to give a good post early velocity. Weekday mornings have the most readers and also the most competition, and the competition wins.

But read the top row again. Ten point eight percent. The single best hour of the week is a one in nine shot at a merely good outcome. Timing is a margin, not a lever.

The constraint that actually matters is you. Post when you can sit in the thread for the next three or four hours, awake and in a mood to answer sharp questions patiently. A perfectly timed Show HN you sleep through is worth less than a mediocre slot you show up for.

Surviving the comments

The first three hours are the launch. Everything else is aftermath.

Reply to everything, including the terse ones. A two-word dismissal answered generously often turns into the most useful thread on the post. Hacker News readers watch how founders handle criticism and they vote accordingly.

Never argue about whether the criticism is fair. Answer the technical content and skip the tone. "Fair, here is why we made that tradeoff" ends well. "I think you have misunderstood what we are trying to do" ends with a detuned thread.

Fix things live. If someone finds a real bug in the first hour, fix it, deploy, and say so in the reply. It is the most reliable way to turn a lukewarm Show HN into a warm one, because it proves you are the person maintaining the thing.

Expect the licence question, the pricing question, and "why not just use X." That last one is not an insult, it is the most common form of engagement on the site, and a good answer to it is often the highest-voted comment on the post.

What is a Hacker News front page actually worth?

Time to set expectations properly, because this is where founders get hurt.

The figure quoted everywhere is 5,000 to 30,000 unique visitors in 24 hours from a front page post. Treat it as a wide, poorly sourced range rather than a forecast: it depends on how high you ranked, for how long, and whether that happened during US waking hours or at 4am.

The better-grounded numbers come from the GitHub side, because stars are public and countable:

• Roughly 1.4 GitHub stars per HN upvote for open-source projects
• About 92 percent of the total star gain lands within 48 hours
• The traffic half-life is about 24 hours, and by day seven you are back to baseline
• HN score explains only about 8 percent of the variance in stars gained, so the post is one input, not the cause

That last point is the honest one. A front page Show HN is a spike, not a slope. It is very good at two things: technical feedback you cannot buy, and early adopters who file real issues. It is bad at the third thing people expect, which is customers, because the HN audience skews toward people who would rather rebuild your product in a weekend than pay for it.

The SEO story is also more indirect than founders assume. The submission link itself is nofollow, so the post passes no authority on its own. What it does is get scraped: dozens of aggregators, newsletters, and mirror sites republish the front page, and those are where the followed links come from. No front page, no scrape, no links. That is a real benefit, but it is downstream of the thing you cannot control.

Compared to the other channels available to you:

The frequency column is the one to stare at. Show HN is effectively a once-per-product event. If it goes badly, your options are to wait about a year before resubmitting the same link or to ship something genuinely new. That makes it a terrible fit for the way indie makers actually work, which is shipping something worth talking about every couple of weeks.

A one-week Show HN plan

Day 1 to 3. Remove every barrier. Make the core thing usable with no signup. If that is impossible, build a seeded demo instance that is. Anything a stranger cannot try in fifteen seconds costs you more upvotes than polish gains.

Day 4. Write the first comment before the title. If you cannot write two hundred honest words about how it works and what is broken, it is not ready. That is useful information on day 4 and painful information on day 7.

Day 5. Test the domain and the account. Turn on showdead. If your domain is new or was bought secondhand, submit the bare link somewhere low-stakes and confirm it does not go instantly dead.

Day 6. Sleep on the title. Cut every adjective, then cut one more. Reread the official guidelines and ask honestly whether you have a thing people can play with or a page describing a thing.

Day 7. Post, then sit there. Submit, leave your comment immediately, stay in the thread for three hours. Saying you launched is fine. Asking for upvotes is the one thing that reliably ends the launch.

Day 8. If it worked, keep answering and write down every piece of feedback that came up twice. If it did not, do not resubmit. Find the real objection in the thread, fix it, and spend the energy on a channel you can use again next week.

Mistakes that kill a Show HN

• Posting a landing page, waitlist, or blog post as a Show HN.
• Requiring a signup before anyone can see the thing work.
• A title with adjectives, superlatives, or the word "revolutionary."
• Asking anyone, anywhere, to upvote it.
• Submitting and then leaving for six hours.
• Arguing with the first critical comment instead of answering it.
• Posting a minor feature release as a launch.
• Reusing your Product Hunt tagline verbatim.
• Deleting the post when it goes badly, then resubmitting the next day.
• Concluding, after one flat post, that the product is the problem.

Frequently asked questions

What is the difference between Show HN and a normal submission?

Show HN is specifically for something you built that readers can try themselves. Normal submissions are for anything else, including your own blog posts and writeups. Posting reading material under a Show HN prefix violates the guidelines and usually gets flagged, while the same link posted normally is completely fine.

How much karma do I need to post a Show HN?

None. Any account can submit. Karma thresholds exist for other things: 31 karma to flag or vouch for posts, 501 to downvote comments. A brand new account can post a Show HN, though a bare account with no comment history attracts more suspicion if anything else about the post looks promotional.

Why did my Show HN get flagged?

Most often because it was not a Show HN: a landing page, a waitlist, a newsletter, or a product nobody could try without signing up. The next most common reasons are a title that reads like an ad and a detectable burst of votes from accounts with no history. Turn on showdead to confirm it was actually flagged rather than simply ignored.

Can I repost my Show HN if it flopped?

Not usefully. Hacker News allows a resubmission after roughly a year if a story got no significant attention, and reposting sooner tends to get killed as a duplicate. If it got zero traction at all, one email to the moderators suggesting the second-chance pool is a reasonable single attempt. Repeated resubmissions are not.

What is the best time to post on Hacker News?

Statistically, quiet slots beat busy ones: Monday 00:00 UTC gives about a 10.8 percent chance of clearing 50 points, versus 2.6 percent at Thursday 06:00 UTC. The difference is real but small in absolute terms, and it matters far less than being awake and available in the thread for the first three hours.

How much traffic does a Hacker News front page post bring?

Widely quoted figures put it between 5,000 and 30,000 uniques in 24 hours, but the range is enormous and poorly sourced. What is measurable: for open-source projects, roughly 1.4 GitHub stars per upvote, with about 92 percent of the total arriving inside 48 hours. Plan for a two-day spike, not a step change.

Is Show HN better than Product Hunt or Reddit?

They do different jobs. Show HN gives the best technical feedback and the strongest signal for developer tools, but you get roughly one shot per product. Product Hunt gives a bigger, more predictable spike and a permanent listing. Reddit gives qualified traffic under a strict frequency cap. Most makers who do well use all three, plus something they can launch on repeatedly, which is worth mapping out before you pick one.

Should I launch on Hacker News if my product is not technical?

Usually not. The audience rewards things engineers can inspect, self-host, or fork. A well-made consumer app can do well, but the odds are worse and the comments are harsher. If your product has no technical surface for people to poke at, spend the day on a channel where the audience matches.

The part that matters

Show HN is a good channel with a bad shape for indie makers. The upside is real and the feedback is the best you will get anywhere for free. But it is once per product, the median outcome is two points, and the single best hour of the week still leaves you with a one in nine chance of a merely decent result.

So do it properly when you do it. Make the thing playable without a signup, write a boring title, put two hundred honest words in the first comment including what is broken, post when you can actually sit there, and answer every reply for three hours. That is most of the gap between the top 6 percent and the median, and it costs nothing but attention.

Then keep a channel where launching does not cost you your one annual shot. Submit your product on makers.page, quote-tweet the launch post with the same plain, technical sentence that works on Hacker News, and rank on real likes instead of a decay formula. The board resets every day, so a flat launch costs you a day and you go again tomorrow. Hacker News gets your one big moment. The rest of the year needs somewhere to go.

Related reading:

• How to launch on Reddit without getting banned
• Your launch flopped. Here's what to do next
• Where to list your indie product