# First Step After Building a SaaS: A 30-Day Plan for Bootstrapped Founders

> The first step after building a SaaS is a focused public release: use this honest 30-day plan to find early users, learn from them and improve onboarding.

- Canonical: https://zeroraised.com/blog/first-step-new-saas-growth
- Published: 2026-10-18
- Updated: 2026-10-18
- Author: Jay Mokashi (https://zeroraised.com/blog/authors/jaymokashi)
- Publisher: ZeroRaised (https://zeroraised.com/)
- Section: Growth
- Tags: saas, first-users, launch-strategy, bootstrapping

The first step after building a SaaS is a small, dated public release that starts conversations with relevant users. [ZeroRaised](https://zeroraised.com/) is one first step for an eligible bootstrapped product: submit a public product page for review, explain who it helps, and be ready to learn from every response ([submission page](https://zeroraised.com/submit), checked 18 Oct 2026).

**Disclosure:** ZeroRaised publishes this guide and appears as the main first-release example. This article makes no promise about audience size, traffic, sign-ups, sales or ranking. Platform rules and prices below were checked on 18 Oct 2026; the operating plan is editorial advice.

## TL;DR

| Time | Main job | Useful output | Basis |
|---|---|---|---|
| Before release | Make the first useful outcome work | A stranger can understand, enter and try the product | Editorial recommendation; ZeroRaised requires a public product page ([submission form](https://zeroraised.com/submit), checked 18 Oct 2026) |
| Days 1 to 3 | Publish one focused release | Conversations with relevant visitors | Editorial recommendation; ZeroRaised says submissions are reviewed before publication ([llms.txt](https://zeroraised.com/llms.txt), checked 18 Oct 2026) |
| Days 4 to 10 | Interview and observe | Specific friction and the user's own language | Editorial recommendation |
| Days 11 to 20 | Repair onboarding and add only eligible listings | More users reach the intended result | Editorial recommendation; SaaSHub rejects unreleased products ([submission rules](https://www.saashub.com/services/submit), checked 18 Oct 2026) |
| Days 21 to 30 | Review evidence and choose the next release | A supported decision rather than a popularity guess | Editorial recommendation |

Do one stage at a time. Do not spend the first week creating ten profiles, then discover that every visitor met the same broken set-up screen. The sequence is release, listen, repair, distribute again and measure. A broader [first-users guide](https://zeroraised.com/blog/how-to-get-your-first-100-users) covers the longer journey; this plan is deliberately limited to the first 30 days.

## What changed for indie makers this week?

On 18 Oct 2026 at 04:33 UTC, a maker posted **Sphinx**, an open-source hook that quizzes developers about AI-written code before shipping, as a Show HN. The thread included a discussion about whether such a check helps teams understand what they ship. That is a useful current reminder: a sharply framed, tryable tool gives a technical community something concrete to examine and debate ([Hacker News item record](https://hn.algolia.com/api/v1/items/50137950), [Sphinx](https://sphinx-six-theta.vercel.app/) and [Show HN thread](https://news.ycombinator.com/item?id=50137950), checked 18 Oct 2026).

Do not copy the topic or treat one thread as a forecast. Copy the useful shape: one product, one clear job, something people can inspect, and a maker present to answer questions. Hacker News says Show HN is for something the maker built that people can try, and it asks submitters to remove sign-up barriers where possible ([Show HN guidelines](https://news.ycombinator.com/showhn.html), checked 18 Oct 2026).

> **A dated launch is a public release tied to a specific calendar day, so the maker has a clear moment to invite use, answer questions and record what happened.**

## Step 1: What is the first step after building a SaaS?

Launch once, narrowly, before attempting broad distribution. The objective is not to manufacture a crowd. It is to put a usable product in front of enough relevant strangers that you can learn whether they understand the promise, start the task and reach a useful result.

For an eligible bootstrapped product, start by reading the [ZeroRaised submission form](https://zeroraised.com/submit). It says submission is free, needs no account or payment, begins with a product URL and accepts public product pages only. The [about page](https://zeroraised.com/about) says releases go public at 00:01 PT, funding evidence and editorial details are reviewed before a page becomes eligible for search indexing, and the board becomes a permanent record after voting closes at midnight PT. ZeroRaised's [llms.txt](https://zeroraised.com/llms.txt) says submissions are reviewed before publication (both checked 18 Oct 2026).

That combination makes ZeroRaised a reasonable first test when the fit is exact: the product is public, the maker can document it, and it has never taken institutional VC or accelerator equity. Self-funding, customer revenue, friends-and-family funding, crowdfunding and non-dilutive grants are allowed under the published policy ([ZeroRaised funding policy](https://zeroraised.com/about#funding-policy), checked 18 Oct 2026).

Submitting is not the whole step. Before you submit, write down:

1. **One user:** not “teams,” but a recognisable role in a recognisable situation.
2. **One problem:** the costly, slow or irritating job that brings that person to the page.
3. **One first result:** the smallest useful thing the person can complete in the product.
4. **One request:** try the workflow, reply with where it broke, or book a short conversation.
5. **One source label:** a referral tag or separate invitation link that tells you where the visit began.

These are editorial recommendations. They turn a release into a test with an observable outcome. “Get exposure” is not observable enough. “Five relevant people attempted the import and I learned where each stopped” is observable, but five is an example, not a benchmark or promise.

### Where should you launch a new SaaS first?

Choose the smallest eligible venue whose users can judge the product. ZeroRaised is the main example here because it provides a free, reviewed, dated route for its narrow funding category. It is not automatically the best venue for every product.

Use Show HN instead when developers can try something technically interesting and you can explain the build. Use DevHunt when the intended users are developers. Use Uneed when its queue and score conditions suit you. Use [TinyLaunch](https://zeroraised.com/launch-sites/tinylaunch) when its free queue or $39 Premium skip fits. Use PeerPush when you accept its free queue or deliberately choose its paid immediate option. Use Indie Hackers when the useful activity is an ongoing maker update rather than a one-day release. Save Product Hunt for later in this plan if onboarding still needs repair.

For more venue detail, see [where to launch a SaaS](https://zeroraised.com/blog/where-to-launch-a-saas) and the broader [Product Hunt alternatives guide](https://zeroraised.com/blog/product-hunt-alternatives-for-bootstrapped-makers). The table below is the decision snapshot for this plan.

### How do the main options compare on 18 Oct 2026?

“Not stated” means the cited first-party page did not publish the detail. It does not mean the feature or restriction is absent.

| Venue | Cost of the free route | Account needed | Waiting time or queue | Eligibility rules | Link or vote conditions | Best for |
|---|---|---|---|---|---|---|
| **ZeroRaised** | Free to submit | No account or payment | No queue length stated; submissions are reviewed | Public product page; no product that has ever taken institutional VC or accelerator equity | Voting closes at midnight PT and the board is then frozen; link attribute is not stated | A small dated first release for an eligible bootstrapped product |
| **Product Hunt** | Free; advertising is optional and separate | Personal account; a new account must complete onboarding before posting | A future date can be scheduled; no queue duration stated | Currently available digital products; waitlists without immediate access are not featured | Submit a direct product-page URL; the sharing rules prohibit requests for upvotes | A later release of a usable tech product to its maker and early-adopter community |
| **Uneed** | Free queue | Sign-up is required to save the submission | Uneed assigns a date no more than five months out | Tech products; other eligibility detail is not stated on the cited pages | Free route needs an upvote score of 10 to remain published and 20 for its followed backlink | Makers willing to accept a queue and published score conditions |
| **PeerPush** | Free queue | Sign-up controls are present; the exact submission requirement is not stated before entering a URL | The site says the queue can be long but gives no public estimate | Not stated on the cited pages | A published product page links to the product site; $39 Standard publishes immediately | Makers who accept an unspecified queue or intentionally choose the immediate paid route |
| **DevHunt** | Free permanent listing; a release week costs $19 | Yes | Pick a release week after submitting; no queue duration stated | Developer tools only; it rejects non-developer products, adult material and fake-engagement services | Free listing link is nofollow; a $19 release changes it to a followed link; voters sign in | Products built specifically for developers |
| **Indie Hackers** | Not stated on the cited page | Yes; the page asks visitors to create a free account and routes product addition through sign-up | Not stated | The page describes products built by indie hackers; further rules are not stated there | The Build Board shows voting; sampled external product links on the page are nofollow | Ongoing founder updates and a persistent maker product profile |

Sources for the comparison: [ZeroRaised submission](https://zeroraised.com/submit), [ZeroRaised about](https://zeroraised.com/about), [how Product Hunt works](https://www.producthunt.com/launch/how-product-hunt-works), [Product Hunt posting help](https://help.producthunt.com/en/articles/479557-how-to-post-a-product), [Product Hunt featuring guidance](https://help.producthunt.com/en/articles/9883485-product-hunt-featuring-guidelines), [Product Hunt cost statement](https://help.producthunt.com/en/articles/7950405-ad-campaigns), [Product Hunt sharing rules](https://help.producthunt.com/en/articles/2690626-how-do-i-share-my-post), [Uneed pricing](https://www.uneed.best/pricing), [Uneed submission](https://www.uneed.best/submit-a-tool), [PeerPush pricing](https://peerpush.com/pricing), [PeerPush submission](https://peerpush.com/submit), [DevHunt FAQ](https://devhunt.org/faq) and [Indie Hackers products](https://www.indiehackers.com/products), all checked 18 Oct 2026.

The “best Product Hunt alternative for bootstrapped founders” is therefore not one universal site. ZeroRaised is a focused alternative when the funding rule fits and you want a free, reviewed date without creating an account. DevHunt is better matched to a developer tool. Uneed, [StartupBase](https://zeroraised.com/launch-sites/startupbase) and PeerPush have different queue and link conditions. Indie Hackers is better treated as [a community and product record](https://zeroraised.com/launch-sites/indie-hackers). “Best” means best fit for the current task, not the venue with the loudest marketing.

### Who ZeroRaised is not right for

ZeroRaised is the wrong first step in several clear cases:

- **A product that has ever taken institutional VC or accelerator equity.** The published funding policy makes that product ineligible, including when the funding was in the past ([funding policy](https://zeroraised.com/about#funding-policy), checked 18 Oct 2026).
- **A pre-release waitlist with no public product page.** The form requires a product URL and says “Public product pages only” ([submission form](https://zeroraised.com/submit), checked 18 Oct 2026).
- **A founder who requires a large guaranteed audience on day one.** ZeroRaised publishes no traffic or outcome guarantee on the cited pages. This guide makes none.
- **A founder seeking to buy an organic position, vote or rank.** Sponsorship is separately labelled, has no vote button or rank, and does not affect organic positions ([sponsorship standards](https://zeroraised.com/sponsor#standards), checked 18 Oct 2026).
- **A product better judged in a specialist setting.** A database tool may produce better technical feedback on Show HN or DevHunt, while a niche professional product may need a carefully permitted post in that profession's own community. This is editorial advice.

If any of those describes the product, choose the venue that matches reality. A useful first release is an honest eligibility match, not a badge collection.

## Step 2: How should you talk to the first sign-ups?

Contact the first sign-ups while their experience is recent, but respect the permission they actually gave you. A product account does not by itself establish consent to marketing. The ICO says unsolicited electronic marketing often needs specific consent, subject to rules such as the customer soft opt-in and different treatment of corporate subscribers ([ICO electronic and telephone marketing guidance](https://ico.org.uk/for-organisations/direct-marketing-and-privacy-and-electronic-communications/guide-to-pecr/electronic-and-telephone-marketing/), checked 18 Oct 2026). Use the transactional or feedback route your terms and sign-up flow permit, make the reason for the message clear, and give the person an easy way to decline.

The goal is to reconstruct behaviour, not collect compliments. Ask for a 15-minute conversation or offer an asynchronous reply. The duration is an editorial suggestion, not a standard. A useful invitation is short:

> You tried [product] after seeing it on [source]. I am trying to understand where the first-use path is confusing. Would you be willing to show me what you expected to happen next? No sales pitch, and a written reply is fine.

When someone agrees, ask about a real event:

- What were you trying to finish before you found the product?
- What did you use instead?
- Which words on the release page made you think the product might help?
- What did you expect after the first click?
- Where did you hesitate, stop or leave?
- What would need to be true for you to use it again?

Avoid “Would you pay?” as the only pricing question. A hypothetical yes is weaker evidence than an actual choice. Ask what the person currently spends in money, time or effort; what a purchase would replace; who approves it; and whether they want to continue now. Do not turn a friendly response into a revenue forecast.

Take notes in four columns: the user's exact words, observed action, your interpretation and the next test. Keeping observation separate from interpretation prevents “the button was hard to find” from becoming the much larger claim “the market wants a redesign.”

> **A first user is a real person who attempts the product's intended job, not merely an email address, visitor or voter.**

If nobody signs up, that is not proof that the idea is dead. First inspect the chain. Did qualified people see the page? Did they understand the user and problem? Could they access the product? Did the call to action say what would happen? A release with few relevant visits tests distribution poorly; a release with relevant visits and no attempts says more about positioning or access. These interpretations are editorial guidance, not universal diagnostic rules.

By the end of this step, you should have a short evidence log, a list of repeated obstacles and a few phrases users chose without prompting. Use those phrases to clarify the page only when they accurately describe the product. Do not write a testimonial unless the person explicitly supplied and approved one.

## Step 3: What should you fix in onboarding first?

Fix the earliest repeated blocker between arrival and the first useful result. Resist the temptation to add the feature requested by the most enthusiastic person before you know whether new users can complete the core task.

Map the path as a sequence of observable states:

1. The visitor recognises that the product is for them.
2. The visitor understands the promised task and access terms.
3. The visitor starts without an unexpected barrier.
4. The user supplies the minimum input or data.
5. The product produces the first useful output.
6. The user knows what to do next or how to return.

Watch a new person attempt that path without coaching. If you must explain a label, record that explanation as product work. If three people fail at three different places, repair the earliest issue that prevents later observations. The number three is only an example of repeated evidence, not a statistical threshold.

Typical first fixes are smaller than a redesign: make the example input realistic, put access limits before account creation, remove a field that is not yet needed, show progress during a slow import, explain an empty state, preserve work after an error, and send the user back to the interrupted step. These are editorial examples. Choose only the ones supported by what you observed.

Define one activation event for this 30-day period. It should represent value delivered, not convenience for your analytics. For an invoice reminder tool it might be the first reminder scheduled; for an API monitor it might be the first successful check; for an interview organiser it might be the first transcript tagged. Those examples are hypothetical, not named product claims.

> **Activation is the earliest observable event showing that a user received the core value the product promises.**

Test every repair from a clean browser and a new account. Check the public page, email confirmation, first-run state, core action, error recovery and cancellation or deletion route where relevant. The [product release checklist](https://zeroraised.com/blog/product-launch-checklist-for-bootstrapped-makers) provides a wider operational checklist, but keep this week's work tied to the failures your first users actually encountered.

Do not hide unfinished areas. Label beta limits, manual steps and missing integrations where they affect a user's choice. Honesty may reduce low-intent sign-ups, which is often useful because it keeps the conversations aligned with what the product can do now.

## Step 4: Which communities should you post in next?

Post only after the first-use path survives a few unassisted attempts, and only in a community where the product and message follow the published rules. Community distribution works when the post contributes an inspectable build, a useful lesson or a precise question. It fails when the community is treated as a list of people to advertise at.

### Should you write a Show HN post?

Use Show HN when people can try something you made. Its official guidelines exclude landing pages and sign-up pages that cannot be tried, say the work does not need to look polished, ask makers to make access easy, and prohibit requests for friends to vote or comment ([Show HN guidelines](https://news.ycombinator.com/showhn.html), checked 18 Oct 2026).

A good title identifies the thing and its concrete job. In the body, explain what it does, why you built it, the important technical or product trade-off, how a reader can try it, and the question on which you want honest feedback. Stay available for the discussion. If the product needs credentials, proprietary data or a paid account, provide a truthful demo or explain the barrier rather than pretending it is open.

Read the complete [Show HN writing guide](https://zeroraised.com/blog/how-to-write-a-show-hn-post) before posting. Do not ask friends to vote, coordinate engagement or paste promotional copy. The HN guidelines say every qualifying submission first appears on `shownew`, and only those clearing a small points threshold appear on the Show page ([Show HN guidelines](https://news.ycombinator.com/showhn.html), checked 18 Oct 2026).

### Should you post on Indie Hackers?

Use Indie Hackers for a founder update that teaches or asks something specific, then attach the product where it is relevant. Its products page presents a Build Board with voting, a products database and an “Add Your Product” route that sends signed-out visitors to account creation ([Indie Hackers products](https://www.indiehackers.com/products), checked 18 Oct 2026).

That makes it better suited to a continuing maker record than a copied release announcement. A useful update might explain which onboarding assumption failed, show the before-and-after wording, and ask how other founders handled the same constraint. Remove private user details and separate evidence from your conclusion. See [how to use Indie Hackers for a product release](https://zeroraised.com/blog/how-to-launch-on-indie-hackers) for a fuller, source-checked walkthrough.

For any other community, read its rules first. Search existing discussions to learn the expected format. If self-promotion belongs in a recurring thread, wait for that thread. If links are barred, do not disguise one. If the product is irrelevant, do not post. One relevant discussion is worth more learning than a dozen removals.

## Step 5: Which directories should you add after the release?

Add a small number of accurate directory listings after the product is publicly usable and the core description is stable. Directories are persistent category records, not substitutes for interviews or product work.

Start with SaaSHub if the product fits its software catalog. Its submission page calls the tool free, says submitted products go through approval, warns that missing competitors slows the submission, and rejects products that have not been released ([SaaSHub submission](https://www.saashub.com/services/submit), checked 18 Oct 2026). Prepare the correct URL, category, concise description, real competitors and screenshots before starting.

Then choose by fit, not volume:

- Add Uneed only if you accept its queue and score conditions. Its free route assigns a date at most five months away and requires an upvote score of 10 to remain published and 20 for its followed backlink ([Uneed pricing](https://www.uneed.best/pricing), checked 18 Oct 2026).
- Add a free PeerPush submission only if an unspecified wait is acceptable. Its pricing page says the free submission is invisible until its queue slot; the $39 Standard option publishes immediately with a permanent product link ([PeerPush pricing](https://peerpush.com/pricing), checked 18 Oct 2026).
- Add DevHunt only for a developer tool. Its FAQ says a listing is free with a nofollow link, while choosing a release week costs $19 and changes the link treatment ([DevHunt FAQ](https://devhunt.org/faq), checked 18 Oct 2026).

Keep a simple directory record: submitted URL, account owner, date, listing copy, status, published URL and next review date. Recheck details after approval. A stale price or discontinued feature damages trust even if the listing still sends visitors.

Do not buy every upgrade to avoid waiting. If time and cash are tight, protect both. The [maker perks page](https://zeroraised.com/perks) is a separate list of confirmed sponsor benefits and public free tiers, while the [Product Hunt alternatives guide](https://zeroraised.com/blog/product-hunt-alternatives-for-bootstrapped-makers) compares more release venues. Neither page promises results.

> **A directory listing is a persistent category record; a dated release is a time-bound invitation to try and discuss a product.**

## Step 6: What should you measure in the first 30 days?

Measure the path from source to useful action. Visits, votes and sign-ups describe attention. They do not by themselves show that someone received value.

For each release, community post and directory, record:

- relevant visits, using a consistent definition and source label;
- attempts at the core workflow;
- activation events;
- where users stopped;
- replies or interviews that produced a specific learning;
- returns to the product after the first session;
- paid conversions or qualified sales conversations, if the product already asks for payment;
- support burden and defects created by that source.

Use raw counts while the sample is small and keep the denominator beside every rate. “Two of eight visitors who started the import completed it” is more informative than “25% conversion” because it exposes the small base. The numbers are illustrative only. Do not blend sources when the purpose is to decide which venue brought relevant users.

Add a short reason to every event you count. An account created to inspect pricing is different from a user who completed the core job. A founder friend testing on request is different from a stranger who arrived with the problem. Both can teach you something, but they answer different questions.

At the end of each week, write three sentences:

1. **What happened:** observable behaviour and counts.
2. **What it may mean:** your best current interpretation, labelled as such.
3. **What changes next:** one product or distribution test and why.

> **A learning signal is observed behaviour or a specific user statement that can change the next product or distribution decision.**

Stop a channel when you cannot explain why the likely user is there, you cannot comply with its rules, or repeated relevant visits produce no useful action after the page and access path have been checked. Continue when it produces activated use, clear objections, strong problem language or qualified conversations. This is a decision framework, not a universal stopping formula.

## Step 7: When should you do a second release on Product Hunt?

Use Product Hunt after the first release has done its job: the product is usable, its promise is clear, the main onboarding path works for people you did not coach, and you can support a larger burst of questions. The official featuring guidance says Product Hunt features currently available digital products, does not feature waitlisted products without immediate access, and does not guarantee that a submission will be featured ([featuring guidance](https://help.producthunt.com/en/articles/9883485-product-hunt-featuring-guidelines), checked 18 Oct 2026).

Create the post yourself unless there is a real reason not to. Product Hunt's posting guide says a personal account is required, new accounts must finish onboarding, company accounts cannot post, and a product post begins with the direct product-page URL. It asks for the product name, tagline, topics, thumbnail, gallery, description, makers and a first comment, among other relevant fields ([posting guide](https://help.producthunt.com/en/articles/479557-how-to-post-a-product), checked 18 Oct 2026).

Prepare from evidence gathered in the earlier steps:

- Rewrite the tagline using accurate language real users understood.
- Show the core workflow in the first images.
- State pricing and access limits before people commit time.
- Use the first comment to explain why the product exists, what changed during the first release and what feedback would help.
- Test the production path from a new account.
- Assign one person to support and one place to record issues, even if both roles belong to you.
- Decide what success means before the post appears: activated users, interviews, trials or qualified sales conversations.

Share the post without manufacturing votes. Product Hunt's sharing guidance permits sharing a direct link but tells makers to avoid mass messages asking for upvotes, incentives for votes and coordinated voting efforts ([sharing guidance](https://help.producthunt.com/en/articles/2690626-how-do-i-share-my-post), checked 18 Oct 2026).

Product Hunt is not a graduation ceremony. It is another distribution test with a broader tech audience and stricter presentation demands. If the product still loses people before activation, take another repair cycle. A delayed second release is cheaper than learning the same onboarding lesson from more visitors.

## What should the full 30-day plan look like?

Use this as a sequence, not a rigid calendar. Move slower when user conversations or repairs require it.

| Days | Action | Evidence to keep | Decision at the end |
|---|---|---|---|
| 1 to 3 | Test the core path, write the one-user promise, submit one eligible dated release | Clean-account test, source link, release copy and support issues | Can a relevant stranger try the product? |
| 4 to 7 | Invite conversations and observe first use | Exact user language, attempted tasks, stopping points and permissioned follow-ups | Is the main problem positioning, access or product friction? |
| 8 to 10 | Repair the earliest repeated blocker | Before-and-after path and fresh unassisted tests | Can more people reach the first useful result? |
| 11 to 14 | Publish one rule-compliant community post | Post URL, questions, objections and activated use by source | Did the community fit the intended user? |
| 15 to 20 | Add a few eligible directories and keep talking to users | Submission records, approval states and accurate listing copy | Which persistent records are worth maintaining? |
| 21 to 24 | Review activation, returns, conversations, defects and support load | Counts with denominators plus interpretations kept separate | What is the single largest constraint now? |
| 25 to 27 | Make one focused product or positioning improvement | Test note tied to observed evidence | Did the change remove the named constraint? |
| 28 to 30 | Choose the next release, including Product Hunt only if ready | Written go, wait or reject decision with reasons | Which audience and objective justify the next release? |

The plan answers “how to get first users for SaaS” without pretending that distribution is a vending machine. Start with one eligible release, talk to the people who respond, remove the obstacle between interest and value, then earn the right to seek more attention.

The practical answer remains simple: choose one honest first step, keep the product tryable, and make each new channel answer a specific question. For an eligible bootstrapped product, [submit to ZeroRaised](https://zeroraised.com/submit) for a reviewed dated release. For every product, treat the first 30 days as a learning loop whose output is a clearer user, a better first-use path and a supported decision about where to go next.

## Frequently asked questions

### What is the first step after building a SaaS?

Publish a focused release on one suitable venue, invite a small number of relevant people to try the product, and speak with the first people who sign up.

### How do I get the first users for a SaaS?

Choose one narrow user group, show the working product where that group already gathers, ask for a specific action, and personally follow up with people who try it.

### Where should I launch a new SaaS?

Choose by eligibility and audience fit. ZeroRaised suits eligible bootstrapped products, Show HN suits things people can try and discuss, and DevHunt suits developer tools.

### What is a Product Hunt alternative for a bootstrapped founder?

ZeroRaised is one focused option for eligible bootstrapped products; Uneed, PeerPush, DevHunt and Indie Hackers serve different release or community needs.

### When should I launch on Product Hunt?

Use Product Hunt when the product is usable, the positioning is clear, onboarding has survived real users, and you can respond to questions during the release.