Should You Launch Before the Product Is Ready? Four Gates to Pass
Launch before the product is finished, but not before a stranger can complete the one thing it exists to do. A rough product with a working core, an honest description and a way to reach you is ready for a first launch. A page with a sign-up form and nothing behind it is not. Hold back the launch chances you can only use once until the core works.
What does ready mean for a first launch?
Ready is not finished. It is a short list of gates, and a product that clears all four can be launched with rough edges everywhere else.
- The core task works end to end for someone you have never met. If the product turns receipts into spreadsheets, a stranger can upload a receipt and get a spreadsheet back without your help.
- Nothing can harm the person using it. No lost data, no exposed accounts, no payment that charges and delivers nothing. This is the one gate you cannot trade for speed.
- The page says what is missing. A visitor should learn about the gaps from you, not by hitting them.
- You can answer within a day. A launch without a maker who replies is a page, not a conversation.
Everything outside those gates, such as polish, settings, a second language and a perfect onboarding, can follow the launch. How to present the gaps is covered in how to launch an unfinished product.
Which launch chances can you only use once?
Waiting costs little where you can launch again and a lot where you cannot. Before you spend a slot, check whether it comes back.
| Place | Can you use it again? | What its pages say |
|---|---|---|
| Show HN | Rarely | Hacker News’s FAQ says a small number of reposts is fine if a story had no significant attention in the last year or so, and that otherwise reposts are buried; the Show HN rules say a point release is usually not substantive enough |
| Smol Launch free launch | No | Its free launch covers a product’s first launch; later ones use a paid tier |
| This board (ours) | Not directly | The rules ask for one launch per product, and a website cannot be launched twice within seven days; a published launch is final, so say plainly what is unfinished before you publish |
| Product Hunt | Not confirmed | Its pages refused our automated check, so read its current rules on repeat launches |
The practical lesson: keep Show HN and any free first launch for the day a stranger can use the core. On this board a published launch is final, so a rough first version is fine only if you launch your product free with an honest description that says what is unfinished; what you learn afterwards belongs on your own page.
What does waiting cost you?
Every week spent polishing is a week without a stranger telling you which part you misjudged. Makers tend to polish what they can see and ignore what visitors trip over, so the list in your head and the list in the user’s head rarely match. Show HN’s guidelines go further: they say a project needn’t be complicated or look slick, and that the community is comfortable with work at an early stage. They also say that if your work isn’t ready for users to try, you should not post a Show HN and should come back once it is.
A workable rule is to fix the launch date first and cut scope, not the date, when it nears. Cut features until the four gates pass, then launch. If the gates cannot be met by the date, move the date, and write down which gate was the cause.
When should you wait even though it is only half finished?
- You handle money, health, legal or other private data. A rough interface is forgivable, a rough security model is not. Review it before strangers arrive, and see vibe coding security if AI wrote much of it.
- It ships through a store that gates access. Google Play’s testing page, read on 13 November 2026, says developers with personal accounts created after 13 November 2023 must run a closed test with at least 12 testers opted in for 14 continuous days before applying for production access. Apple’s TestFlight page says the first build offered to external testers may be sent to App Review. So “ready” for those includes the store’s clock. The Google Play publishing guide covers the steps.
- The core task needs your help every time. If you must be in the loop to make it work, it is a service, so launch it as one, or wait.
- You would be embarrassed by the data you would collect. If you cannot say what you store and why, write that first.
How can you test readiness this week?
Give the address to five people who have never seen the product, say nothing about how it works and watch them try the core task, on a call or in person. Write down where each one stalls. If two of the five stall at the same step, that step is your next job, not a reason to launch tomorrow. If all five finish, you are ready, whatever the colour of the buttons.
Five is a convenient number, not a statistic, so repeat the test after each fix. If you do not know where to find five strangers, look in the communities where the problem is discussed, or at the ideas board for people describing what they wish existed. The what is an MVP guide explains how small a first version can be.
Frequently asked questions
Is it bad to launch a product with bugs?
Small bugs outside the core task are normal at launch, and a visible list of known issues builds trust. Bugs that lose data, expose accounts or break the core task are different, and you should fix those first. The test is whether a stranger can still finish what the product is for.
Can I launch just a landing page?
You can post a landing page for a product that is not built, but say so plainly, and expect fewer places to accept it. Hacker News’s Show HN rules list sign-up pages and landing pages as things not to post. A launch board is more forgiving, if the description says the product is early.
How rough can a first version be?
Rough in looks and features, but working at its core. A visitor should be able to try the one promised thing in minutes, without you. Hacker News’s guidelines say a Show HN need not look slick, so design polish is not the bar. A broken main path is.
Should I wait for more features before launching?
Usually not. More features rarely fix the problem that matters, which is whether a stranger understands and finishes the core task. Launch with the smallest thing that works, collect what people ask for and add those. Features nobody requested are guesses.
What if I launch too early and it flops?
A quiet first launch costs you one slot and teaches you what confused people. Repeat chances exist on some platforms and not on others, so save the one-shot ones. Read the notes on relaunching a product before you decide what to do next.
Written by Post your build ideas with AI assistance and edited for this site. Where a page names another product, the facts came from that product’s own pages on the date shown above; terms change, so check them before you rely on them. About the publisher.