Show HN Launch Strategy and Comment-Thread Defence

A Show HN reaches the front page roughly one time in twenty, and the decision is made inside the first forty minutes in /newest, where four to six votes from readers who owe you nothing separate page one from page nine. Our work is aimed at that window and at the six hours after it, because the thread, not the point count, decides whether the traffic becomes accounts or a post-mortem. What happens before the post goes up: We read the product the way the thread will read it: no account, cold browser, no context, looking for the point at which a reader gives up. If that point exists, the launch is postponed. Show HN has a requirement most launches quietly break, which is that the thing has to be usable immediately. A waitlist, an email gate, a book-a-demo button or a video standing in for a working page invites the comment that ends a launch, and that comment arrives inside four minutes. Title work is written, not brainstormed. Three to five candidates with the reasoning for each, checked against what has and has not survived on comparable submissions in the Algolia HN Search API over the previous two years. Superlatives and marketing verbs get a submission flagged by readers long before a moderator sees it. The thread is the deliverable: We do not write your comments. A founder's own voice, awkwardness included, reads as real and a rewritten one does not. What we build instead is an objection brief: the twelve to eighteen questions the thread will actually ask, ranked by likelihood, each with the facts needed to answer in two sentences. Pricing. What you do with the data. What happens when the upstream API breaks. Why not Postgres and a cron job. Why that licence. Who tried this in 2019 and why they stopped. We hold a live channel for the first six hours. Target for a first reply to any top-level comment is under ten minutes. A hostile top comment answered well converts better than the launch itself; the same comment left sitting for three hours sets the tone of the whole page and then stays indexed. What we will not do: We do not post from our own accounts, vote on your submission, ask anyone to vote, or organise a group to arrive at the same time. Ring detection costs the account, not the post. We do not message moderators asking for a re-rank; the second-chance pool exists and works on its own terms. No paid distribution, no press release, no syndication. Who this is not for: Health and wellness products carrying clinical claims. The thread will audit the evidence in public, that page will outrank your own inside a week, and if the evidence is thin the audit becomes the search result. Lead-capture real estate products. Anything whose demo is a form. Companies who cannot put an engineer in the thread on the day. We say all of this in the first call. What the numbers have actually looked like: A top-five position held for three hours has produced fifteen to thirty-five thousand sessions across the launches we have run. A mid-page finish, which is the more common outcome, produces two to four thousand. Sign-up conversion for developer tooling has run one to three percent of sessions. For anything that needs a sales conversation, treat the traffic as awareness and the thread as the durable asset, because it is quoted back to you years later. Not included: Product changes, landing page build, documentation writing, or any other launch channel. We will tell you plainly which of those is blocking the launch, and we would rather delay than post into a gate.