Talk to people before you launch again. In the reports we read where builders give numbers, first users mostly came from people answering others directly, in public threads or one to one, in places where the problem was already being discussed. Launch sites brought real customers for some of them and a day of attention for others. It often takes weeks of daily effort, and some people do all of it and still hear nothing.
You’re in a big club
In November 2026 we counted 124 Reddit and Hacker News posts from builders complaining about marketing. 61 said they couldn’t advertise or bring the product to market. 41 launched to nobody. None complained about how their content looked. The Reddit ones come from a public archive. An AI reader did the counting, with no second check, and we picked posts that used marketing words, so treat that as a hint and not a study. We kept the totals, not the full list. One post about a year of work and almost no users drew 676 comments.
One person on Hacker News put the problem in a line: “the skills for building and the skills for getting attention are almost completely disjoint”. You got good at the first one. Nobody taught you the second.
What builders with real numbers report
Advice on this is everywhere and mostly untested, so we looked for people who gave numbers. These are their own accounts. Nobody checked them, and people who failed rarely write the post, so read this as what some people report, not what will happen to you.
| Who | What they did | What they report |
|---|---|---|
| Buildpad, a two-person team selling a tool to founders | Seven months of projects marketed from guides got nothing. Then they asked a subreddit first, messaged those people when the first version shipped, posted a few times on Reddit, and worked founder communities on X every day | 100 users in about two weeks with no money. Free until about 150 users, when a Product Hunt launch brought 475 new users in a day and the first payers. 4,900 users and $9,000 in the six months after launch |
| PostKing, which sells an AI tool for marketing content | Answered questions in communities, and mentioned the product only when someone asked for a tool | 42 of its first 100 customers from Reddit and 38 from Product Hunt and Peerlist. Directories brought 3 sign-ups and nobody paying |
| A builder on LinkedIn | Messaged people directly about a problem they shared, not about the tool | 35 users in two weeks |
| An iPhone app maker | Posted short developer diary videos on LinkedIn while building | First 100 users within days of launch, and a short spell in the App Store’s top paid list |
| A builder giving talks | Spoke at small local meetups, again and again in the same scene | First users came from the conversations afterwards. One talk did little, five made a network. No count given |
| A 10-day-old side project | $500 on five versions of a Reddit ad, plus a Product Hunt launch | 753 visits and 6 sign-ups from the ads, and 8 visits from Product Hunt |
| A Product Hunt launch | Launched on Product Hunt | 400 sign-ups and 1 paying customer |
| A founder buying ads early | Paid for ads early on | $2,000 spent, no customers |
A thread on Indie Hackers asked what distribution flopped. Of the first 18 stories, 7 were directories and launch lists, 5 were Reddit and communities, and 3 were cold email. None blamed how the product looked. Reddit is on both lists, and we can’t tell from that thread how those people posted. The reports that worked describe weeks of answering people before mentioning a product.
What we take from it
This part is our reading, not a finding. First users mostly came from people answering others directly, in public threads or one to one. Launch sites brought real customers for some and a day of attention for others. Directories brought almost nothing, and ads taught little without a working message, though one builder still got half of 12 sign-ups from them.
Two caveats. Much of this comes from people selling to other builders, and a commenter doubted that it works unless your customers are founders. The founder agreed it suits them, because founders are so active online. If you sell to gardeners or dentists, the same idea means their groups, their events and their inboxes. And another commenter’s answer to the whole question is that nobody really knows, and that it looks like a lottery you can only win by staying in it. We think that’s closer to the truth than most guides admit.
What to do, in order
- Write one sentence: who it’s for, and what it does for them. If you can’t, stop here. More reach won’t fix that.
- Find where those people already complain. Search Reddit, Hacker News, Facebook groups and forums for the complaint in their words, not your product’s category. “My invoices eat every Sunday” finds people. “Invoice automation” finds vendors.
- Talk to ten of them. Ask what they do about it today, what it costs them and what they’ve already tried. A click saying they’d use it costs nothing and means little. What they already spend, in money or hours, tells you much more.
- Answer questions for a couple of weeks before you mention what you made. Read each community’s rules first, and say you built it when you do. We list the rules for the main places.
- Make something people can see without talking to you: a free version, or a 30-second video of it working. A plain screen recording is a fine first version.
- Get the first ten on by hand. Message them when it ships, watch one of them use it, and fix what stops them.
- Count people who came back or paid, not visits. One builder found about 3,700 direct visits but only 140 that did anything, and puts most of the rest down to bots. Check your own numbers for the same thing.
- Give it four weeks on one group of people before you change anything big. That’s our judgment, not a rule.
Should they pay from day one?
Builders disagree. The two-person team above stayed free until about 150 users, because they wanted feedback more than money. Others say only a card number tells you anything. We’d charge sooner than feels comfortable if your buyers are businesses, since it’s the one answer nobody can give just to be polite. Either way, don’t count free users as proof. One builder found that of about 20 people in a friends-of-friends Discord, one matched the customer they had in mind.
What to skip at first
- Ads, until one message already works by hand. They spend money faster than they teach you.
- Directories as a way to find customers. A launch can bring customers, but it’s one day, so do it after you’ve talked to people.
- Six channels. Six quiet channels teach you less than one busy one.
- Another round of features. One blunt piece of advice on Hacker News was to stop building, because building is the part you already know how to do.
- A marketer, before you know what you’d tell them to do.
A question most first-time builders never ask
Who already has your customers? One builder’s approach is to stop selling to restaurants one at a time and sell to whoever already sells to restaurants: the supplier, the bookkeeper, the software they already use. It doesn’t suit everything, and it’s harder if you sell to consumers. It can still save you months.
What it costs by hand
Very little money, and a lot of time. A phone or a free screen recorder will do, and Buffer has a free plan for up to 3 channels (read 3 Nov 2026). The hard part is keeping going every week while you’re also building.
Where Lou fits
The weekly video part is what Lou by Luven AI is being built to take on: paste a link, approve the videos, and they post. It won’t have the conversations above for you, and those matter more. Lou isn’t open yet and hasn’t been proved on real products, so do this by hand today. We’re in the same spot as you: Lou has no users yet either. We’ll write up what worked for us and what flopped.
How we write these: with AI help, from the sources we link, with dates. Counts marked as ours were made by an AI reader with no second check. If something’s wrong, email everest@luven.ai and we’ll fix it and say so.
