Hello, fellow founders and to-be founders! Captain Super here — part mascot, part mentor, ready to help you name the villains that keep tripping you up. Today's villain is one of the sneakiest: The Feedback Mirage.

Before Dropbox was Dropbox, it was a three-minute video.

No product. No signup flow. No code. Just Drew Houston explaining, in plain language, what the thing would do if it existed. He posted it. Within 24 hours, over 70,000 people joined the waitlist.

That’s validation. Not a nice conversation with five people who said “yeah, I’d probably use that.” Not a Slack channel full of “this is such a good idea!!” reactions from your friends. A number. A behavior. Evidence that strangers wanted something enough to raise their hand for it — before it existed.

Here’s the uncomfortable stat behind why this matters: most founders who build products nobody wants aren’t lazy, unoriginal, or bad at their jobs. They just let excitement substitute for evidence. They let the dopamine hit of building override the discomfort of finding out, early, that the idea might be wrong.

This issue is your pre-flight checklist — how to know if an idea is worth building, before you spend the weeks (or months) building it.

Naming the Villain (Evaluation vs. Validation)

Here’s the trick the Feedback Mirage plays on you: it dresses up evaluation and calls it validation. They sound like the same thing. They are not.

Evaluation is opinions. It’s interviews, surveys, “what do you think of this idea?” conversations. It’s useful — it helps you understand a problem, sharpen your pitch, spot obvious dealbreakers. But it never confirms that anyone actually wants what you’re building. People are generous with opinions because opinions cost them nothing.

Validation is behavior. It’s someone giving you money before the product exists. It’s someone switching away from their current workaround to use your half-built thing. It’s a pre-order, a paid pilot, a waitlist signup that required an email address, not just a thumbs-up emoji.

The test is simple: did it cost them anything? Time, money, effort, switching pain — if the answer is no, you collected an opinion, not a signal.

This is the exact trap Dropbox’s founder avoided. He didn’t ask 50 people “would you use cloud file syncing?” He put a 3-minute video in front of strangers and asked them to raise their hand — for real, with an email address — for a product that didn’t exist yet. 70,000 people did. That’s not a nice compliment. That’s proof.

Old Playbook vs. New Playbook

Here’s where it gets a little confusing, because the internet will give you two opposite answers depending on which corner you land in.

The old playbook says: talk to 15-20 people before you write a line of code. Interview them. Understand their workflow. Test willingness to pay. Only then, build.

The new playbook says: skip most of that. Building is so cheap and fast now that shipping a real product is the validation. Have a handful of conversations, then build and put something in front of real users within two weeks.

Both camps are right — just not for the same situation. Here’s the rule of thumb:

• Build-first works when building is genuinely cheap for your idea. A simple app screen, a no-code workflow, a landing page with a working demo behind it — if you can get something real in front of users in days, do that instead of interviewing for weeks. You’ll learn more from watching 50 strangers use a rough version than from 20 hypothetical conversations.

• Interview-first still wins when you don’t know the buyer or the problem space well. If you’re entering an unfamiliar industry, a regulated market, or a niche with its own language and politics, a few real conversations up front will save you from building the wrong thing entirely — because you don’t yet know enough to build even a rough version that tests the right assumption.

The villain doesn’t care which playbook you use. It just wants you to stop at “people seem interested” and skip the part where you actually test that.

The Pre-Flight Checklist

Alright, cadets. Before you’re cleared for takeoff, run through this. Don’t skip a step just because the earlier ones felt good.

1. Write the problem in one sentence - no solution allowed.

Try this test: describe the problem to someone without using the word “app,” “platform,” “tool,” or “dashboard.” If you can’t do it, you’re solution-first, not problem-first — and solution-first ideas are the ones that get built beautifully and used by no one. Example of doing it wrong: “Freelancers need a tool to track invoices.” Example of doing it right: “Freelancers lose an average of $6K a year chasing late payments and don’t find out a client is a slow payer until after the work is done.” Notice the second one gives you something to go validate — the first one just gives you something to go build.

2. Look for evidence people are already paying for a workaround, not just complaining about the problem.

Complaints are cheap — go find them anyway, but weight them. One recent framework scored problems by mention volume and severity: a problem with 80 mentions rated as highly painful can be a stronger signal than one with 200 mentions rated as mildly annoying. Where to actually look: Reddit threads in your niche, G2/Capterra reviews of adjacent tools (read the 2-3 star reviews — that’s where the workarounds live), App Store reviews, and job postings (if companies are hiring someone to do the thing manually, that’s a very expensive workaround). If you can’t find at least a handful of people already spending time or money to patch this problem, it might be a “nice to have” you’re the only one who feels strongly about.

3. Set your finish line before you start testing in writing.

This one matters more than it sounds like. Once you’ve run a test and seen a number, your brain will retroactively decide that number was good. Protect against this by writing your threshold down beforehand and sharing it with a co-founder or friend who’ll hold you to it. Concrete examples of thresholds: “50 waitlist signups where the signup form requires a real answer to ‘what would you pay for this’ — not just an email.” Or: “3 businesses willing to do a 30-minute paid pilot at $200.” Vague thresholds (”good response”) let the villain win by default.

4. Ask for something that costs them something pick a real test, not a poll.

A few concrete formats, roughly in order of how much signal they give you:

• Landing page test: a page with your value prop and a waitlist or pre-order button. Track conversion rate, not just signups — a 2% conversion from cold traffic is a very different signal than 20%.

• Fake-door test: put a button or feature in front of existing users (if you have any adjacent product) and see who clicks it, even though it doesn’t do anything yet.

• Concierge MVP: manually deliver the outcome yourself, by hand, for 3-5 customers, before writing any product code. If they’ll pay you to do it manually, that’s one of the strongest signals available — and it also teaches you what to actually automate first.

• Pre-sale: ask for money before the thing exists. Even a small deposit. Nothing filters politeness out of a “yes” faster than asking for a credit card.

5. Cap how long you’ll spend before showing someone and treat a blown deadline as information.

If your first testable version is creeping past 6-8 weeks, that’s not a scheduling problem, it’s a diagnostic: it usually means one of three things — the scope is too big, you’re relying on tech that’s more complex than the problem needs, or you’re quietly gold-plating a v1 because showing someone something rough feels scary. The fix in each case is the same: cut scope until you’re testing one assumption, not five. A rough version that tests the core value prop beats a polished version that took so long to build you’re now emotionally unable to kill it.

6. Watch what people do after the initial reaction, not the initial reaction itself.

“This is amazing!” is evaluation — it’s free and it’s nice and it tells you almost nothing. Look instead for: did they come back and use it a second time without being reminded? Did they tell someone else about it without being asked? Did they ask when it’s available, or ask for a specific feature (both are signs they’ve mentally already started using it)? Did they get annoyed when you took it away or paused the pilot? Irritation at losing access is one of the most underrated validation signals there is — nobody gets mad about losing something they didn’t actually want.

Close-

So there it is, the Feedback Mirage, named and grounded. It’s not a villain that jumps out at you. It’s quiet. It shows up disguised as kindness, as encouragement, as a friend saying “I’d totally use that.” And it will happily let you spend six months building something beautiful that nobody asked for, because being nice was easier than being honest.

You don’t beat it with more confidence. You beat it by asking for something that costs people something: money, time, a workflow switch, a real yes, before you write a single line of code.

Here’s your assignment before the next issue lands: pick one idea you’re sitting on right now, and run it through step 1 of this checklist. Write the problem in one sentence, no solution allowed. If you get stuck, that’s not a failure, that’s the checklist working exactly as it should.

Next time, we’re putting this into practice: a real founder, a real idea, and the validation test that either saved them months or should have.

Until then, stay grounded, cadets.

— Captain Super 🦉

We sat down with Zeeshan for the very first episode of the Behind the Launch Podcast, and the conversation ended up being less “here’s how to rank in ChatGPT” and more “here’s why most brands are optimizing for the wrong audience.” Here’s what stuck with us.

From editing freelancers to editing the internet

Serplock didn’t start as an AEO tool. It started as a content-rating engine — Zeeshan’s answer to a problem every content marketer knows too well: freelance writers turn in drafts, and someone has to sit there and turn “gibberish” into something publishable. That editorial bottleneck was, in his words, the real deal-breaker. Skip it, and all the time and money spent on the writer goes to waste.

When AI entered the picture, Zeeshan didn’t reach for it to write content — he was skeptical of that from the start. Instead, he used it to review content: feed in a URL, get back what’s wrong with the intro, where the internal links should go, what the piece is missing. A rudimentary tool, by his own admission. But it mattered, because as AI started writing more of the internet’s first drafts, the job market shifted underneath marketing teams. As Zeeshan put it, one of the first conversations he had in a fractional marketing role wasn’t about hiring content writers — it was about hiring editors, because AI was already doing the writing.

That’s the seed AEO grew out of: not “how do we get cited by an LLM,” but “how do we make sure whatever gets published is actually good enough to survive scrutiny — human or machine.”

The knowledge graph is the whole game

Here’s the part worth screenshotting. Zeeshan’s core thesis is that SEO and AEO aren’t competing disciplines — one is the philosophical foundation the other stands on. What’s actually changed is where you point your content.

Old SEO logic: chase search volume. Write the “best video editing software” listicle because people search for it, even if it has nothing to do with your actual product.

Serplock’s approach: build a knowledge graph — a structured universe of everything your brand needs to teach both humans and AI agents about itself. Features, use cases, capabilities, customer impact. Every piece of content earns its place not by keyword volume, but by relevance to the brand at the center.

The reasoning is simple once you sit with it: LLMs form their understanding of you the same way a person forms an impression at a dinner party — through everything said about you, good and bad. A lost customer review, a bad license dispute, an off-brand blog post about a product category you don’t actually compete in — all of it gets folded into how an AI describes you when someone asks. Zeeshan’s example: an email marketing platform that’s written a lot of content about cold emailing might get lumped in as a cold-emailing tool, even if that’s not the positioning it wants. The fix isn’t more content. It’s more disciplined content — everything orbiting the brand, not chasing traffic.

SEO fundamentals didn’t disappear — they just got a promotion

Zeeshan is refreshingly uninterested in the “is AEO real or just SEO with a new name” debate. His take: it doesn’t matter, because the fundamentals SEO built — readability, structure, crawlability, writing so bots and humans can both parse you — are exactly what AEO needs too. Bottom-line-up-front writing, in particular, gets a shoutout: it’s what AI Overviews already do by default, and readers have been trained by them to expect it everywhere now.

The uncomfortable question underneath all of this, which Zeeshan raises almost as a throwaway line, is the one worth sitting with: in two or three years, how many humans will actually be visiting websites at all? There’s a related paradox he points to — a growing share of top-ranking Google content is now AI-generated, which means AI is increasingly training on its own output. Net new information isn’t necessarily entering the system. Marketers optimizing purely for algorithmic citation, without adding anything genuinely new, might be optimizing for a feedback loop that eventually has nothing left to feed on.

What actually worked for early distribution

Away from the philosophy, a few tactical notes from Serplock’s early growth that are worth stealing:

• Listicles still convert distribution fast. Slower-burn SEO content builds authority over time, but listicles were Zeeshan’s fastest early lever for getting the first wave of visitors in the door.

• Comparison pages punch above their weight, especially when you can credibly stack your product against a recognizable competitor and highlight a genuine edge.

• Product-led content, sourced from real demo calls, outperformed generic long-tail SEO plays back at Mailmodo — because it answered the actual questions prospects were asking, not the questions a keyword tool guessed they might ask.

The takeaway for founders building in public

Positioning was always the job. AEO just made it impossible to fake. If your content orbits your brand instead of chasing traffic, if your editorial process is tight enough to survive both human readers and LLM summarization, and if you’re adding something to the internet that wasn’t already there — you’re doing AEO whether you call it that or not.

That gap between what got built and why is the entire reason Prodigent exists.

The problem: the “why” disappears

As an engineering leader, Tim kept hitting the same wall. A product or feature would move through its lifecycle, and the reasoning behind it would move with it, scattered across Slack threads, unrecorded meetings, and docs nobody opens again.

“One of the biggest challenges that I always faced was the changing nature of a product or feature through its entire lifecycle. I wanted some way to capture all of that knowledge throughout every stage, but also without too much noise.”

That last part matters. Tim wasn’t looking to build another place to dump documentation. He wanted the reasoning captured without adding a task nobody wants to do. That distinction became the design brief for Prodigent.

The real fight wasn’t the pitch. It was the switching cost.

Nobody had to convince Tim’s prospects that losing institutional knowledge was a problem. Every product team has felt it. The resistance showed up somewhere else entirely.

“Every product team has an established way of working, both in terms of processes and tools used. Trying to get people to consider other tools or other processes is an uphill battle.”

This is the part of building in product management tooling that doesn’t make it into most launch posts. Solving a real problem isn’t enough. Teams won’t tear up a workflow they’ve already built their week around, no matter how good the reason is.

How he got around it: stop asking teams to switch

Tim’s response was to stop competing for a spot in the toolchain and instead work underneath it.

“Focusing on making the product adaptable to any workflow, and integrate with any other tool, so that it could slot into existing processes and toolchains.”

Instead of becoming a new destination, Prodigent sits alongside the tools a team already uses and captures the reasoning that normally falls through the gaps between them. Adoption doesn’t require a migration. It requires nothing to change except that the “why” finally gets recorded somewhere.

Where this goes next

Tim’s immediate goal is narrow on purpose: get more product teams using Prodigent, and let real usage shape what it becomes next.

The longer view is bigger. As generative AI takes on more of the actual writing of software, Tim expects tools that define requirements and translate context between people and AI to matter more, not less. He sees Prodigent eventually moving past planning, specifying a product or feature in enough detail to help produce the code itself, or sitting inside an engineer’s IDE and directing the work based on the requirements and reasoning it already has.

The takeaway

Prodigent isn’t solving a knowledge problem with more documentation. It’s solving it by making the capture invisible and the fit frictionless. If you’re building in a crowded workflow category, the lesson isn’t “make something better.” It’s “make something that doesn’t ask anyone to change what they’re already doing.”

Today’s villain: The Leaky Bucket.

If you’re a SaaS founder, you will meet it. You’ll close a great month — new signups, new logos, MRR chart pointing up and to the right. Feels like momentum.

Then you look closer and realize half of what you added, you also lost. You weren’t growing. You were refilling a bucket with a hole in the bottom, as fast as you could pour.

This isn’t a “you’re bad at sales” problem — that one’s simple to fix. This villain is sneakier: it hides inside your best month, disguised as good news.

In this edition, Cap breaks down the math nobody wants to look at, the two very different holes in the bucket, a handful of real teams who caught it, and what they actually changed.

The Math Nobody Wants to Do

Here’s the number that should worry every SaaS founder more than it does: the average SaaS company sits at somewhere between 10-14% annual churn, and the widely used bar for “good” is under 5%.

Yet the majority of SaaS companies never hit that bar. Most founders don’t know which side of that line they’re on, because they’re not looking.

The math compounds in a way that punishes you quietly. A company starting with 1,000 customers at a steady 5% monthly churn rate loses close to half its base by the end of the year — even while acquisition keeps humming along in the background, making the dashboard look fine.

And it’s not just customer count. Losing one customer paying $5,000 a month hurts far more than losing ten customers paying $100 a month, even though “customer churn” would report ten losses in one case and one in the other.

If you’re only watching logo count, you can miss the fact that your highest-value accounts are the ones quietly walking.

Meanwhile customer acquisition itself has gotten more expensive — costs are up significantly over the last several years — which means the leaky bucket isn’t just a growth problem anymore, it’s a burn-rate problem. You’re paying more to fill the bucket at the exact moment it’s leaking faster.

Two Very Different Holes

Founders often talk about “churn” like it’s one thing. It isn’t. There are two distinct holes in the bucket, and they need completely different patches.

Voluntary churn is a customer actively deciding to leave — they cancel, they don’t renew, they tell you (or don’t) why. This is the one founders think about: wrong product fit, bad onboarding, a competitor, a budget cut.

Involuntary churn is quieter and dumber: a card expires, a payment fails, and the customer didn’t mean to leave at all — they just got dropped. This category alone can account for a large chunk of total churn — estimates run from roughly a fifth to nearly half, depending on the business.

And it’s often the cheapest hole to patch: automated payment retries, expiry reminders, and updated-card prompts recover a meaningful share of these customers without a single retention conversation.

If you’re only building retention plays for voluntary churn — onboarding improvements, customer success calls, feature requests — you might be ignoring the easiest win in the building: just catching failed payments before they turn into cancellations.

When you hit it, existing customers alone grow your revenue even with zero new signups. It’s the single biggest lever mature SaaS companies pull, and it starts with treating retention as a growth channel, not a defensive one.

Reframe the Problem

The founders who get this right stop asking “how do we get more signups” and start asking “why did the ones we already had leave.” That second question is less exciting, but it’s where the real leverage is.

It costs far more to replace a customer than to keep one, and every percentage point of churn you shave off compounds — it’s not a one-time fix, it’s a permanently higher floor under your growth. Some estimates suggest a modest cut in churn can roughly double your effective growth rate, simply because you stop losing ground you’d otherwise have to re-win.

The Playbook: How Real Teams Patched the Bucket

Find your Red Flag behaviors, not just your Red Flag customers

Groove was stuck at a 4.5% monthly churn rate and didn’t know why. Instead of guessing, they dug into behavioral data and identified specific “red flag” patterns — users with unusually short sessions, users whose login activity had gone quiet, and users showing other signs of struggle.

Each pattern got its own targeted outreach: a setup-help email for short-session users, a check-in for disengaged ones, a support nudge for struggling accounts. The result was a churn reduction of roughly 71% — not from one big fix, but from catching the right people at the right moment, before they’d already decided to leave.

Share

Fix who you sell to, not just how you retain them

SmartReach was losing customers at a rate of 27% a year. Instead of only optimizing onboarding or support, they went further upstream — refining exactly who they sold to in the first place, and declining to close deals with poor-fit prospects even when it meant saying no to short-term revenue.

Segment before you intervene

Mention, a social listening tool, grew from a few hundred customers to over 200,000 and discovered a blanket retention strategy doesn’t scale. They separated users by plan type, recognized that paid and trial users needed different treatment than free users.

Watch usage, not just the calendar

StatusBrew found that waiting for a cancellation event to react was already too late. By monitoring usage patterns and proactively reaching out to accounts showing early signs of disengagement, rather than waiting for a renewal date to force the conversation, they cut churn by roughly 20%. The shift was from reactive (”why did you leave”) to proactive (”we noticed you’ve gone quiet — everything okay?”).

Fix the boring stuff: failed payments

It’s not glamorous, but for a lot of SaaS businesses, the highest-ROI churn fix has nothing to do with product or support. Automated dunning — retrying failed payments, prompting for updated card details before an expiry, sending a reminder ahead of renewal — recovers a real slice of “churned” customers who never actually decided to leave.

Shorten the path to the first “win”

A recurring pattern across churn research: users who hit a meaningful milestone in their very first session stick around at dramatically higher rates than users who don’t — some data points to a gap as large as 60-75% retention for activated users versus 10-20% for those who never got there.

What Not To Do

Don’t celebrate new MRR without netting it against churned MRR. A big new-customer month means nothing on its own. Put both numbers on the same chart, every time, or you’re just watching half the movie.

Don’t treat all churn as one problem. A customer who churned because a credit card expired needs a completely different fix than one who churned because your product didn’t solve their problem. Lumping them together means you’ll build the wrong solution for both.

Don’t guess at why people leave. “They probably wanted more features” or “it was probably the price” are stories founders tell themselves instead of doing the unglamorous work of actually looking at usage data and asking churned customers directly what happened.

Don’t treat retention as someone else’s job. It’s tempting to hand churn off to a support team once you have one. But in the earliest stage, the founder is usually the only person who can see the whole picture — acquisition, product, and retention — clearly enough to catch the pattern.

Close

The Leaky Bucket doesn’t announce itself. It hides behind a good month, a rising chart, a Slack message that says “we hit a new MRR record.” It’s patient — it’ll let you celebrate for a while before the math catches up.

The founders who beat it didn’t do it with one clever trick. They did it by refusing to look away from the number that felt worse to look at — the one going out the back door, not the one coming in the front.

They separated the customer who forgot to update a card from the customer who genuinely walked away. And they started treating their existing customers as a growth channel worth investing in, not just a base to defend.

Growth that lasts isn’t the fastest pour. It’s the tightest bucket.

Captain Super, signing off. Now go check your churned MRR, not just your new MRR. 🦉

That’s the story of two brothers, David and Felix Heikka.

In the summer of 2024, the two brothers were sitting in an apartment when they built two products that nobody wanted.

But this story isn’t about their failure. It’s about how that failure helped them build a company called Buildpad, which is now used by over 100,000 founders.

Before we jump into the success story, let’s talk about the failure first because that’s where it all began.

Two brothers, two very different lives

David and Felix Heikka were born in the north of Sweden, two years apart. After that, their paths split for the better part of a decade.

David finished high school and became a professional poker player, moving to Asia to play. Five years later, he came back to Europe and co-founded a SaaS business that did well.

Felix took a different route entirely. He joined the army, moved further north in Sweden, and stayed for five and a half years, working his way up to NCO — non-commissioned officer.

While Felix was still in uniform, David was in Budapest, learning to code and building a side project. On their weekly calls, the two brothers kept coming back to the same idea: what if they built something together?

In March 2024, the timing lined up. Felix left the army and booked a one-way ticket to Budapest.

He had no startup experience. No technical background. What he had was a brother with a plan, and a willingness to bet on it.

Neither of them knew yet how hard the next seven months would be.

The first product: AIConvert

The side project David had been building in Budapest was AIConvert — an AI-powered lead-generation form, built around a problem he’d run into in his own SaaS business.

When Felix arrived, the plan came together fast. David would keep building. Felix would learn marketing from scratch and take that side of the business.

They were confident. More than confident.

Every business that ran sales calls was going to want this, they thought.

So they built. And then they marketed — hard. Paid ads. SEO. Cold DMs to prospective customers. Every channel they could learn how to run, they tried.

The signups didn’t come.

When one did trickle in, they celebrated it. Not because it was a milestone worth marking on a roadmap, but because it was rare enough to feel like a win. They set ambitious targets and missed almost all of them.

Felix, meanwhile, was living off savings. There was no salary. No funding. Just runway, quietly shrinking.

This was supposed to be the easy part — they’d built something, they understood the problem, they were putting in the hours. It was still going nowhere.

Five months of silence

By June 2024, they hadn’t given up on AIConvert, but they’d started looking sideways for something — anything — else.

That’s how TinderRoast happened. Someone had gone viral on X with an AI tool that roasted people based on their post history. David and Felix saw an opening: what if the same idea worked on dating profiles?

It didn’t. TinderRoast never took off.

And they kept hearing nothing back. No real results, no matter which lever they pulled.

This is the part of the story that’s easy to gloss over, because it doesn’t have a clean turning point yet. There’s no product to point to. No user calling in with praise. Just two brothers, one shrinking runway, and a growing sense that something about the way they were working — not just what they were building — had to change.

They weren’t lazy. They weren’t undisciplined. They were doing exactly what founders are told to do: build, launch, market, repeat.

It just wasn’t working.

The idea that came from frustration

The idea for what would become Buildpad didn’t arrive as a flash of inspiration. It arrived as an annoyance.

David was using AI tools to help with development, and he kept running into the same wall: every time he opened a new session, he had to re-explain the project from scratch. The AI had no memory of what they’d already discussed, what had already been tried, what the actual goals were.

Memory in LLMs, as a real feature, didn’t exist yet. But the idea stuck. David and Felix started talking it through, and the shape of an “AI cofounder” began to form.

This time, though, they didn’t go straight to building.

That decision — to pause before writing a single line of code — is the hinge the entire story turns on.

The biggest change they made

After two failed products and seven months of close-to-nothing, David and Felix had learned one thing clearly: they needed to solve a real problem, not one they’d assumed was real.

So instead of idea, build, launch, hope — they tried something closer to problem, talk to people, validate, build.

They found the subreddit where their target audience — early-stage founders — actually spent time. They posted a simple survey asking for feedback on the idea.

The response was positive.

For the first time in seven months, they had something they hadn’t had with AIConvert or TinderRoast: actual evidence, from actual people, that the problem was real before they’d built anything to solve it.

Only then did they start building the MVP.

100 users

They shipped the MVP in September 2024. Within two weeks, they had 100 users.

On paper, a hundred users doesn’t sound like much. It isn’t, for most startups.

For David and Felix, it was enormous.

For the first time, people were actually interested. Not because of an ad. Not because of a cold DM. Because they’d asked people what they needed before they built it, and then built that.

45 days of listening

They didn’t rush from there straight to a big launch.

The MVP had brought in real users, and real users meant real feedback — the kind you can’t get from a survey. So David and Felix spent the next stretch, roughly 45 days, doing something less exciting than shipping features: listening, and adjusting based on what they heard.

They talked to users. They took in criticism. They rebuilt parts of the product around what people actually said, not what they assumed people wanted.

It’s a section of the story with no dramatic beat to it. Nobody goes viral. No milestone gets hit. It’s just two brothers doing the unglamorous work of making the product better before they asked for more attention.

That patience is also why what came next didn’t feel like luck.

Product Hunt

In November 2024, seven months after Felix had landed in Budapest, they launched on Product Hunt.

It went better than they’d imagined. They landed at #4 Product of the Day, with more than 500 upvotes. In the first 24 hours, 475 new users signed up. Within a week, they’d crossed 1,000 users.

And then the number that mattered most: after seven months of building products that made almost nothing, they got their first paying customers.

Not signups. Not interest. Money.

For two brothers who’d spent the better part of a year celebrating single signups, watching people choose to pay was the real payoff — bigger, in its way, than the Product Hunt ranking itself.

The unglamorous part

The launch ended. The traffic spike faded, the way launch spikes always do.

What came next was less a celebration and more a return to work — the kind that doesn’t show up in a highlight reel.

David and Felix kept talking to users. They shipped updates. They posted publicly about what was working and what wasn’t, sharing both the wins and the failures rather than only the polished version.

The big launch, it turned out, wasn’t the finish line. It was closer to the starting gun for the actual company.

20,000 → 100,000

A year after that Product Hunt launch, Buildpad hit 20,000 users.

The brothers later described the shift in blunt terms: they’d gone from celebrating every single signup to watching hundreds join in a single day.

But the story doesn’t end on that number. Buildpad’s next chapter came with another failure baked into it.

The rebrand that didn’t work

In October 2025, with Buildpad at around 20,000 users, David and Felix made a big call: they rebranded the entire product as aicofounder.com.

The logic made sense on paper. Users had been describing Buildpad the same way for over a year — it felt like having an AI cofounder. The brothers figured: why not own that category outright? Become the AI cofounder. Pick up the search traffic that comes with owning a term people were already searching.

Nine months later, they reversed it.

Not because the product had failed — quite the opposite. Buildpad grew from around 20,000 users to more than 100,000 founders during that stretch, rebrand and all.

The problem was smaller and stranger than a growth number: nobody, including the brothers themselves, could agree on how to write the name. AICofounder? AI co-founder? aicofounder.com? Their own team used different versions of it. Their users typed it a dozen different ways.

David later put it plainly: when your own team isn’t sure how to write your company’s name, something is off.

A category name looks appealing from the outside — it explains itself instantly. But very few companies ever manage to make a generic term function as a real brand. For everyone else, chasing the category means spending your energy defending a name you can never fully own, instead of building equity in one that’s actually yours.

The lesson wasn’t really about naming conventions. It was a smaller version of the same lesson from two years earlier: the obvious move isn’t always the right one, and sometimes you don’t find that out until you’ve already spent nine months proving it.

The part that actually matters

The interesting part of Buildpad’s story isn’t that David and Felix eventually reached 100,000 founders.

What actually changed things wasn’t a better ad, a smarter growth hack, or more hours in the day.

It was a different question.

Instead of asking, “How do we convince people to want this?” — they started asking, “Do people actually have this problem?”

They asked that question on a subreddit, for free, before writing a single line of code for what became Buildpad. The answer changed everything that came after it.

If you’re building something right now and it feels like AIConvert felt — a lot of effort, almost no signal — that’s not necessarily a sign you’re doing it wrong. It might just be a sign you haven’t asked the question yet.

Hello, fellow founders and to-be founders! Captain Super here — part mascot, part mentor, ready to help you name the villains that keep tripping you up. Today’s villain is one of the sneakiest: The Trust Loop.

If you’re a first-time founder, you will almost certainly run into it. You’ve got a product, maybe even ready to launch. But the moment you do, people hesitate to trust it, because you’ve got no users yet. No users means no proof. No proof means no customers. And no customers means... no proof.

You post on LinkedIn. You DM your network. You get a few “looks interesting, keep me posted” replies — then silence. Nobody wants to go first, and you’re stuck holding a product that works with no way to prove it to the people who haven’t tried it yet.

This isn’t just “sales is hard” — that has a fix. This villain is trickier: the thing you need to beat it is the thing it’s blocking you from getting.

In this article, Cap breaks down why this villain’s gotten stronger, the secret to outsmarting it, and a 5-move combo to land your first believers.

Why This Is Worse Than It Used to Be

This isn’t a new problem, but it’s gotten harder to solve. A few years ago, “move fast, be scrappy, figure it out” was enough to get a first wave of believers. That doesn’t work the same way anymore.

Buyers are more skeptical by default. Everyone’s pitching an AI-powered something, every landing page promises to save you hours, and most of it doesn’t hold up. So people have learned to filter hard, and “trust me” isn’t a strategy anymore — it’s a red flag.

Social proof has also quietly become table stakes instead of a nice-to-have. People expect to see logos, testimonials, or case studies before they’ll even take a call, not after. The bar moved earlier in the funnel, and most first-time founders are still building for the old bar.

And zero-to-one is lonelier now, because “just be early and scrappy” doesn’t land the way it used to. Being early used to be a badge of honor. Now it just reads as risky, until you give people a reason to believe otherwise.

Reframe the Problem

Here’s where most founders get stuck: they treat “proof it works” as one big monolithic thing. Either you have it — the case studies, the logos, the glowing testimonials — or you don’t, and until you do, you’re stuck waiting on the sidelines.

That’s the wrong mental model. Proof isn’t a wall you’re either on the right side of or not. It’s a ladder. Different people need different rungs of it, at different stages, and you don’t need to reach the top rung to start climbing.

Your very first user doesn’t need a case study to trust you — they need to trust you, specifically, as a person who’ll actually show up if something breaks. Your tenth user needs something a little more concrete: a specific result, even a small one. Your hundredth user is the one who finally wants the polished case study and the recognizable logos.

So the real unlock at the earliest stage isn’t manufacturing traction you don’t have yet. It’s selling trust instead of traction — giving people a reason to believe in you before you have anything else to show them. Traction comes later. Trust is the thing you can actually build on day one.

The Playbook: How Founders Actually Break the Cycle

Knowing you need trust instead of traction is one thing. Here’s what it actually looks like in practice.

Borrow credibility you haven’t earned yet

You don't have customers, but you might have advisors, mentors, or early feedback from people with names your audience recognizes. "Built with input from [respected person/company]" does real work here — it tells people someone whose judgment they already trust has looked at this and didn't walk away.

Do the work by hand before you productize it

Before you had software, you probably had a spreadsheet, a Slack channel, and a lot of manual effort — and that's fine, that's how this is supposed to start.

Airbnb's founders famously went door-to-door in New York, personally photographing hosts' apartments because the listings were performing badly and they needed proof that better photos would change behavior.

Trade broad claims for radical specificity

"Increases productivity" convinces no one anymore, because everyone says it — it's the "made with love" of B2B marketing. "Saved this specific team 6 hours a week on reconciling invoices" convinces people, because it's checkable, narrow, and obviously not fabricated. The more specific the claim, the harder it is to dismiss as marketing copy.

Go deep on 1-3 customers instead of chasing 20 logos

A single case study done properly — with a name, a role, a real number, a direct quote — will outperform a wall of unnamed "beta users love it" every time.

Make the first customer’s risk close to zero

Pilots, money-back guarantees, "pay only if this works" — these aren't just sales tactics, they're proof-substitutes. Superhuman's early growth leaned heavily on this: instead of a self-serve signup, they did onboarding calls with almost every early user, personally walking them through the product.

It didn't scale, but it made the first users' risk (of wasting time on a confusing new tool) close to zero, which mattered more at that stage than having a slick self-serve flow.

What not to do

It’s tempting, when you’re stuck in this loop, to shortcut your way out. Don’t. A few things that feel like solutions but actually make the trap worse:

Don’t fabricate logos, testimonials, or “beta users”

Vague claims like “our beta users love it” with no names, no numbers, and no way to verify anything are a fast way to lose the trust you’re trying to build. People can smell manufactured proof, and once they catch one exaggerated claim, they’ll assume everything else on the page is exaggerated too.

Don’t wait for “enough” proof before you launch

There’s a version of this trap that looks like diligence but is really just perfectionism — always one more case study away from being ready, always waiting for a bit more traction before you’ll show up publicly. At some point the only way to get proof is to launch without enough of it and build it in public, imperfect as that looks.

Don’t confuse borrowed credibility with earned credibility

Using an advisor’s name or a mentor’s endorsement is a legitimate first rung on the ladder, but it’s not a substitute for eventually delivering your own specific, verifiable results. If you lean on borrowed trust forever instead of converting it into real proof, you’re just delaying the same problem.

Here’s the close:

Close

The Trust Loop isn’t a villain you defeat once and never see again, fellow founder. It’s one you outlast. There’s no single moment where you’ll wake up with “enough” proof and watch the loop shatter for good. Instead, you find the seam: a borrowed cape, a secret-identity grind, X-ray vision on vague claims, depth over a cape-full of logos, a force field for the first customer’s risk. Small moves that let a few people trust you before the evidence exists.

And somewhere in doing that, the frame flips. Your early customers stop looking like the reward you get after you’ve found proof. They start looking like what they really are: the proof itself. Not because they prove your product is flawless — most of them will find bugs, ask for things you haven’t built yet, churn before you’ve cracked retention — but because they prove something bigger. That someone was willing to go first.

That’s the whole mission at this stage. Not eliminating the Trust Loop. Just finding the two or three people willing to step into it with you.

Captain Super, signing off. Now go find your first believer. 🦉

The second someone lands on your page, three questions are already running in their head. Is this real, or another demo that breaks on the third prompt? Is this actually useful, or just ChatGPT with a UI slapped on top? And is it worth five minutes of my time to find out?

Most AI landing pages ignore this. They explain features. They list benefits. They act like the visitor already believes the product works, and just needs to be convinced to buy.

That’s backwards. Your landing page’s first job isn’t to sell. It’s to prove.

Before someone reads a single feature, they need to see the thing actually work. A real output. A real result. Something that isn’t a claim.

Everything else on the page, the copy, the CTA, the tools you use to build it, only matters once that trust exists.

The anatomy of a converting landing page for an AI product

Strip away the templates and every high-converting AI landing page has the same five parts, in the same order.

1. Proof above the fold

Not a headline. Not a hero image. Proof. This is the single biggest difference between AI landing pages and regular SaaS ones. With normal software, you can describe what it does and people believe you. With AI, they don’t believe you until they see it.

That means: a real generated output sitting in the hero section, not a screenshot of a dashboard. A live embedded demo they can try without signing up. A before-and-after that took the product 10 seconds to create. If your hero section is just a headline and a “Get Started” button, you’ve made the visitor do the work of imagining whether this actually works. Don’t make them imagine it. Show them.

2. A headline that names the outcome, not the technology

“AI-powered content generation platform” is a description. It tells the visitor what category you’re in. It doesn’t tell them what changes in their life if they use it.

Compare that to something like “Turn a rough idea into a first draft in 40 seconds.” That’s not smarter copywriting, it’s a different question being answered. One answers “what is this.” The other answers “what do I get.” Visitors only care about the second one.

A useful test: read your headline out loud. If someone could nod and still not know why they’d use the product tomorrow, it’s still describing the technology.

3. The skepticism section

Every AI product visitor is silently running the same two questions: is this just a wrapper around GPT, and will this actually work on my specific case, not just the polished demo example.

Most landing pages never address this directly, hoping the visitor won’t ask. They always ask, they just ask on the way out the door instead of on your page. Put a section that answers it head-on. What’s proprietary about how the product works. What happens with edge cases. Why it’s not just a prompt with a UI wrapped around it. This section does more to reduce bounce than another testimonial will.

4. Specific proof, not vague trust signals

“Trusted by 500+ teams” is a number with no texture. It doesn’t tell the visitor anything they can picture using themselves.

“Cuts first-draft time from 2 hours to 40 seconds” is a number they can picture, because it’s specific enough to imagine in their own workflow. The difference between vague and specific proof is usually just one level of detail: not “saves time,” but how much time, doing what, compared to what.

Real output samples work the same way. A specific, slightly messy real example beats a polished, generic one, because it reads as true rather than as marketing.

5. A CTA with no risk attached

“Book a demo” asks for a meeting on someone’s calendar before they’ve even decided the product is real. That’s a big ask this early.

“Try it free, no card required” costs the visitor two minutes and nothing else. Early in the funnel, your only job is to get them to experience the product firsthand, because that’s what actually builds trust, not more copy. Save the higher-commitment CTA for later in the page, or for retargeting once they’ve already seen it work.

That’s the shape. Everything else, design, animation, color, is decoration layered on top of this structure. Get these five right and the page converts even if the design is rough. Get the design right and skip these five, and it still won’t.

Breakdown: 3 AI landing pages that get this right

1. Flow (wisprflow.ai) — a voice-to-text AI tool

What’s above the fold: no dense explainer. Just a light hero animation that visually shows the product turning voice into refined text, without needing a paragraph to explain it. The visitor sees the transformation happen before they read a single feature bullet.

What to steal: if your product does one clear “before to after” transformation, animate that transformation in the hero. Don’t describe it, show the input becoming the output.

2. Snowglobe (guardrailsai.com/snowglobe) — an AI simulation/testing tool

What’s above the fold: a before-and-after product comparison that communicates the value instantly, paired with a distinct visual system that repeats throughout the page so nothing feels like generic AI-template filler.

What to steal: the before/after format is one of the fastest ways to prove value without asking the visitor to read anything. If your product fixes or improves something measurable, show the two states side by side, not as a claim, as a visual.

3. Composio (composio.dev) — an AI agent tooling platform

What’s above the fold: rather than one static hero, the page transitions visually from wireframes to pixels to grainy gradients, and uses tabbed explanations to communicate features without overwhelming the page, which matters a lot for a technical AI product with several capabilities to explain.

What to steal: if your product does multiple things and you’re tempted to list all of them above the fold, use tabs or progressive reveal instead. Show one capability fully, then let the visitor choose to see more, instead of dumping everything at once.

The pattern across all three: none of them lead with a paragraph explaining the technology. They all lead with something visual that proves the transformation, and they let the visitor’s eyes do the trusting before their brain has to.

Share

How to build a converting landing page (with the tools to do it)

Step 1: Capture your proof first

Before you write a word of copy, get 10 seconds of your product actually working. A real generated output, cropped tight, no voiceover needed.

Tools: Loom (free tier) for a quick screen capture, or your OS’s built-in recorder if you want zero extra sign-ups.

Step 2: Write the headline around the outcome, not the tech

Feed your product’s actual output into an AI tool and ask for 3 headline options that say what changes for the user, not what category you’re in. Cut anything that could describe a competitor’s product just as well.

Tools: ChatGPT or Claude for the drafts, Hemingway App to strip out anything too dense or passive once you’ve picked one.

Step 3: Build the page structure

Use the five-part anatomy from earlier: proof above the fold, outcome-led headline, a skepticism section, one specific proof point, and a low-risk CTA.

Tools: Carrd if you want something live in an hour with zero design decisions. Framer if you want the page to feel a bit more polished and want built-in motion for that “before to after” moment. Google Stitch if you’d rather start from an AI-generated draft and hand-edit clean code from there.

Step 4: Add the skepticism section

Write two or three sentences that answer the question every AI product visitor is silently asking, is this just a wrapper, will this work on my actual use case. Don’t dodge it, answer it directly.

Step 5: Add specific proof, not vague trust signals

Swap “trusted by 500+ teams” for a number the visitor can picture in their own workflow. If you don’t have that data yet, use a real before-and-after instead.

Tools: Canva (free tier) if you want to turn that before-and-after into a simple visual instead of a paragraph.

Step 6: Add a low-risk CTA and a way to capture interest

“Try it free, no card” beats “book a demo” this early. If the product isn’t ready for self-serve yet, a waitlist works, but keep the ask small.

Tools: Tally for a free, no-friction waitlist or email capture form.

Step 7: Publish and measure

Get the page live on a free subdomain first. Don’t buy a custom domain until you know the page converts.

Tools: Carrd and Framer both host free out of the box. Add Google Analytics or Plausible (free trial) so you know if anyone’s actually converting before you touch the design again.

Step 8: Send it to 10 real people before you touch the design again

Not friends who’ll say it looks great. People who’d actually use the product. Watch where they hesitate, that’s your next edit.

Start convert with your landing page

Your landing page doesn’t need to be perfect. It needs to make someone believe your product might actually work.

That’s it. That’s the whole job.

Most founders spend their first weekend on colors, fonts, and animations, and their first month wondering why nobody’s converting. Flip the order. Get the proof right first. Get one honest sentence about what changes for the user. Everything else, the polish, the custom domain, the perfect gradient, can wait until you know the page actually works.

Build it this weekend. Send it to 10 people. Watch where they hesitate. Fix that. Ship again.

That’s the whole loop, and it’s the same one you’ll run for every page you ever build.

As a kid, he made kites. He built bicycles. He bred parrots and sold them. If there was a way to turn something into money, he found it. That instinct never left him. It just took a brutal first startup to point it in the right direction.

Today Rashid runs a small portfolio of bootstrapped SaaS products sitting at roughly $42K in monthly recurring revenue. He is not technical. He has no cofounder drama left to manage. And the whole thing exists because his first company nearly ended him financially.

Here is how that happened.

The company that took everything

Rashid’s first real swing was an SEC-registered robo-advisor. On paper, it sounded like the kind of idea investors love: fintech, automation, a real regulatory moat.

In practice, it was a machine for burning cash. Getting SEC-registered is expensive. Staying compliant is expensive. Running a robo-advisor without VC backing is expensive in ways that compound every month you stay open.

Share

Rashid didn’t have VC backing.

He drained his savings. He maxed out his credit cards. He went deep into debt trying to keep the thing alive. And eventually, it didn’t work. The company that was supposed to make him money took nearly all of it instead.

Somewhere in the wreckage of trying to raise money for that company, Rashid ran into a problem that had nothing to do with robo-advisors: he could not find the right investors. There was no clean way to search for the right angels or VCs, track outreach, or keep them updated. He was cobbling it together by hand, the way most founders do.

He filed that frustration away. He didn’t know it yet, but he’d just found his next company.

A second failure, for a completely different reason

Before he got there, Rashid took one more swing. In 2018, he teamed up with a couple of cofounders to build Investor Hunt, a database to help founders find investors.

This time the product wasn’t the problem. The team was.

It was the founder story nobody puts on a highlight reel: someone didn’t want to share the money, someone else didn’t want to do the work, and the whole thing quietly fell apart. The app shut down. Rashid knew the idea had legs. He’d seen enough interest to believe people wanted this. But belief in an idea doesn’t fix a broken cap table or a mismatched team.

So he let it go. Mostly.

He kept one thing: the investorhunt.co domain. He wasn’t done with the idea. He just needed to do it without the same people.

Building Angel Match out of his own scar tissue

Rashid didn’t have a technical cofounder this time, and he didn’t go looking for one. Instead, he found a remote developer and built Angel Match from scratch, a fundraising platform to help founders find the right investors, manage outreach, and send updates.

It was the same problem he’d hit while raising for the robo-advisor. Except now he was building the tool he’d wished existed back then.

The early version was rough. Lots of bugs. Rashid designed the site himself. His developer built it on Node.js, hosted on Digital Ocean, with Stripe running payments and Klaviyo running email. Nothing fancy. Just enough to ship.

He eventually went back and quietly restored Investor Hunt too, using that domain he’d held onto. It runs today as a second product in the portfolio.

The plateau nobody could explain

Angel Match found revenue fast, then got stuck. For a long stretch, it sat around $5K in monthly recurring revenue and would not budge, no matter what Rashid tried.

The cause turned out to be almost embarrassingly simple. Every time the team updated the app, they were deleting old pages on the website. Each deletion reset whatever search ranking those pages had built up. The site could never compound in organic traffic, because they kept quietly erasing its progress.

It took hiring an in-house SEO expert to catch it. Once they stopped deleting pages and started building for search on purpose, growth started compounding instead of resetting to zero every few weeks.

It is a small, unglamorous detail. It is also the entire difference between a product stuck at $5K MRR and one pushing past $40K.

Where the portfolio stands now

Angel Match is the anchor of the business, sitting around $37K in monthly recurring revenue after dipping slightly from a September peak of $41K. Rashid had paused Meta ads for a month to evaluate whether they were actually profitable and to change who was running them. Ads resume soon, with a goal of pushing the product to $100K MRR.

Investor Hunt, the product that once died from cofounder conflict, now runs quietly at around $4,500 MRR under Rashid alone.

Pur Social, a social media scheduling tool, sits around $490 MRR.

Journalist Hunt, a database of 100,000 journalists built to help founders land media coverage, is newer and smaller at around $180 MRR, but already pulling in close to 100 organic clicks a day.

What Rashid would tell you if you’re just starting out

Rashid is blunt about what he’d change if he had to start over. He would not chase a big, shiny idea with unicorn potential. The odds are stacked against that path, and it depends on things mostly outside your control: fundraising, compliance, how fast you can hire, how fast you can grow.

Instead, he’d start small. Solve one clear, boring, painful problem. Put a paywall behind it early and see if anyone actually pays. Build several of these small tools in parallel instead of betting everything on one. Whatever sticks, double down on distribution.

His filter for a good idea is simple: if people are already solving the problem with spreadsheets and manual workarounds, you’re most of the way to something worth building.

He is also honest about the part that doesn’t photograph well. The hardest thing wasn’t the strategy. It was not quitting for years, while pushing something uphill that barely seemed to move. His first cofounder used to put it this way: healthy, rational people would have quit a long time ago. They kept pushing anyway.

Most of the people who started alongside Rashid did eventually quit. Some are in corporate jobs now, and Rashid doesn’t think less of them for it. Different risk tolerances lead to different lives. But he doesn’t regret the years he spent refusing to be one of them.

That’s the whole playbook. Everything below is the data behind why it works and the exact mechanics of doing it.

Why “Just Post About It” Doesn’t Work at Zero

Cold outreach is failing at a rate that should change how every solo founder thinks about their first customers. B2B cold email reply rates dropped from 6.8% in 2023 to under 5% in 2024, and current benchmarks put B2B SaaS specifically at 2-4%. That means 95 to 98 out of every 100 cold messages you send get ignored outright, and the trend is getting worse, not better, as inboxes fill with AI-generated outreach.

Now look at the other side of the ledger. Warm introductions convert at 40-60% meeting rates. Referred leads close 3-5x faster than cold prospects and carry meaningfully higher retention once they’re in. One dataset put it starkly: a warm intro converts roughly 10-20x better than a cold list.

That gap is the entire story of your first 10 customers. Most “how to get customers” advice pictures cold outreach at scale, more volume, more templates, more automation, hoping the numbers eventually work in your favor. The data says that’s the losing strategy at zero audience specifically. You have no reputation for a cold message to borrow from, so it lands exactly where all cold messages land now: ignored.

The Concierge MVP - What “Doing Things That Don’t Scale” Actually Looks Like

Every founder story you’ve heard about scale started with the opposite of scale. The pattern shows up so consistently across the companies people idolize that it’s worth looking at closely, because it’s not folklore, it’s the actual documented approach.

Airbnb. Brian Chesky and Joe Gebbia had a handful of users in New York while they sat in Mountain View. When Paul Graham asked about their traction, Chesky admitted there wasn’t much. Graham’s advice wasn’t “post more” or “run ads.” It was: go to your users, in person, and get to know them one by one. So they flew to New York and went door-to-door for about a month, personally helping hosts take better photos and improve their listings. That single stretch of unscalable, in-person work is what taught them what Airbnb actually needed to be.

Stripe. The Collison brothers didn’t wait for merchant accounts to self-serve. They personally sat down with early users and manually set up traditional merchant accounts behind the scenes so it looked instant. It’s now known in startup circles as the “Collison install,” and it’s still cited in accelerator cohorts today as the standard for how far founders should go to get a handful of real users working.

DoorDash. No app at launch. Founders printed flyers, walked them door-to-door, and cold-called restaurants one at a time to build the initial marketplace. They made deliveries themselves so they’d understand the mechanics of their own business before they ever tried to automate it.

The common thread across all three: founders resisted this work because the early numbers look pathetically small. Ten users. Then twenty. It feels like it isn’t working. What actually happened is compounding they couldn’t see yet, and information you can only get by being in the room with your first users, not blasting them from a distance.

At zero audience, this is your unfair advantage, not your burden. You’re small enough right now to give every single early customer an experience no company at scale could ever give them again. That’s the thing to actually be doing instead of posting and waiting.

The Community-and-Referral Engine

Once you understand why cold fails and why manual, high-touch work is the actual unlock, the question becomes: where do you find the first people worth that level of effort? This is where community and referrals do the heavy lifting, and the data on both is specific enough to build a real system around.

Go to where the problem is already being talked about, not where your product category is. Reddit threads, niche Slack and Discord groups, industry Facebook communities, and competitor review sections are where people are actively venting about the exact pain you solve. Searching for the complaint gets you further than searching for your product category, because people describe problems in plain language long before they know a solution category exists.

Show up as a person before you show up as a founder. This matters more than it sounds like it should. Communities can tell the difference between someone who’s been useful for weeks and someone who joined yesterday to drop a link. The trust you build by answering questions and being genuinely helpful is what makes your eventual mention land as a recommendation instead of an ad.

Then, once you have even 3-5 users, ask each one for a single direct introduction. This is where the numbers get compelling. Warm introductions convert at 40-60% meeting rates versus 2-5% for cold outreach, referred leads close 3-5x faster, and one analysis found referrals generate 36x more revenue per lead than cold calls, with 37% higher retention on top of that. A single warm intro from an existing happy user is worth more than a hundred cold messages, and it costs you nothing but the ask.

This is also why the sequence matters. Cold outreach still has a role, but it’s for discovery, testing which pain points and phrasing actually land, not for closing. Once you know what resonates, you stop spending effort on cold volume and start spending your credibility on warm intros to the people who actually matter.

The Repeatable System

Here’s how Sections 1 through 3 come together into something you can actually run this week.

Step 1: Find the complaint, not the category.

Search Reddit, niche communities, and competitor reviews for the specific pain your product solves, in the words people actually use to describe it. Build a list of 20-30 people clearly stuck on this problem right now.

Step 2: Show up before you sell.

Spend one to two weeks being genuinely useful in the communities where these people are. Answer questions, share what you know, build a track record before you ever mention what you built.

Step 3: Go door-to-door, digitally.

Message people one at a time, like Airbnb and Stripe did in person. Not a pitch, an offer to help: “Saw your post about X, built something for it, want to try it and tell me what’s wrong with it?” Expect roughly 1 in 10 to convert. That means 100 messages gets you 10 customers, and that ratio is normal, not a sign you’re doing it wrong.

Step 4: Give an experience no company at scale could give.

Do the unscalable thing. Set it up for them personally. Get on a call. Fix their specific problem by hand if you have to. This is where you learn what to build next, and it’s the same move that built Airbnb, Stripe, and DoorDash.

Step 5: Ask each customer for one direct introduction.

Once you have 3-5 real users, ask each for a single specific person to introduce you to, not a public share. This is your highest-leverage move: warm intros convert 10-20x better than cold outreach, and this is how you get from 5 to 10 without needing an audience at all.

Step 6: Track the source of every one of your first 10.

Not the channel in general, the specific thread, community, or person. Most solo founders find one channel quietly did most of the work. Find yours early and go deeper there instead of spreading thin for the next 50.

None of this needs an audience. It needs 100 direct conversations with people who already have the problem, treated like Airbnb treated its first hundred hosts: like the most important users you’ll ever have, because for a little while, they are.

Your Way to Your First 10 Customers

There’s no shortcut here, and that’s actually the good news. Cold outreach isn’t broken because you’re doing it wrong, it’s broken because the entire channel decayed while nobody was watching. Warm trust didn’t get harder to build, it just moved to a different place: direct conversations, small communities, and the specific person willing to vouch for you.

Airbnb didn’t scale their way to their first hundred hosts. Stripe didn’t scale their way to their first real users. They went narrow, went manual, and treated ten people like ten people instead of a number to hit. That’s the whole unlock. You don’t need reach right now. You need 100 honest conversations with people who already have the problem, and the discipline to do the unscalable thing while you’re still small enough that it’s actually possible.

Your first 10 customers were never going to come from an audience you don’t have yet. They come from this. And once you have them, you’ll have something no cold list, no ad, and no audience could ever hand you at this stage: proof, told by someone else’s mouth, that what you built actually works.

The result of decades of that default is stark. Less than 1% of post-consumer textiles ever get converted into new garments. The other 99% ends up landfilled, incinerated, or shipped somewhere out of sight, out of mind, and out of the loop for good.

Where the idea came from

Pushpanjali Verma didn’t stumble on this problem from the outside.

She’s quick to point out that the problem doesn’t affect her personally in any direct way. She simply spent enough years inside the industry to see exactly where the seams come apart.

That’s the origin story behind Niche by Nature: a vertical marketplace built on circular economy principles, designed to fix fashion’s linear supply chain.

What it’s not:

• Not a recycling app

• Not a resale platform

What it is:

• A marketplace that sits at the intersection of brands, textile recyclers, textile sorters, and consumers

• A system that tries to make all four talk to each other in a way that actually closes the loop

The hard part isn’t the idea. It’s explaining the idea.

Ask Pushpanjali what’s hardest about building this, and she won’t say funding or hiring. She’ll say explaining circular economy as one coherent business.

What’s actually happening: it’s building the connective tissue between players who already exist, who already do their piece of the puzzle, but who’ve never had a reason or a system to coordinate with each other.

That’s a pitch that’s easy to say and hard to land in one sentence. The kind of problem that scares off some founders and pulls in others. Pushpanjali is clearly the second kind.

Why the timing matters

The EU has made circular economy one of its biggest regulatory pushes right now:

• Extended producer responsibility rules

• Mandatory textile collection targets rolling out across member states

That means real deadlines and real compliance pressure landing on fashion brands, whether they’re ready or not.

Pushpanjali is betting this regulatory wave turns “nice to have” into “have to have” for exactly the kind of infrastructure Niche by Nature is building. Large scale. Genuinely difficult to execute. Hard to ignore, for the same reasons.

What’s kept her going

Not a big raise. Not a viral moment. Not a glowing press writeup.

When asked what’s meant the most so far, Pushpanjali points to something simpler: strangers from across the globe showing up to help without ever meeting her in person. People joining the effort on trust alone, before there was a product to point to or a case study to share.

That’s the part of early-stage building that never makes it into a pitch deck. Nobody puts “random people on the internet believed in this before there was proof” on a slide. But it’s usually the actual difference between founders who keep showing up on the hard weeks, and founders who quietly stop.

Where it’s headed

The goal right now is deliberately narrow: connect brands, textile recyclers, textile sorters, and consumers well enough to hit product-market fit. Not five separate products chasing five different customers. One working loop, proven with real transactions, before anything else gets added.

Niche by Nature is being built by two female founders, on what they call, in their own words, a very long journey. It’s long because circular supply chains don’t get rebuilt in a quarter, and because coordinating four different stakeholder groups takes more trust-building than a typical two-sided marketplace.

It’s worth watching for the same reason. The regulatory wave behind it isn’t slowing down, and someone is going to build the infrastructure fashion brands will eventually be required to use. Pushpanjali is betting it’s her.

Follow Pushpanjali’s build: LinkedIn · nichebynature.com

None of those steps are hard on their own. Stacked together, they add up to a part-time job that most founders never have time for. Most lean teams don’t have the headcount for it either. Shadrack didn’t.

“As a founder myself, I know how hard it is to grow when you do not have a big SEO team or endless time,” he says. So he built SerionFlow: a tool that turns search gaps into branded pages on your own domain, then keeps improving them using real Google and Bing search data.

The idea wasn’t to replace an SEO team. It was to give founders and small teams a repeatable system for doing the work an SEO team would normally do, without needing to hire one.

The trap he wanted to avoid

The obvious risk with a tool like this is that it turns into another AI content generator. Type a keyword, get a page, publish, repeat. Shadrack didn’t want that.

“SEO pages need trust, brand voice, structure, quality, and a real reason to exist,” he says. Generic, auto-generated pages might get indexed, but they rarely earn a reader’s trust, and search engines have gotten better at spotting them too.

So instead of a one-click generator, SerionFlow works as a full loop. It scans the brand to understand voice and positioning, finds real search opportunities, generates pages for review rather than auto-publishing them, publishes approved pages under the user’s own domain, refreshes sitemaps, submits updates to Google and Bing, and then uses performance data to improve pages over time.

The reframe that changed everything

For a while, Shadrack thought of SerionFlow as an AI content tool. That framing kept limiting how he built and talked about it. Every conversation about the product ended up being about writing quality, when writing was never really the hard part.

The shift came when he stopped thinking about content and started thinking about workflow. Find the opportunity. Generate the page. Publish on-domain. Submit to search engines. Track performance. Improve using real Search Console data.

“That changed the product, the messaging, and the way I explain it to people,” he says. Once he framed SerionFlow as a replacement for a messy, manual process rather than a writing tool, it became easier to explain what the product actually does and who it’s actually for.

Early traction

SerionFlow launched publicly on January 15, 2026, after months of building and private testing with startups across different niches. That private testing period mattered. It let Shadrack see how the full loop, from opportunity discovery through published pages, actually held up across different brands and industries before opening it up more broadly.

Along the way, SerionFlow was accepted into both the Google for Startups Cloud Program and the MongoDB startup program. Shadrack calls it a huge milestone, and it’s the kind of validation that matters most in the early days: outside programs backing the product before the founder has a big case study to point to.

Where he wants this to go

Ask Shadrack what success looks like in a year, and he doesn’t talk about page counts or traffic charts. He talks about trust.

“I want to see customers using SerionFlow to consistently discover search opportunities, publish useful pages, track real results, and build a stronger search presence over time,” he says. “The goal is not to flood the internet with more generic pages. The goal is to help businesses create useful, brand-matched pages around real search intent and keep improving them with actual data.”

That distinction, between flooding the internet and building a real search presence, is the line Shadrack keeps coming back to. It’s easy to build a tool that generates a lot of pages fast. It’s harder to build one that generates pages worth having, and then keeps them improving long after they’re published.

For founders and small teams who know they need SEO but don’t have a team to run it, that’s the bet SerionFlow is making: that a review-first, data-driven workflow beats a pile of generic pages every time.

Check out SerionFlow at serionflow.com.

Fifteen months earlier, the same founder launched a side project on the same platform. It got featured. It pulled in half the upvotes, 300. It converted 91 paying customers.

Nothing about his effort, his product quality, or his strategy changed between those two launches. What changed was Product Hunt itself. That single data point is the best argument we’ve seen for why “how many upvotes did you get” is close to the least useful question you can ask about a launch.

We pulled together eight documented launches, real names, real numbers, mostly from the founders’ own published accounts, to figure out what actually predicts a launch working versus just looking like it worked.

Myths this debunks

• “You need a well-known hunter to get featured.” Nitro’s team launched three separate times, never once used a hunter, and hit the top 5 all three times.

• “A polished demo video is basically required.” Dub.co skipped the conventional wisdom on video production spend and still won #1 Product of the Day; several other founders in this research reported scrappy screen recordings outperforming agency-made videos.

• “More upvotes means more customers.” Tomas Blatak’s two launches prove this false in the most direct way possible: double the upvotes, roughly 1% of the customers.

• “Mobilizing your whole network at once maximizes your shot.” useDaily.ai’s founder did exactly this, a single coordinated email blast across three platforms, and got throttled by the same algorithms he was trying to trigger.

• “The launch is the growth event.” In every strong outcome here (Dub.co, Arc, Customerly), the actual growth event was the weeks or months of community-building before launch day. Launch day just cashed it in.

What a vote actually cost

The pattern: the founders who spent months building real relationships before launch day spent very little on launch day itself, and it worked. The founder who spent his launch-day effort on volume and simultaneity, rather than relationships built earlier, got flagged by the exact systems he was trying to activate.

How far ahead did they actually start

If there's one number to take from this whole piece, it's this one. Nobody who started less than a few weeks out had a clean win. The founders with a multi-month head start are also the ones who didn't need to spend money on launch day.

Which founder are you closer to right now

• No community yet, first launch, feeling behind: you’re closer to Hurree or useDaily.ai’s starting position. Hurree still came out fine because they focused their limited time on operational discipline (24-hour coverage, tracked links) rather than volume. useDaily.ai didn’t, and got throttled. If this is you, spend your remaining time on one or two authentic relationships a day rather than one big blast at the end.

• Relaunching an existing product with real improvements: you’re in Nitro’s position. Their advantage wasn’t a clever trick, it was simply showing up again, better, without needing a hunter. If you have a product that’s meaningfully improved since a past launch, a relaunch is a legitimate strategy, not a sign of desperation.

• You have months of runway before you need to launch: you’re in Dub.co’s or Customerly’s position, and the highest-leverage thing you can do isn’t anything on this list of launch-day tactics. It’s building the pre-launch waitlist or community now, so launch day becomes a formality rather than a gamble.

• You already tried once and it didn’t work: you’re in Tomas Blatak’s or Customerly’s earlier position (their first live chat launch also underperformed). Both cases show one underwhelming launch doesn’t predict the next one, especially if the platform, product, or your community has changed since.

Share Behind the Launch

The eight launches, in detail

1. Tomas Blatak (Leadspicker): the same founder, two launches, opposite outcomes

What happened:

• June 2023: Blatak launched a side project, an AI-powered VC sheet tool, under his company Leadspicker. Featured, 7th Product of the Day, 300 upvotes, 91 paying customers.

• September 2024: he launched Leadspicker itself, the flagship product. Not featured, but most-upvoted for most of the day, finishing with 612 upvotes. Paying customers: 1.

What changed in between: In September 2023, Product Hunt featured an average of 47 products per day. By January 25, 2024, the featured rate had dropped sharply. By September 2024, only about 16 products per day were being featured. Non-featured launches, regardless of upvotes, typically see only 100 to 500 visitors and 1 to 15 signups, versus 1,000 to 5,000 visitors and 10 to 150 signups for featured ones.

Takeaway: Upvotes and rank are not revenue, and platform rules can shift under you between launches. Check whether “Featured” status is realistically in reach before treating a launch as a customer-acquisition bet rather than a credibility play.

2. Lachlan Chavasse (useDaily.ai): when doing “everything right” backfires

What happened: Chavasse messaged roughly 1,000 LinkedIn connections, brought a QR code to events, and collected about 60 confirmed sign-ups via a Google Form. He locked in his launch date a week ahead but didn’t finalize his site and graphics until the last minute. At 8am on launch day, he sent one email with three links at once: Product Hunt, LinkedIn, and X.

What worked: Genuine engagement: 100+ upvotes, about half with comments, all personally answered.

What didn’t: The simultaneous, coordinated push read as inauthentic to the algorithms. Product Hunt throttled the points his upvotes were worth and LinkedIn barely surfaced the post. He finished stuck at #26 all day, with fewer than 50 real points despite the 100+ raw upvotes, and was never featured.

Takeaway: Mobilizing your network isn’t the same as forcing everyone to click the same link at the same moment. Stagger the ask.

3. Steven Tey (Dub.co): the teaser waitlist that turned into #1 of the Month

What happened: Dub.co launched March 21, 2024, after over a year of building an open-source, self-hostable Bitly alternative. Months before launch, the team used Product Hunt’s teaser page feature to build a waitlist directly on the platform. On launch day, Tey posted live vote counts publicly: 330 upvotes and 101 comments at hour 4 (already #1 of the Day), 420 by hour 5, climbing through #5 and #2 Product of the Week before finishing with 1,085 votes and 211 comments as #1 Product of the Day, later also #1 of the Month.

What worked: Months of community-building before launch day, radical transparency about live numbers during the day, and personally replying to nearly every comment, a habit Tey has kept up long after.

Takeaway: The pre-launch waitlist is the actual campaign. By the time launch day starts, the outcome is mostly already determined by the relationships built beforehand.

4. Ksenia Khriptovich (Nitro, Alconost): three launches, zero hunters, compounding results

What happened: Nitro is a B2B translation platform, notoriously hard to promote on Product Hunt against consumer-friendly tools. The team launched three times as the product matured: “Nitro,” then “Nitro 2.0” after adding Google Docs/Sheets support, then “Nitro 3.0” after adding JSON and iOS string file support. Nitro 2.0 placed #4 with 335 upvotes on launch day (500+ within the week). Nitro 3.0 was a genuine nail-biter, swinging between 6th and 9th place before breaking into the top 5 and defending it for 12 more hours.

What worked: Never using a hunter across all three launches, warming up team PH accounts a full month before each one, and genuinely improving the product between launches rather than relaunching the same thing.

Bonus: the launches also generated a client for Alconost’s separate localization service.

Takeaway: You don’t need to buy your way in with a hunter. A slow, honest presence built across months, then relaunching as the product genuinely improves, compounds.

5. Arc.dev (Arc 3.0): the win that came after 26 tries

What happened: Arc, a remote-talent marketplace with a 300,000+ developer email list, had already launched 26 times before Arc 3.0. Weeks ahead, the team pitched the product in person to 30 makers at a Product Hunt meetup in Taipei and ran an internal “Arc Team Challenge” to mobilize the whole company with assigned roles. They spent $770, their largest single campaign cost, on a freelance-animated launch video.

Result: 813 upvotes, 127 comments, their best launch ever, 4x the performance of any of the previous 26.

Takeaway: An underperforming launch isn’t a signal to quit. The compounding came from in-person relationships built months ahead and a company-wide effort with assigned ownership, not from a single launch-day trick.

6. Mangools: one team, three tools, three different launch styles

Takeaway: Even a missed or reactive launch can produce real signups. But deliberate planning, especially cross-promotion across assets you already own, clearly outperforms leaving it to chance.

7. Luca Micheli (Customerly): comments, a S.M.A.R.T. goal, and a cat named Katy Purry

What happened: Customerly’s first Product Hunt attempt (a live chat tool) underperformed. For their second attempt, a free knowledge base tool, they set one explicit goal: 100 signups from Product Hunt. Their prep centered on a private “Founders Club” Facebook group built over the prior three months. Their launch content included a viral story about a cat solving a problem using their product. On launch day, they asked for sincere comments, not upvotes, paired the free tool with a 50%-off coupon for their paid products, and ran a 90-person Zoom webinar mid-launch.

Result: They beat their goal, ending with 200+ new company signups in two days, still generating leads five months later, finishing at #2.

Takeaway: One measurable, unglamorous goal beats “hit #1.” A launch-day discount on a paid product can convert curiosity into revenue immediately, and a low-stakes, fun story can outperform a straight feature list for shares.

8. Hurree: treating launch day as a 24-hour operation, not an event

What happened: For their first Product Hunt launch, Hurree staffed shifts so their UK-based team covered the full 24-hour window, not just their local working day. They tracked a clear engagement dip as their GMT day wound down and a distinct uptick around 17:00 GMT as other time zones woke up. They used two separately tracked links, a landing page and a direct free-trial signup, to see which converted better. After launch, they sorted every piece of feedback into praise, suggestions, and criticism to shape their roadmap.

Takeaway: Your launch doesn’t run on your local clock. Plan deliberate coverage for the hours your usual team is asleep, and use separately tracked calls-to-action to learn what actually convinced people.

Why we built SuperLaunch differently

Almost every founder in this piece was, in some way, at the mercy of an anonymous upvote count and an opaque algorithm deciding whether their work got seen. Blatak’s story is the starkest version: the exact same effort produced wildly different outcomes because of a platform change no founder could see coming or control.

That’s the specific problem we built SuperLaunch around. Instead of anonymous votes, named SuperReviewers actually test a product and leave accountable feedback. Instead of a single 24-hour window where timing and algorithm luck decide everything, founders get a multi-week visibility window. It’s not a claim that we’ve solved launch unpredictability entirely, but it’s a direct response to the exact failure mode this research keeps surfacing.

You wrote the tagline twelve different ways. You lined up your first comment. You set an alarm for midnight PST because that’s when the Product Hunt day “starts,” and somehow that detail felt important enough to lose sleep over.

You launched.

You refreshed the page every few minutes for the next six hours. You watched the upvote count. You watched the comment count. You watched your rank.

Twenty-four hours later, it was over.

Maybe you finished #27. Maybe #41. Maybe you got a dozen comments, half of them from people you personally messaged. Maybe the “Maker” badge is the only thing that feels real about the whole day.

Now what?

Almost every Product Hunt guide on the internet stops right here. They’ll tell you how to pick a launch date, how to build a hunter list, how to write a tagline that doesn’t sound like everyone else’s tagline. None of them tell you what to do the morning after, when the traffic has dried up and you’re staring at an analytics dashboard that looks nothing like the screenshots in those guides.

This article starts there.

Your Launch Probably Didn’t Fail

Let’s get one thing out of the way first.

A lot of founders define “failure” as not hitting Top 5. Or not getting featured in the newsletter. Or not crossing some round number of upvotes they’d quietly decided mattered.

None of that is actually a measure of whether your product is good, or whether people want it, or whether you have a business.

Product Hunt measures one thing: how your launch performed on Product Hunt, on that specific day, against whoever else happened to be launching at the same time. That’s it. It’s not a referendum on your startup.

A few things that genuinely affect your rank and have nothing to do with your product:

None of this is an excuse to ignore real feedback. If people tried your product and bounced, that’s signal worth taking seriously. But the leaderboard position itself? That’s noise dressed up as a verdict.

The founders who build real companies aren’t the ones who nailed launch day. They’re the ones who treated launch day as one data point and kept moving.

Stop Doing These Things

Before we get into what to do, here’s what not to do in the days right after a quiet launch. These are the traps that waste time and momentum.

↳ Assuming nobody wants the product. A flat Product Hunt day tells you about Product Hunt traffic, not market demand. You need more signal before drawing that conclusion.

↳ Bolting on random features immediately. This is the classic panic move. A weak launch doesn’t mean your roadmap is wrong. It means your distribution day was rough.

↳ Changing your pricing overnight. Pricing decisions deserve data, not panic. Give yourself at least a week of real usage before touching it.

↳ Obsessing over the final rank. It’s already final. There’s nothing left to extract from staring at it.

↳ Quietly giving up on marketing. This is the costliest mistake on the list. A bad launch convinces a lot of founders that “marketing doesn’t work for us,” when really one channel, on one day, didn’t work.

↳ Comparing yourself to VC-backed launches. A team with a marketing hire, a PR budget, and 50,000 X followers is not running the same experiment you are. Different inputs, different outputs. Comparing your day-one numbers to theirs isn’t useful information, it’s just a way to feel bad.

Action box: Before you change anything about your product, pricing, or strategy, give yourself 72 hours to actually look at the data. Section 3 tells you exactly what to look at.

Day 1–3: Diagnose

Resist the urge to react. First, understand what actually happened.

Here’s the checklist:

Traffic

• Where did visitors come from? Direct Product Hunt traffic vs. social shares vs. people who found you elsewhere that day.

• How long did people stay on the page?

Signups

• How many people signed up relative to visitors? This is your top-of-funnel conversion rate, and it matters more than your PH rank ever will.

Activation

• Of the people who signed up, how many actually did the core action your product is built around?

Retention

• Did anyone come back the next day? The next week? Early retention, even from a small number of users, tells you more than a launch-day spike ever will.

Comments

• Read every single one. Not for ego. For patterns. Confused comments point at unclear positioning. Skeptical comments point at trust gaps. Genuinely excited comments point at who your real audience is.

Emails and DMs

• Anyone reach out personally? Those are often your most motivated early users, and they deserve a real reply, not a template.

User interviews

• Pick five people who signed up and ask if they’ll do a 15-minute call. This single step usually produces more useful information than the entire launch day combined.

Product Hunt tells you how many people looked. It doesn’t tell you why they left, or why a few of them stayed. Only you can find that out, and only by asking.

Week One: Improve the Product

With three days of real data, you’ll usually find the same handful of friction points showing up again and again. Fix those. Resist the urge to do anything bigger.

Focus on:

• Landing page clarity. If multiple comments asked “wait, what does this actually do?”, that’s your answer. Rewrite the headline before you touch anything else.

• Onboarding. Where did signups drop off before activating? That’s your highest-leverage fix this week.

• Pricing friction. Not a full pricing overhaul. Just check whether the price was confusing, hidden, or mismatched with what people expected.

• Bugs that came up under real traffic. Launch day is the first time most products see meaningful concurrent usage. Fix what broke.

• Your call-to-action. Is it obvious what someone should do next on your site? Half of post-launch traffic loss is just unclear next steps.

That’s the whole list. Don’t rebuild your product based on one rough day of feedback. Tighten what you have, then go get more people to look at it.

Week Two: Restart Distribution

This is the part most “how to launch on Product Hunt” guides skip entirely, and it’s the most important week in this whole plan.

Product Hunt was never supposed to be your only distribution channel. It’s one channel, with one specific audience, active for one day. Treating it as your entire go-to-market plan is the actual mistake, not your rank.

Here’s where else to put your energy in week two:

• Reddit. Relevant subreddits for your category, plus broader startup communities. Reddit rewards honesty and punishes anything that reads like a pitch. The upside is an audience that already trusts peer recommendations over ads.

• LinkedIn. Especially if your buyer is a working professional rather than a hobbyist builder. A founder-written post about what you learned from launch outperforms a polished announcement almost every time, because people engage with people, not press releases.

• X (Twitter). Building in public, sharing real numbers, and replying to people in your space all compound here. Every reply is a small chance to be discovered by someone in that person’s audience too.

• Founder communities and Discord servers. Indie Hackers, niche Slack and Discord groups, and founder-only communities are full of people who’ll actually try your product and tell you the truth about it. The feedback tends to be more useful than launch-day comments, because there’s no leaderboard incentive distorting it.

• Facebook groups. Underrated for certain categories, especially local services, parenting, hobbies, and small business tools, where the audience often isn’t active on X or Reddit at all.

• Directories. There are dozens of smaller directories beyond Product Hunt that take submissions year-round, not just on one launch day. Each one is a small, low-effort shot at incremental traffic and a backlink.

• Newsletter submissions. Niche newsletters in your category often have far higher intent-to-attention ratios than big platforms, since their readers opted in specifically for that topic.

• Show HN. A different audience than Product Hunt, skewed more technical, and worth a separate attempt if your product fits. Treat it as its own launch, not a rerun of your PH post.

• Alternative and ongoing launch platforms. Some platforms are built specifically for the weeks after a Product Hunt launch, not just one big day. This is exactly the gap SuperLaunch was built to sit in: a place where your tool stays visible and gets structured feedback from real reviewers over days and weeks, not just one 24-hour spike.

• Direct email outreach. Slower, but the highest-intent channel available to you. A short, specific email to someone who fits your ideal user beats a hundred passive impressions.

Each of these contributes a little. None of them will single-handedly replace a Top 5 Product Hunt finish, and that’s fine. The goal isn’t one big spike. It’s several small, repeatable ones that stack.

What we actually tried

We’ve run SuperLaunch through some of this ourselves, so here’s what happened, without the highlight-reel version.

Reddit, with no pitch at all. A few days after our own launch, distribution had stalled. We posted in five to seven relevant subreddits offering free, honest feedback on AI tools and SaaS products from a small community of reviewers. No pitch. No paid angle. Just an offer framed around “real feedback, not cheerleading,” aimed at early-stage founders and vibe coders who were tired of empty encouragement.

One post, late at night, turned into 17+ tool submissions, 18+ comments, and over 900 views within a single hour, along with direct messages from founders asking to join. That post did more for our cold-start problem than launch day itself. After submitting the same post on about 4-5 subreddits, we got 15K+ views from Reddit alone.

What we learned: the offer mattered more than the platform. “Free honest feedback, no strings” is a message Reddit users actually trust, because it doesn’t read like marketing.

Direct outreach instead of waiting to be found. Rather than relying on people discovering us organically, we started reaching out to founders directly, on LinkedIn and in founder communities, the same way we’d want someone to reach out to us. Slower than a viral post, but it’s the channel that’s produced the most consistent, qualified submissions over time.

What we learned: distribution doesn’t have to be passive. Asking directly, one founder at a time, works even when it doesn’t scale the way a Reddit post does.

A small paid test on X, almost by accident. We ran an X Ads experiment with a budget of $6. That’s not a typo. It generated over 85,000 impressions and led to 10-12 product submissions.

What we learned: on a platform where the algorithm is actively looking for content worth amplifying, even a tiny, almost throwaway budget can outperform what you’d expect, if the offer and the targeting are right. We’re not claiming this is repeatable at scale, but it’s proof that you don’t need a real ad budget to find out whether a channel is worth more investment.

Learn from Other Founders

You’re not the first founder whose Product Hunt launch undersold the product. Here’s what happened to a few others, and what they did next.

A few patterns show up across almost every one of these:

↳ Unnamed SaaS founder Called the launch “just a brief moment,” not the real growth driver. Growth actually came from Discord, Reddit, and niche craft forums discovered after launch. No exact numbers shared, but the founder said these later communities mattered more than launch day. Lesson: a quiet launch can still be useful if it pushes you to find where your real users hang out.

↳ Starter Story Got awareness and backlinks from Product Hunt, not much direct traffic. Growth actually came from Reddit and organic SEO. Roughly 2,000 visitors and 100 subscribers from the PH launch itself; real traction came later. Lesson: Product Hunt is often a visibility event first, an acquisition channel second.

↳ EmailEngine Strong credibility and backlinks from launch, minimal direct conversion. Growth actually came from open-source community traffic. Customer acquisition came from open-source channels, not the launch. Lesson: a launch can pay off in trust and SEO long before it pays off in revenue.

↳ NoteForms Traffic, credibility, and backlinks from launch, but zero direct sales. Growth actually came from long-tail SEO and brand visibility over time. Valuable backlinks, no immediate sales. Lesson: search traffic frequently outlasts launch-day hype as a real channel.

↳ ChatPal Strong launch metrics, but the founder credited the real value to the focus the launch forced, not the leaderboard spot. Growth actually came from founder-led content and continued engagement with the launch community on X. About 2,000 site visits, 137 downloads, and 50+ backlinks within 36 hours. Lesson: a good launch is also a forcing function. The content habits it builds can outlast the spike itself.

↳ BaseTemplates Launch started as what the founder called “invisible growth” before becoming a real inflection point. Growth actually came from ongoing brand awareness rather than a single spike. Grew toward $10K+/month over time. Lesson: the launch can be a catalyst. The compounding happens in what you build after it.

↳ Questflow Reported a real SEO boost in the window after launch. Growth actually came from community activation that spilled into organic search. Founders described it as “a serious bump” alongside 5,100+ page views. Lesson: launch attention can convert into durable search visibility when it earns mentions and links.

↳ History Search Used the launch as a path from early idea to first revenue, with content explaining the product alongside it. Growth actually came from content that clearly explained the product and the problem it solved. 1,000 signups in the first two days. Lesson: clear explanatory content keeps working long after the launch-day traffic fades.

A few patterns show up across almost every one of these:

• Product Hunt frequently works as an awareness and credibility layer, not the acquisition engine itself.

• Reddit, SEO, open source, and founder-led content show up repeatedly as the channels that actually convert.

• A quiet launch isn’t a dead end. It’s often the moment a founder finally figures out where their real users are.

Build Long-Term Distribution

The founders who keep growing after a flat launch all share one habit: they’re not relying on a single event to carry their distribution. They’ve built assets that keep working without them refreshing a leaderboard.

The assets worth investing in:

• A blog. SEO content compounds. An article you write this month can still be bringing in visitors a year from now, long after a launch-day spike has flatlined.

• A newsletter. Owned audience, no algorithm in between you and your reader. Every issue is a guaranteed touchpoint, not a hope that the feed cooperates.

• A community. Discord, Slack, or even a simple group chat of your most engaged users. These people become your earliest evangelists and your most honest critics.

• SEO. The long, slow channel that almost nobody mentions in launch guides, and almost everyone wishes they’d started earlier. It’s the one channel that keeps working while you sleep.

• YouTube. Especially powerful for products that benefit from being seen in action, where a single demo video can keep surfacing in search for years.

• An X audience. Built through consistent, specific posting, not just launch-day threads. The accounts that grow are the ones that show up daily, not the ones that post once and disappear.

• A LinkedIn audience. Particularly valuable for B2B and professional tools, where buyers are more likely to trust a founder’s post than a company page.

• An email list. The one channel you fully own, regardless of what any platform decides to change tomorrow.

A launch is an event. These are assets. Events end. Assets compound.

Your 30-Day Recovery Plan

Here’s a practical, week-by-week roadmap for the month after a launch that didn’t go the way you’d hoped.

Week 1: Understand what happened Review traffic, signups, activation, retention. Read every comment for patterns. Talk to 5 early users. Fix the highest-friction onboarding step.

Week 2: Restart distribution Publish an honest launch recap. Post in 5-7 relevant Reddit and founder communities. Submit to 2-3 directories beyond Product Hunt. Send 10 direct outreach messages to ideal users.

Week 3: Build proof and content Collect 2-3 testimonials from early users. Publish one piece of content explaining the problem you solve. Reach out personally to 10 more potential users or partners.

Week 4: Ship and measure Release the fixes from Week 1. Tell your existing users what changed and why. Measure traffic, signups, and retention against Week 1’s baseline. Repeat what worked.

Pick one metric from Week 1 (signups, activation rate, or retention) and track it every week through Week 4. If that one number is trending up, your recovery is working, regardless of what your Product Hunt rank says.

Final Thoughts

A Product Hunt launch lasts one day.

Distribution lasts as long as you keep showing up.

The founders who win aren’t the ones who peaked on launch day. They’re the ones who treated launch day as the start of a longer conversation with their market, not the end of one.

If your launch didn’t go how you hoped, you’re not behind. You’re exactly where most successful founders have been at some point. The only real difference between the ones who recover and the ones who don’t is what they do in the 30 days right after.

Now you know what to do with yours.

If you're past your Product Hunt launch and looking for ways to keep momentum going, SuperLaunch is built for exactly this stage.

It's a place where your tool stays visible for days and weeks, not just one, and gets structured feedback from a community of real reviewers who'll actually use it and tell you what they think.

No leaderboard to obsess over, just ongoing distribution and honest feedback while you keep building. Worth a look if you're still figuring out where your next users are coming from.

He held onto one conviction through all of it. Support isn’t a cost center, it’s the reason people stay. Treat someone well when something breaks, and you’ve earned more loyalty than any discount ever could. So when the cracks started showing, scaling beyond what his team could carry, he didn’t go looking for a shortcut. He went looking for something that actually understood the problem.

He didn’t find it. The tools he tried either guessed wrong when it mattered most, or worked well but cost more every month than the last, growing more expensive in direct proportion to how much they actually helped. There was a gap between what existed and what the job actually demanded. Steven decided to build the thing himself.

Nobody tells you how hard the first yes is

Sarrai launched into a market already full of people trying to solve the same problem, each in their own way. Steven doesn’t resent that. He thinks there’s room for a lot of different ideas here. But it meant he couldn’t just put something out into the world and wait.

He’d believed, the way a lot of builders do at the start, that good work speaks for itself. Build it well enough, and people find their way to it. It took a few rough, humbling months to learn that isn’t how it works anymore. He had to go out and ask. Post in forums. Show up on Reddit, on Slack, on LinkedIn. Request feedback from strangers who owed him nothing. “The first customer is the hardest one to get,” he says. “And let me tell you, they were right.”

What kept him going through that stretch wasn’t traction. It was a conviction he’d carried since his support days: that somewhere, a knowledge expert was on holiday, unreachable, while their team scrambled without them. He’d lived that stress himself, on both sides of it. The person who has the answers but can never fully switch off. The person who needs an answer and has nowhere to turn. He built Sarrai so that knowledge could finally live somewhere besides one person’s head, getting a little sharper every time someone had to step in and help.

Share

The moment it clicked

Then, one ordinary day, he posted about something new he’d shipped. Within hours, his inbox had several people in it, asking for demos, wanting to talk. Not because he’d pitched anyone. Because the work spoke for itself, finally, after months of speaking for it himself. “The biggest validation I have ever received,” he calls it.

It’s a small moment in the scope of a company. But for a founder who’d spent months wondering if any of this was landing, it was proof that it was.

What he’s still building toward

Sarrai is young, and Steven hears that more than he’d like. People want to see real users before they commit. He expects that to ease in time, maybe another year and a half of doing the work before the hesitation fades on its own.

But when he talks about what success actually looks like, it isn’t a valuation or a headline. It’s simpler than that. An easier conversation with the next person he tries to convince. Enough coming in that the work pays for itself, and pays him too. The kind of quiet, sustainable proof that the thing he built out of his own exhaustion is now holding someone else up.

Steven shares more of the build at @Steven_Lemmens and on LinkedIn.

In one room, small business owners know AI matters — but most of them don’t need another tool. They need someone to help them see where time, money, and follow-up are quietly leaking out of the business, and how AI can support better systems without replacing their own judgment.

Brandon Russell

In the other room: teachers. Told to use AI, often overwhelmed, undertrained, handed tools with zero context for how to actually use them. Brandon teaches business full-time in a public school. He sees that gap up close, every week.

“I have lived on both sides,” he says. He spent years as a business owner and consultant before becoming a teacher — so he knows what it’s like to build something new while still carrying a full-time job, a family, and the slow pressure of trying to create a better future one piece at a time.

4THDMC | EVOLVE is his attempt to build the bridge he kept wishing existed.

The hardest lesson: clarity beats complexity

Here’s the part most founders skip when they tell their story — the part where the idea was too true.

Brandon had the experience, the tools, the philosophy, the credentials. The problem wasn’t a lack of substance. It was that the business sounded like consulting, education, AI tools, content, strategy, and personal reinvention, all bundled into one explanation. Every piece of it was real. None of it was simple.

“The hard part was not creating more,” he says. “It was reducing. Choosing the front door. Deciding what gets said first and what waits.”

His takeaway is one every early founder should tape above their desk:

“If people cannot repeat what you do in one sentence, they probably cannot refer you either.”

The shift that gave the business a spine

The real turning point came when Brandon stopped treating AI as the product.

He’s not selling AI tools. He’s redesigning the work around them — workflow audits, revenue-leak diagnosis, follow-up systems, SOPs, and AI-assisted operations for small businesses. For teachers, that means tools that cut workload while keeping professional judgment exactly where it belongs: with the teacher.

“That one distinction gave the company a clearer spine,” he says. “It’s not about chasing tools. It’s about building systems that keep working after the excitement wears off.”

What success looks like from here

Brandon isn’t chasing a headline number. In the next 12 months, success means a real consulting pipeline — signed clients, actual case studies, business owners using systems his team helped build — alongside a stronger user base on the educator side.

Personally, it comes down to one specific, unglamorous target: $4,600 a month, for three straight months. Not for vanity — for freedom. That number is the difference between a side project he works on after hours and the main road.

“The bigger version of success is simpler,” he says. “People who felt behind because of AI now feel capable, clear, and less trapped in their work.”

4THDMC | EVOLVE was formed as an LLC in April 2026, and Brandon is building it the way most real businesses actually get built — piece by piece, after hours, some days feeling like momentum and some days feeling like pushing a boulder uphill. The company philosophy doubles as the only instruction he’s really following:

Share Behind the Launch

He was tired. Not the kind of tired that a weekend fixes. The kind that builds up over years of writing code for other people’s dreams while yours sit in a notes app gathering dust.

He told the internet he was quitting his job to build products. If nothing worked in twelve months he would go back to employment.

That person is Andrey Azimov.

And he documented every single day of it publicly.

Who is Andrey Azimov?

Andrey wasn’t a household name before his year of building. He was a developer with a regular job, a few side project ideas, and a growing frustration with the gap between where he was and where he wanted to be.

He didn’t have a big audience. He didn’t have a safety net. He didn’t have a co-founder or investors or a roadmap.

He had a year. A thousand dollars. And the internet watching.

The deadline that changed everything

Most people who want to go indie never do because the timing is never right. There’s always a better moment — after the next raise, after the kids are older, after the savings are bigger.

Andrey removed the option of waiting.

He started building. And he started sharing.

Every week he posted updates. What he was building. What was working. What wasn’t. How much money he had left. How he was feeling about it.

The transparency wasn’t a strategy. It was accountability. If he stopped posting people would notice. If he failed quietly he could pretend it didn’t happen. Doing it publicly meant the only way out was through.

What he built

His first product was Hack My Month — a simple tool that helped people plan and track monthly goals. Not a revolutionary idea. Not a technically complex build. Just something useful, built fast, and put in front of people.

It didn’t make him rich. But it made him something more valuable at that stage — it made him a person who ships.

Then came Sheet2Site. A tool that let anyone turn a Google Sheet into a simple website without code. The idea was almost embarrassingly simple. A spreadsheet. A URL. Done.

It found users immediately.

Sheet2Site became his proof of concept — not just that the tool worked, but that he could build something people actually wanted to use. That he wasn’t just dreaming about being an indie maker. He was one.

The real story behind the year

When people talk about Andrey’s story they focus on the products. But the products were almost beside the point.

What he was really building during that year was evidence. Evidence that he could do this. Evidence that shipping was possible. Evidence that the life he wanted — working on his own terms, on his own products, without asking anyone for permission — was within reach.

The public documentation forced him to stay honest. There were hard weeks — weeks where nothing shipped and the doubt crept in. He wrote about those too. Not because vulnerability is a content strategy but because pretending everything is fine is exhausting and nobody learns anything from a highlight reel.

The deadline forced him to stay moving. When you have twelve months and a counter everyone can see, paralysis is not an option.

Where he is now

Andrey crossed the finish line. He didn’t go back to employment.

He showed that you don’t need a perfect idea. You don’t need a big budget. You don’t need the right moment.

You need a deadline. And the willingness to do it in public.

Three things he did that most founders won’t

He made failure a real option — and then worked like it wasn’t.
By setting a hard deadline with a return-to-employment clause he took the “I’ll figure it out eventually” escape hatch away from himself. The deadline was real. That made the work real.

He shipped imperfect things and learned from real users.
Sheet2Site was a simple idea. Hack My Month was a simple idea. Neither required months of planning. Both got built, launched, and in front of people fast. The feedback loop from real users taught him more in weeks than planning would have in months.

He made the journey the content.
Most people wait until they succeed to tell their story. Andrey told it as it happened — including the uncertain weeks, the shrinking bank balance, the moments of doubt. That honesty built an audience that was genuinely invested in what happened next.

What founders can learn from Andrey Azimov

Set a deadline you can’t ignore. Vague goals produce vague results. A specific end date with real consequences changes how you work every single day.

Start with your constraints, not despite them. $1,000 and one year sounds limiting. It was also exactly the pressure that kept him focused and shipping.

Document the process not just the outcome. The story of building is more interesting than the story of having built. Share it while it’s happening.

Ship something simple first. Sheet2Site was a spreadsheet connected to a URL. It worked. It found users. It proved the concept. You don’t need complexity to create value.

Go public with your commitment. Accountability to an audience is one of the most underrated productivity tools available to any founder. Use it.

Andrey Azimov writes and builds at https://www.andreyazimov.com/ . His year of building is documented publicly and worth reading in full.

Every Sunday, Behind the Launch tells the story of one founder building something real. No press releases. No polished narratives. Just the honest version.

If someone you know is building something worth featuring — drop it in the comments. I read every one.

This week on SuperLaunch.io

Every week our SuperReviewers community tries new AI tools and shares honest feedback. Here’s what we reviewed this week.

Prodigent.ai ⭐ 4.2 / 5

A structured async knowledge layer that captures your team’s decisions, questions, and context in one place — so nothing gets lost in Slack threads. Best for product and engineering teams who want a dedicated space for decisions without the chaos of chat.
👉 superlaunch.io/product/Prodigent

ClankerView ⭐ 4.4 / 5

Sends AI personas to browse your website and returns a detailed structured UX critique in minutes. Frictionless onboarding, generous free credits, and report quality that goes well beyond a generic audit. Best for founders who want honest UX feedback before launch.
👉 superlaunch.io/product/ClankerView

DaedalMap ⭐ 4.0 / 5

A geographic query engine that lets you ask natural language questions over climate, demographic, and hazard datasets through an interactive map. Polished interface, zero signup friction to explore. Best for researchers, journalists, and urban planners working with geographic data.
👉 superlaunch.io/product/DaedalMap

Want your tool reviewed by our community? List it for free at superlaunch.io

The founder had just lost his tenth customer that month. He was frustrated — understandably so — but what caught Karanvir’s attention wasn’t the frustration. It was the honesty that followed.

All that information existed. It was sitting in Stripe, in Intercom, in his analytics tool. Nobody had connected it.

That conversation became the seed of ChurnAutopsy. But first, Karanvir had to ask the same question to every SaaS founder he could find.

The Exit Survey Lie

The answer he kept getting back was consistent: no, they didn’t really know why their customers were leaving.

It wasn’t that the data didn’t exist. It was that until recently, AI wasn’t capable enough to read all that fragmented context together and produce something genuinely useful. “That window opened,” Karanvir says simply, “and I walked through it.”

He looked at what founders were doing when a customer churned and found three camps. The most diligent ones were doing it manually — opening Stripe in one tab, Intercom in another, trying to piece together a story. Effective, but completely unsustainable past 20 or 30 churns a month. The majority were reading the exit survey and filing it. And the third group wasn’t doing anything at all — just absorbing the Stripe notification, feeling a moment of frustration, and moving on. The knowledge of why that customer left would simply disappear.

Version One

The first version of ChurnAutopsy was, by Karanvir’s own admission, almost embarrassingly simple. You uploaded a CSV of churned customers, and the AI read everything together and gave you a plain-English explanation of why each one left. No integrations. No automation. Just: here’s what it means.

What he didn’t expect was the reaction.

That sentence told him the core idea was right.

What changed the product significantly came from watching early users. Founders didn’t just want to understand one customer in isolation — they wanted to know if there was a pattern. When you look at 20 churned customers and find that 15 of them hit the same friction point in week two, that’s not a one-off. That’s a product problem you can actually fix. The pattern detection feature — now the most strategically valuable part of the product — was never in the original plan. It emerged entirely from watching what founders were trying to do that the product wasn’t letting them do.

The Real Culprit

After running hundreds of churned customers through the product, a pattern emerged so consistently that Karanvir gave it a name: silent onboarding failure.

The story goes like this. A customer signs up with genuine intent. They hit a point of confusion in the first two weeks — something not obvious enough, a feature that requires unexpected setup, an integration that doesn’t work the way they assumed. They open a support ticket, or sometimes don’t even do that. They log in a few more times, trying to make it work, doing so less frequently. They’re not angry. They haven’t decided to leave. They just quietly drift, until one day they look at their credit card statement and cancel.

This pattern shows up in roughly 65–70% of the churned customers ChurnAutopsy analyses. And almost none of them report it as the reason they left.

The case that crystallised this came from a B2B analytics founder who was convinced his churn was a pricing problem. He’d done the maths, looked at the exit surveys, talked to former customers, and landed on the conclusion that he was priced 20–30% too high. He was about to run a repricing exercise.

When they ran his last 10 churned customers through ChurnAutopsy, 8 of them had opened a support ticket in weeks two or three asking about a specific data integration. All 8 received a link to documentation. None of them ever successfully completed the integration. Their login frequency dropped sharply the week after that support ticket. They all cancelled within 60 days. And all of them selected “too expensive” on the way out.

The integration worked fine. It just required a step that wasn’t obvious. One confusing moment in week two was silently killing 80% of his churned customers, and none of them had consciously registered it as the reason they left.

He fixed the onboarding for that integration. He never changed his pricing.

The Silence Nobody Talks About

Building ChurnAutopsy wasn’t without its own difficult period. Karanvir is candid about the early months.

What kept him going was a deliberate decision to focus on conversations over metrics. Even when people weren’t ready to pay, he kept hearing versions of the same thing: this is a real problem, I just haven’t found the right solution yet. That consistent signal from people who understood the space was enough.

The moment that fully shifted it: a founder used a demo on his actual churned customer data and immediately said he’d been about to lower his prices to fix a churn problem that was actually caused by a confusing integration setup. He’d been ready to give away revenue to solve the wrong problem. One conversation made everything before it feel worth it.

There’s a harder truth Karanvir has come to about the early-stage building experience that he wishes someone had told him before he started.

“What nobody says publicly is that the silence is just how it works,” he says. “It’s not a sign you should pivot or give up. It’s the normal experience of building something that doesn’t have distribution yet. The founders who make it through this period aren’t the ones who avoided it. They’re the ones who found a way to keep showing up during it without requiring external validation to do so.”

What’s Next

In 12 months, Karanvir wants ChurnAutopsy to feel less like a tool and more like a team member — the one who’s always watching your retention data so you don’t have to. Connect Stripe once. Get an autopsy report within 60 seconds of every cancellation. Get a Monday morning digest of churn patterns and at-risk customers without ever logging in.

The target customer is the SaaS company in the £10K to £500K MRR range — too serious about retention to keep guessing, but not yet ready for Gainsight’s enterprise pricing and complexity.

The bar he’s set for himself is specific: a founder should be able to say they haven’t lost a customer they didn’t understand in six months. Not just knowing why customers leave — knowing early enough to do something about it.

And for any founder reading this who just lost their first five customers and has no idea why, his advice is equally direct.

Fix that one thing, he says. Don’t reprice. Don’t rebuild features. Don’t run win-back campaigns. Understand what specifically broke in the first month, and fix it.

The answer has been there all along. You just needed someone to connect the dots.

He was broke. His music career had stalled. He’d had a taste of freedom — travelling through Asia, working online — and now he was watching it slip away as his bank account drained. He needed to make money. And he needed to make it fast.

So he opened his laptop and told the internet he was going to build 12 startups in 12 months.

That person is Pieter Levels.

Who is Pieter Levels?

Born in Amsterdam in 1987, Pieter Levels studied music production at the Conservatorium van Amsterdam and worked as a techno music producer and DJ before becoming an entrepreneur. He taught himself to code in his late twenties — no computer science degree, no bootcamp, no formal training.

Today he is one of the most well-known indie makers in the world. He has built a $3.1M per year portfolio of bootstrapped products — including Nomad List, Remote OK, and PhotoAI — without funding, employees, or an office. He has over 500,000 followers on X where he shares everything — revenue numbers, failures, experiments, and ideas — in real time.

He is, more than almost anyone else, the person who made “one person, one laptop, real business” feel possible.

This is how he got there.

The challenge

The rules he set himself were brutal in their simplicity: one new product, every month, for 12 months. Ship it. Launch it. Move on. No exceptions.

He wasn’t trying to build 12 successful companies. He was trying to find out which ideas had life in them by putting them in front of real people as fast as possible. Most would fail. That was fine. Failure at speed costs almost nothing. Failure after six months of building costs everything.

Project number seven was a spreadsheet.

From spreadsheet to $3M

Within days, the spreadsheet had more traffic than he knew what to do with. People were sharing it, adding to it, arguing about which cities deserved their rankings. He had touched something real — a question thousands of people were quietly asking with no good answer.

Over the next months, the spreadsheet became a website. The website became a paid membership community. The community became the largest crowdsourced database of cities for remote workers in the world, with over 100,000 members.

He built all of it alone. His tech stack was vanilla PHP, jQuery, and SQLite — not React, not microservices, not the cutting-edge frameworks developers spend months debating. Just simple tools, used well, by one person who cared more about shipping than impressing.

Meanwhile he launched Remote OK — a remote jobs board built inside a single PHP file — which grew into the world’s largest platform of its kind. Then PhotoAI, an AI image generation tool, which became profitable within weeks of launch.

The real story behind the numbers

When people discover these numbers, the first instinct is to look for the trick. The famous audience that explains it. The lucky timing. The viral moment that made it easy.

But Pieter has been consistent about what actually happened: it took over 70 attempts, not 12. Years of shipping things that went nowhere. Nomad List wasn’t a stroke of genius — it was the seventh thing he tried in a year when he was running out of money and desperately needed something to work.

What separated him from the thousands of developers who tried the same 12-startups challenge and disappeared wasn’t talent or luck. It was one decision he made from the very beginning: to build everything in public.

He shared his revenue numbers openly. He wrote about what was working and what wasn’t. He made his entire business — the good months and the terrible ones — an open book on Twitter.

That transparency became his most powerful growth engine. Every honest post about a struggle built trust with people who would eventually become customers. Every revenue milestone he shared brought thousands of curious builders to discover his products. He didn’t build an audience to sell to. The audience found him because of how he built.

We tell one founder story every week. Some are famous. Most aren’t yet. All of them are honest.

Three things he did that most founders won’t

He shipped before it was ready. Nomad List was a spreadsheet. Remote OK was a PHP file. PhotoAI was an experiment. None of them were polished at launch. He put the simplest possible version in public and let real users pull it forward. The version in your head is not the version that succeeds — only the version people can actually use tells you anything true.

He built for himself first. Every product he’s built started as something he personally needed. That’s not a coincidence — it’s the source of the authenticity that made early users trust him. When you’re solving your own problem, you know exactly how good the solution needs to be.

He made transparency his competitive advantage. In a world where most founders only share the wins, Pieter shared everything. The struggles. The near-failures. The experiments that went nowhere. In doing so he built something most funded startups spend millions trying to manufacture: genuine trust.

Where he is now

He recently built an AI flight simulator in 30 minutes using Cursor. It went viral. He added a Stripe payment link for pixelated F-16 jets. It started making money within days.

The 12-startups challenge he started in 2014 has now been attempted by thousands of builders worldwide. Most don’t complete it. But the ones who get somewhere tend to share one thing in common — they stopped waiting for the right moment and started shipping.

Pieter Levels never waited. That’s still, a decade later, the whole story.

Pieter writes at levels.io and shares everything at @levelsio on X. His book MAKE — about bootstrapping startups without funding — is at makebook.io.

If someone you know is building something worth featuring — comment on this post. I read every comment. :)