How to Launch a Startup

"How do I launch a startup" gets typed by two different people, and most guides answer only one of them.

One has an idea and a notebook. The other already has a product sitting on a server, untouched by a single stranger.

The guides that answer this query treat both the same way: nine steps, ten steps, starting with "find an idea" and ending, eventually, with something called launch, buried near the bottom after incorporation and funding took up most of the page.

That order is backwards for the second person, and this post is written for them. Launching a startup is one dated, specific moment, and almost everything that makes that moment work has to already be true before it arrives.

Launch Is a Moment, Not the Whole Company

Most "how to start a startup" guides spend most of their length on what happens before a single user exists: the idea, the market research, the funding, the legal structure.

Launch shows up as step eight or nine, covered in a paragraph, right after "build your product" and right before "grow your business."

That ordering makes sense if nothing exists yet. It stops making sense the moment a founder already has something built, even roughly. What's missing at that point isn't another planning step. It's a dated event where strangers get to try the thing, and a way to know within days whether that went well.

CB Insights' most recent post-mortem analysis of failed, venture-backed startups found that 43 percent died from poor product-market fit, ahead of bad timing at 29 percent and unsustainable unit economics at 19 percent. CB Insights' most recent post-mortem analysis runs the full breakdown against 385 companies where a cause could be identified.

Almost none of those founders failed because they skipped a step on a nine-step list. They failed because nobody outside the building tried the product early enough, honestly enough, to say it wasn't working.

Launch is the event that produces that information, on a specific day, at a specific cost. Treat it as one milestone inside a longer process, never the finish line, and never something to delay until the product feels finished. It rarely does, and waiting for that feeling is its own failure mode.

Pick a bar for what counts as launched before choosing a date, something checkable rather than a feeling: ten strangers using the core loop without your help, say, or the first paying customer who found the product without a personal introduction.

Write that bar down now, whichever one fits the product. Everything below is what has to be true on the other side of it.

What Has to Be True Before Day One

Before touching a directory, a launch platform, or any paid placement, five things need to already be in place. Skip one and launch day turns into a fire drill instead of a data point.

1. A product a stranger can use without you standing over their shoulder: the core loop, working end to end, with no step that requires you to explain it out loud first.
2. A landing page with one action on it, and analytics wired to that action before the first visitor arrives. Without that, the day produces a feeling instead of a number.
3. A name and a one-line description short enough to fit a directory's character limit, tested on someone who has never heard you explain the product out loud.
4. One channel already warmed up before launch day: an email list, a community you post in regularly, a newsletter that already knows your name. Cold directory traffic alone converts at directory rates, which is a weak first bet.
5. A way to check three numbers by the end of day one: total visitors, conversions, and which source sent the good ones. Decide what "good" means in advance, in writing, before the numbers exist to argue with you.

Skip the fourth item and the launch depends entirely on strangers arriving through a directory homepage. That's the most fragile plan on this list, because it depends on placement nobody controls and on a day nobody chose the timing of.

For the full sequence, in order, with the specific numbers attached to each step, the launch checklist that sets the numbers walks through it from a blank page to launch day. What matters here is one level up from that: what has to already be true, and where the real risk sits before the submission forms open.

The riskiest gap, in practice, is usually the warm channel. A founder with zero list and zero community relationship is betting the whole launch on people who have never heard of them clicking a link on a page they didn't choose to visit that day. That bet loses more often than it wins.

Building that channel comes first, before the first submission goes anywhere.

Build These Three Things Before You Touch a Directory

Three things, built in order: a landing page with tracking already wired in, a submission-ready one-liner and image set, and a first traffic source that isn't a stranger.

The landing page first, because everything else points at it. Set it up with one call to action, wire an analytics event to that action, and load the page once from a phone on cellular data before sending anyone to it. A three-second load kills more launches than a bad headline does.

Run the math on that page before day one, not after. Unbounce's Conversion Benchmark Report, drawn from more than 57 million conversions across 41,000 landing pages, puts the median SaaS landing page conversion rate at 3.8 percent, against a 6.6 percent median across all industries. Unbounce's Conversion Benchmark Report breaks the number down by traffic source too.

Eight hundred visitors on launch day, at the SaaS median, is roughly 30 conversions. Half that rate, a real possibility on a first attempt, and it's 15. A plan for day one built around 100 signups means the page needs work before the day, not a bigger traffic push on the day itself.

Second, the submission kit: a one-line description, a 240-character version of the same line, a logo at three sizes, and one screenshot that shows the product doing something rather than a marketing illustration of it doing something. Directories and launch platforms all want a version of this kit, and building it once, correctly, saves five separate scrambles later.

Third, a first traffic source that already trusts you. An email list, even a small one, converts differently than cold directory traffic does. KickoffLabs' benchmark data on waitlist pages puts a healthy signup conversion rate at 20 to 40 percent of visitors, with anything under 10 percent flagged as a page or offer that isn't connecting. KickoffLabs' benchmark data on waitlist pages sets a similar bar for referral rate: 15 to 20 percent is healthy, under 5 percent means the incentive isn't working.

Set that against the 3.8 percent SaaS median above. A list of 400 warm names, even converting at the low end of that range, outperforms 800 cold visitors landing on a directory front page. Build the list, or the community relationship, before the launch date, not on it.

Choosing a Launch Day and What Happens On It

Pick a day with nothing else competing for attention: no product deadline, no travel, no second job pulling focus that week. Tuesday through Thursday tends to outperform Monday and Friday, since Monday is catch-up and Friday is already checked out.

The day itself runs something like this, hour by hour, for a founder who did the prep above.

1. Early morning: submit to the chosen directories and launch platforms first, before traffic starts, so listings are live when people arrive. The directories worth your time is the shortlist worth working from instead of submitting everywhere at once.
2. Mid-morning: send the email to the warm list. This is usually the highest-converting traffic of the day, and it should go out once the landing page and tracking are confirmed working, never before.
3. Midday: check the three numbers from the before-day-one list. Visitors, conversions, source. Check once, note it, close the tab; refreshing every ten minutes changes nothing except your stress level.
4. Afternoon: respond to every comment on every platform where the product was posted, personally, in your own voice. A Product Hunt alternative timed to your stage often matters more here than the biggest platform does, because a smaller, more targeted crowd of makers who actually comment beats a large one that scrolls past.
5. Evening: tally the day. Total visitors, total conversions, conversion rate by source, and one sentence on what surprised you, written down while it's fresh. Memory rounds everything up.

Paul Graham made the underlying point in February 2009, before most of today's launch platforms existed: the main value of whatever a founder launches with is as a pretext for engaging users. Paul Graham wrote in February 2009 that launching teaches a founder what they should have been building, and that they haven't really started until they have.

The product on day one isn't the final product. It's the thing that gets real reactions instead of imagined ones.

Where the audience is other founders and builders rather than a general one, launch platforms built for solo founders tend to convert better than general directories, since that audience already understands what it's looking at without an explainer paragraph.

Where Money Buys Certainty on Day One

Everything above is free or nearly free, and it should be tried first. Every channel in this list shares one weakness, though: none of them guarantee that a specific number of people see the listing today. A directory ranks by algorithm or by submission time. A community feed buries a post within hours.

One place online sells visibility on launch day outright instead of leaving it to be earned or gambled on. The Board is a public leaderboard where rank is sorted purely by the amount paid, and every listing on it is disclosed as a paid placement, not an editorial pick or an algorithmic one.

Reach on that leaderboard follows rank the way reach follows any ranked placement: a top position gets seen far more than a listing far down the list, and the price to hold a top position only rises within the current season once somebody takes it. For a launch day specifically, that means a listing bought a week ahead of the date is already collecting impressions before the first email goes out.

None of this replaces the prep earlier in this post. It's what to add once that prep is done and a founder wants one guaranteed source of eyes instead of another submission that might or might not get noticed on the day it matters most.

The First 90 Days: What to Watch and When to Change Something

Launch day produces a number. The next ninety days decide whether that number matters.

Track three things weekly, not daily: total signups or sales, conversion rate by source, and retention at week one. Checking daily on a small sample size mostly produces noise, and a founder chasing noise instead of a trend.

Click-through rate is the metric most people misread in these first weeks. A high CTR on a launch-day post can mean the headline promised something the product doesn't deliver, which shows up two weeks later as churn rather than as success. What counts as a good click-through rate breaks the benchmark down by channel, so a number that looks impressive on day one doesn't get mistaken for traction it isn't.

Set the decision rule before day one, so it isn't improvised under pressure in week three.

1. If visitors are healthy but conversion stays flat for two weeks straight, the problem is the page or the offer, not the traffic. Change the headline or the price before adding another channel.
2. If conversion is fine but week-one retention sits below a third, the problem is the product's first ten minutes, not the marketing. Fix onboarding before spending on a second launch.
3. If both numbers look fine but nobody outside the initial wave has heard of the product a month later, the launch was a spike, not a channel, and a repeatable second channel still needs to be found.

Ninety days out is also long enough to know whether the specific worry in the CB Insights data above applies here: whether what got built actually matched a market need, or whether launch day revealed a gap the pre-launch conversations missed. That answer changes the roadmap, not just the marketing plan.

More on distribution, and what it costs once the initial spike fades, sits under everything else filed under Launches, the category this post belongs to.

The next move after day one isn't a bigger launch. It's picking one number from the three above, watching it for two weeks, and being willing to act on what it says.