Part 1 of 2: How to earn attention from Reddit, Hacker News and Product Hunt when you have no ad budget
TL;DR
- In the AI era, ideas and even whole products get cloned in a weekend. Distribution is the part AI can't copy for you.
- A launch is not a day. By the time you hit "publish," about 80% of the work should already be done.
- One rule beats every tactic: give value before you take value.
- On Reddit, Hacker News and Product Hunt, patience and reputation beat polish and hype. Fast, promotional launches get removed, banned or ignored.
- You can build your first real audience for roughly zero dollars if you create something useful around your product before the product exists.
Your idea is not the moat. Distribution is.
Try this thought experiment. You have a good idea for an AI product. How long before someone else ships something nearly identical?
A year ago the honest answer was a few months. Today, with Claude Code, Cursor and a dozen other tools, it's closer to a long weekend. You can polish the model, sand down the UX, tune the prompts. Competitors can do all of that too, and they'll do it faster than you'd like.
What they can't copy overnight is who already knows you exist: the communities that trust you, the search results that surface you, the people who will say your name when someone asks "what should I use for this?" That's distribution, and it's still stubbornly human work.
I come from a growth background and have worked on go-to-market for several AI products: a music generator, an AI companion app, a financial-data platform, and a few others. I've done launches with money and without it. Some of them worked. Some of them got my accounts banned. This two-part series is the playbook I wish I'd had at the start, with the failures left in.
Part 1 is about the launch itself and the three communities most founders obsess over: Reddit, Hacker News and Product Hunt. Part 2 covers search, PR, creators and the art of making noise.
Launch day is the last 20%
When I picture "launching a product," my brain reaches for the standard toolkit: buy promotion, record a slick demo video, line up a wave of creators to post the same demo on the same day with the same script, and manufacture the feeling that this thing is blowing up.
That's a valid playbook, and I'll get to it. But it hides the most important fact about launches: the visible event is the tip of the iceberg.
Take Product Hunt. Whether your product climbs the daily ranking has very little to do with what you do that day. It depends on whether, months earlier, you quietly gathered people who care. We live in an attention economy where everyone is distracted by something else, so the traffic and noise you capture on launch day are a function of the groundwork underneath.
The same goes for offline. A great launch is rarely purely online. It leans on the founder friends you met at a conference, the developers you traded favors with, the investor who happily retweets your news. These threads connect to each other, and none of them can be pulled out of thin air the week before.
So the first reframe: stop asking "how do I launch?" and start asking "what can I build in the 90 days before the launch that makes launch day easy?"
The one rule: give value before you take value
If I could only give you one principle, it would be this: share the value, not your product.
In marketing terms: give first, collect later. Put something genuinely useful in front of people before you ask for anything. Don't expect anyone to use your product in return. They have alternatives, they owe you nothing, and they can smell an agenda from the first sentence.
I'll keep returning to this idea, because it shows up everywhere: in how you post on Reddit, how you write comments, how you build an audience, how you approach creators, even how you make friends at conferences. The tactics change by channel. The rule doesn't.
It also changes how fast you should move. Value-first marketing is slow by design. If you're in a hurry, you'll be tempted to skip straight to the ask, and the channels below are very good at punishing that.
Product Hunt: a ranking, not a strategy
Product Hunt is where most founders start, so let's be honest about what it has become.
The best example I know of a good Product Hunt launch is Gamma, the AI presentation tool. As I remember it, they spent a long time building up a large group of loyal early users before they launched, back when Product Hunt's credibility was still intact. On launch day they rallied those users, and the upvotes were real: actual customers vouching for a product they already liked. That's the model. The ranking was the result of earlier work, not a substitute for it.
Today the picture is murkier, for three reasons:
- It's commercial. Some well-known hunters will, for a fee, endorse your product or help push it up the list. Founders also blast every group chat they're in, begging friends to upvote. A #1 spot on the daily chart may mean the product is great. It may also mean the founder had the best hustle.
- Timing is brutal. If several products launch on the same day, you can get lost in the pile. Pick a quieter day.
- The return has dropped. My honest view is that, for most teams, Product Hunt is no longer a high-ROI launch channel. The effort to climb the chart often outweighs the customers it brings.
My suggestion: treat Product Hunt as an optional credibility badge, not your launch. If you already have a base of users who will show up and vouch for you organically, go for it, and choose a calm day. If you don't, spend those weeks building the base.
Hacker News: earn your seat, then be interesting
My team and I have launched on both Product Hunt and Hacker News, and I find Hacker News the more community-driven of the two. It's run by Y Combinator, and it is famously hard to game. If you're thinking about creating a stack of accounts to push a post up the front page, don't bother.
Three things I'd tell anyone planning a Hacker News post:
1. Start building your account months ahead. If you want to share an article or a launch there someday, begin now. Spend three or four months commenting thoughtfully on other people's posts and becoming a recognizable voice. You can't suddenly decide on launch week to do something promotional. The community leans technical and pure, and it reacts badly to anything that smells like marketing.
2. Reputation is what keeps a post alive. If you've made real friends in the community and respected regulars comment on your post, it stays visible far longer. Without that, it sinks quickly.
3. Interesting beats technical. You don't need to publish hardcore engineering to get noticed. When we were working on an AI music product, ChatGPT went down one day. We wrote a song about the outage and shared it. It caught fire on Hacker News, a small but very real win. The bonus came afterwards: a number of people who write technical blogs saw it and linked to our song from their own posts, which handed us several natural backlinks for free.
That last part is worth sitting with. We didn't set out to win backlinks. We made something people wanted to share, and the SEO benefit came as a by-product. That's value-first marketing in miniature.
Reddit: three plays that work (and the one that got my accounts banned)
If you have a small budget or none at all, Reddit is the channel I'd look at first, but only if you can start two to three months early to build up karma and a credible profile.
One warning before the tactics: don't outsource your posts to someone else's accounts. Only the founder really knows what the product feels like, and many "we'll post for you" accounts just generate generic AI-written content that readers sniff out immediately.
Play 1: Comparison posts, in small subreddits
Write a genuine comparison of your product and similar tools. Be fair. Cover the pros and cons of your competitors as well as your own, and write as a user of AI tools rather than as the company.
Then post it in small, niche subreddits with a few hundred to a few thousand members, ideally ones whose rules aren't strict about self-promotion. Avoid the giant communities. Their moderators have seen every flavor of disguised promotion and will remove it.
Why bother if the audience is tiny? Because the goal isn't Reddit traffic. It's two long-tail effects:
- SEO. Reddit ranks well on Google, often in the top five for searches that include a product name. A comparison thread can outrank the same comparison on your own blog.
- Credibility for people who are already curious. Someone hears about your product, isn't sure whether to download it, and searches "your product + reddit" to see what real people think. Having a fair, readable thread there helps.
The mistake people make is wanting both: a comparison post and thousands of eyeballs, so they post it in a huge subreddit and get it removed. Pick one goal.
Play 2: Comments, not posts
Posts are slow and risky. Comments are underrated.
- List the questions your target users ask, especially the ones tied to a real pain your product solves.
- Watch for new threads asking those questions, ideally within the last 24 to 48 hours, and the fresher the better.
- Reply as a helpful community member: "there are a few ways to approach this. First, ... second, ..." Leave out the link.
This only works if your profile looks like a real community member's and not a billboard. Participate in discussions that have nothing to do with your product.
Play 3: Posts that share value, not products
If you do want to post, make it beneficial to others first. In my experience, the posts that have done well for me, reaching tens of thousands of views and thousands of upvotes, all shared something useful and weren't ads.
Building an AI music product? Share your favorite playlists, artists or recent releases in the right music community. Building an AI travel product? Share routes. Weaving your product into a post in a natural way is tricky, so don't force it.
And here's the pattern that works: be consistent. Post the same kind of thing on a schedule, say one or two playlists a week. For the first four weeks, mention nothing about your product. Let people get used to you as the person who brings good music. Only when the community knows you can you say, "By the way, I'm building this, and it tries to solve that." By then it feels like a friend sharing news.
Reddit communities care deeply about their atmosphere. If people recognize you as someone who adds value, they upvote you and thank you in the comments. If not, you're gone.
The lesson is the one I started with: you can't extract value from a community before you've put some in.
Build the audience before the product
The cheapest launch I've ever run started months before there was a product to launch.
We were building a data platform, something like a Bloomberg-style terminal where people can explore all kinds of data. The question every founder in that position asks is: where do the first users come from?
My answer was to build a media publication in parallel with the product. While the platform was being developed, the publication used its data to write analysis: price trends, macroeconomic outlooks, that kind of thing. Readers found the articles useful enough to forward to friends, and they spread well in several markets. At the bottom of each piece, we invited readers into a private community (a Discord server, for readers outside China).
Then the loop did its work:
- Articles based on the product's data prove the data is valuable, without sounding like an ad.
- Community collects the readers who care, in a place we own.
- Beta comes next. Readers of the analysis naturally wonder whether they can see the raw data and draw their own conclusions. Now they can, with invite codes and referral mechanics that encourage a second wave of sharing.
- Official launch arrives last. By now the community has hundreds of engaged people who know the product and are happy to help you launch it.
The whole thing cost almost nothing. It worked because the content and the product were tightly connected: the articles were made from the product's data, so every reader was already a qualified prospect.
If you take one idea from this section, take this: find the valuable thing your product makes possible and give that away publicly, long before you ask for a signup.
A 90-day pre-launch checklist
- Months 3 to 2 before launch: Create or revive your Reddit and Hacker News accounts. Comment daily on threads where you can genuinely help. Pick two or three small subreddits to become a regular in.
- Months 2 to 1: Start a recurring, value-only post series in one community. Publish content built from your product's insight or data. Begin collecting an email list or private community.
- Month 1: Write your comparison posts. Open a small beta with invite codes. Quietly line up the supportive people (users, friends, fellow founders) who will show up on launch day.
- Launch week: Choose a quiet day if you use Product Hunt. Post where you've earned the right to. Reply to every comment.
- Always: Share value first. Never buy upvotes, farm karma or cross-vote your own accounts.
Closing thought
Everyone wants the shortcut: the viral post, the front-page ranking, the one tweet that changes everything. Sometimes shortcuts exist. But the launches that hold up are the ones that look like overnight successes and turn out to be slow, boring groundwork on the inside.
Give first. Be patient. Show up in the places where your users already are, as one of them.
In Part 2, I'll cover the other half of the playbook: starting SEO on day one with long-tail keywords, using PR to build authority (including with AI search), working with creators without wasting money, and what to make of the rising wave of attention-grabbing stunts, including the cautionary side of the story.
