In the first week of May, 3,913 products launched on Product Hunt.
Not that month. That week. And here’s the part that should stop you: the median product got one vote. More than a third got zero. Only 8.6 percent cleared ten votes. The top one percent of launches took a third of all the attention on the platform.
Almost every one of those founders did the work. They shipped something real, wrote the copy, made the graphics, picked the launch day. Then they walked into a room where four thousand other people were talking at the same time, and most of them were never heard at all.
That’s not a launch problem. That’s the defining condition of building right now.
The two cost curves split
Everything about making software got cheaper in the last two years. A founder with no design background ships a product that looks funded. Copy, landing pages, onboarding flows, support docs, all of it compressed from weeks to an afternoon. The cost of going from idea to working product fell off a cliff.
The cost of getting a stranger to notice went the other direction. Median SaaS customer acquisition cost has climbed roughly 260 percent over nine years, and the typical company now spends about two dollars to acquire one dollar of new annual recurring revenue, up 14 percent since 2023. At seed stage the median CAC sits around $1,248 per customer.
Read those two facts next to each other and the strategic picture gets obvious. The thing you’re good at got cheap. The thing you avoided got expensive. And because building is where the satisfying feedback lives, most founders keep pouring effort into the side of the equation that no longer differentiates them.
Your competitor can rebuild your product in a month. They cannot rebuild eight months of you showing up in the same place in front of the same people.
A launch is not a distribution plan
Here’s the pattern I see over and over. A founder spends four months building. Two weeks before launch they start thinking about distribution, which in practice means making a list: Product Hunt, a few directories, LinkedIn, maybe a Reddit post, an email to everyone they’ve ever met. Launch day arrives. There’s a spike. Signups, congratulations, a screenshot of the traffic graph.
Then it decays, because it was always going to. A launch is an event. Distribution is a system. Events give you a spike and a hangover. Systems give you a line that goes up slowly and doesn’t reset.
The reason the list approach fails isn’t that any single channel is bad. It’s that doing eight channels badly for two weeks teaches you nothing. You end up with eight data points, all of them noise, and no idea which one to double down on. Meanwhile the founder who picked one channel and stayed in it for a quarter knows exactly what a good week looks like, what their audience responds to, and what to do next Monday.
One channel, ninety days
The rule is exactly what it sounds like. Pick one channel. Commit ninety days. Don’t evaluate it before then.
Choosing is the part people overthink, so narrow it with three questions.
Where does your buyer already gather? Not where you’d like them to be. Where they actually are, in a form you can observe. If your customers are operations managers at construction firms, that is probably not X. It might be a trade association newsletter, a LinkedIn group, or a conference floor.
What can you do every week without dreading it? This is the underrated criterion. A channel you hate is a channel you’ll quit in week five, which is right before most of them start working. If writing drains you, don’t pick content. If cold outreach makes you sick, don’t build a plan that depends on fifty emails a day.
What compounds? Some channels leave something behind. A post that keeps getting found, a partner who keeps referring, a community where your name accumulates. Others reset to zero the moment you stop paying. Early on, when you can’t afford to rent attention forever, prefer the ones that stack.
Then pick one metric and only one. Not signups, traffic, impressions, and follower count. One. Usually it’s qualified conversations started, because at your stage that’s the thing that turns into revenue.
What working looks like at day 90
Not a big number. A repeatable one.
Ask whether you can explain why last week’s results happened. If you can say “these three posts did well because they answered the question buyers ask on the second call,” you’ve found something. If your best week was good and you have no idea why, you found luck, and luck doesn’t scale.
Ask whether week twelve took less effort than week one for the same result. That’s the signature of compounding. If it took more effort for the same result, the channel is fighting you.
If both answers are bad after ninety honest days, switch channels. That’s a real decision made on real evidence, and it’s worth infinitely more than the eight-channel scatter you’d have run instead.
The uncomfortable part
If distribution is now the scarce asset, then it deserves the treatment you currently give the product. Standing weekly time. A metric you actually watch. Iterations. The willingness to be bad at it for a while in public.
Most founders won’t do that, because building is comfortable and distribution feels like asking for something. Which is exactly why it’s still the advantage.
Four thousand products launched that week. The ones people remember weren’t the best built.
