The cold-start problem nobody warns you about: 8 things that actually moved the needle before we had users
Every indie founder gets told to "build in public" and "find your first 100 users." Nobody tells you what those first weeks actually feel like: you ship something you're proud of, you post it, and it lands in total silence. Not criticism. Not feedback. Silence.
The silence isn't a signal about your product. It's a structural problem. Every platform that could distribute you — Product Hunt, Reddit, Discord, X, Instagram, YouTube — decides how much reach to give you based partly on how people are already reacting. With nobody reacting yet, there's nothing to amplify.
I've spent the last while working on distribution for early-stage products across a lot of channels, and I've watched which cold-start tactics survive contact with reality. Here are eight, ordered roughly by how much they mattered.
1. Pick one channel and be genuinely present in it for a month
The single most common mistake I see is founders spreading across six platforms and being invisible on all of them.
Reach compounds within a channel, not across them. Thirty comments in one subreddit over a month makes you a recognisable name there. Five comments each across six platforms makes you nobody, six times over.
Pick the channel where your users already argue about the problem you solve. Spend a month contributing without mentioning your product at all. It feels agonisingly slow and it is the highest-return thing on this list.
2. Fix the thing people check before they decide you're real
When someone hears about your product, they look you up. What they find in the next ten seconds decides whether they try it.
That check is unforgiving and mostly non-verbal. An X account with 40 followers and no posts since March. A Discord with 12 members and no messages. A YouTube channel with three videos at double-digit views. None of that says "bad product" — it says "abandoned project," and people won't invest time in an abandoned project.
Before any launch, audit what a stranger sees. Concretely: does your most recent post have a plausible engagement rate for your follower count? A profile with 8,000 followers and four likes per post reads as bought or dead, and both are fatal. You can sanity-check the ratio with any engagement rate calculator — the number matters less than whether it's wildly out of line with your size.
3. Warm up the room before the launch, not during it
Product Hunt taught me this the expensive way.
A launch is a ~24-hour event with a feedback loop measured in minutes. Early velocity determines placement, placement determines traffic, and traffic determines the rest of your day. If you arrive at 00:01 with nobody who knows you exist, you spend the whole window climbing from zero — and the window closes before you finish.
Everything that decides a launch happens in the two weeks before it. Tell people individually. Not a broadcast post — actual messages to actual humans who'd care, asking them to look on the day. Thirty people who genuinely turn up beats a thousand-follower announcement into the void.
One more thing worth knowing, because it catches people out: anything that arrives in a sudden block on launch day tends to get stripped by the platform. Real interest arrives unevenly across the window. Engineering a spike at minute one looks exactly like the thing the platform is filtering for.
4. Community platforms have a different physics — use it
Reddit, Discord and Telegram behave nothing like feed platforms, and that's an advantage for a small product.
On a feed platform you're competing with everyone for a slot. In a community you're competing with whoever else is in that room right now, which might be eleven people. A thoughtful comment in a 4,000-member Discord reaches a meaningful fraction of the people who care about your category. The same effort on X reaches nobody.
The catch is that communities have a minimum viable size before they self-sustain. A Discord under about 50 members feels dead to anyone who joins, so nobody stays, so it stays under 50. The same floor exists on Telegram, where an empty channel reads as abandoned within seconds. Getting past that threshold — by any means, including inviting every person you've ever spoken to about the problem — is the whole game. Below it nothing works; above it, the room starts generating its own conversation.
5. Ship something useful that isn't your product
This is the tactic I'd most want to go back and start earlier.
Build one small free tool adjacent to what you sell. Not a lead magnet with a form in front of it — an actual tool that solves a small problem in ten seconds with no signup. A calculator. A checker. A generator.
Two things happen. People link to it, because a useful tool is one of the few things on the internet people still link to voluntarily. And it puts you in front of people at the moment they have the problem, which is a far better introduction than an ad.
We run a few ourselves. The one that gets used most takes a username and tells you how much of a following looks bought, which takes about four seconds and no signup — and it reaches more people than anything we've ever written about ourselves.
6. Be specific in public, not polished
"Building in public" has degraded into posting MRR screenshots. The version that works is narrower: post the specific thing you learned this week that cost you something to learn.
Nobody engages with "excited to share our progress!" People engage with "we lost three days to a caching bug that only appeared over 2MB, here's the exact symptom." That post finds the four people who've hit the same wall, and those four people remember you.
Specificity is also self-selecting. Vague posts attract vague attention; precise posts attract exactly the people your product is for.
7. Treat the first ten customers as research, not revenue
Your first ten paying customers are worth more as information than as money, and almost everyone under-mines them.
Talk to every one. Not a survey — a conversation, ideally live, where you ask what they were doing before they found you and what nearly stopped them buying. The near-misses are the valuable part. Every one is a sentence on your landing page you haven't written yet.
We learned more from a handful of customers who emailed us because checkout confused them than from any analytics dashboard. They were telling us, individually and for free, exactly which step was broken.
8. Decide deliberately how you feel about paid signals
I'll be upfront: I work in this space, so read this with that in mind.
There's an entire industry selling early engagement — followers, upvotes, members, views — and indie founders are one of its biggest customer bases. It's rarely discussed honestly on either side. The vendors pretend there's no downside; the purists pretend nobody does it, while a good fraction of the launches you admire had help.
What I'd actually say, having watched it closely:
It solves exactly one problem — the cold-start floor from #2 and #4. An empty room that looks abandoned. It doesn't create demand, it doesn't retain anyone, and it won't rescue a product people don't want. If your issue is "nobody wants this," this makes the graph look better and changes nothing.
Where it fails is expectation mismatch. People buy launch-day impact and discover the delivery is paced over days, because anything arriving in a block gets stripped by the platform. That's not vendors being slow, it's what safe delivery looks like — but almost nobody says so upfront, so the buyer feels cheated.
The honest version of the advice: if you use it, use it as a floor rather than a strategy. Get a community past the dead-room threshold, or a profile past the "abandoned" read, then let real activity carry it. If you're relying on it to look successful rather than to stop looking abandoned, you've already lost the thread.
Plenty of founders will disagree with the whole category, and that's a legitimate position. I'd just rather see the trade-off argued openly than pretended away.
What I'd tell myself at the start
Most cold-start advice is about tactics. The thing I actually got wrong was sequencing.
I tried to be visible everywhere before I was interesting anywhere. I launched before I'd told anyone. I built the product before I'd built anything anyone would link to. Every one of those is the right activity in the wrong order, and the wrong order costs months.
If I started again: one channel, one free tool, thirty real conversations, and a launch date set after all three exist — not before.
Happy to compare notes if you're in the middle of this. The first-hundred-users problem is one of the few where other founders' specifics are actually transferable, and most of what's written about it is too abstract to use.
