Nobody Told Me About Growth
I'm a builder. I thought if the product was good enough, people would find it. I was wrong.
I love building things. It’s what I do. I see a problem, I get excited, I start coding, and a few weeks later there’s something real. A working product, a demo, something I can show people. That part I’m good at.
The part I was terrible at was everything that comes after.
My routine used to look like this: finish building something, post it on Twitter, maybe put it on Product Hunt, drop it in a few Discord servers. Then wait. Sometimes a few people would try it. Sometimes nobody would. I’d tell myself the product just wasn’t good enough yet, go back to building, add more features, try again. Same result.
I never questioned the process. I just assumed that’s how it works. You build something great, and the world eventually notices.
That’s the biggest lie I believed.
Before I actually started learning about this, the whole world of “growth” felt incredibly overwhelming to me. People talk about growth hacking, influencer marketing, SEO, paid acquisition, PLG, ABM, demand generation, performance marketing, content marketing, affiliate programs. Founders go on podcasts, post Twitter threads, do Product Hunt launches. I’d see all of this happening and think: these are all different things and I don’t understand any of them.
Then I started talking to people who actually do this for a living, and the first thing I realized is that it’s all much simpler than it sounds.
All of it is just distribution. Getting your product in front of the people who need it. The reason it sounds overwhelming is because there are dozens of channels you can use to do that, and people talk about each channel like it’s its own discipline with its own jargon. But once you see the full picture, it clicks. Growth isn’t some mysterious dark art. It’s a set of channels, and your job is to figure out which ones work for your product.
That reframing alone made the whole thing feel less scary.
The map
So what are the actual channels? One conversation that helped me see the full picture was an interview with Chang Chen, founder of Hockey Stick, on a YouTube channel called 课代表立正. Her team has worked with 100+ AI companies in Silicon Valley on go-to-market. She walked through over 20 channels and how they map to different types of products. Here’s what I pieced together from that conversation and others:
Organic channels don’t cost money per click but cost time and effort. SEO, where you create content that ranks on Google. Programmatic SEO, where you generate pages at scale using your product’s own data or capabilities. Organic social content on Twitter, LinkedIn, TikTok. Reddit and community forums. Product Hunt launches. Blogging. These are channels most builders default to because they feel free. But they’re not free. They cost a ton of time, and they only work for certain types of products.
Paid channels are where you spend money to get attention. Google Ads, Facebook Ads, Instagram, Pinterest, TikTok ads. Influencer and creator marketing. PR and media coverage. Billboards and offline ads.
Sales-driven channels are for higher-priced products. Founder-led sales early on, where the founder personally closes deals. Then building a sales team with inbound (PR, events, conference sponsorship to generate demand) and outbound (SDRs doing cold emails and cold calls to find and qualify leads). ABM (account-based marketing) for targeting large enterprises, where you customize your pitch for each stakeholder in the company: the decision maker, procurement, finance, IT. Each person needs different marketing material. You’re essentially running a campaign per company, not per segment.
Community and referral overlaps with everything else. Building a community on Slack or Discord. Running hackathons that turn participants into advocates. Word of mouth. PLG (product-led growth), where the product itself is designed to spread from one user to the next. And affiliate marketing, which I used to think was simple but actually comes in three distinct types: your existing users referring friends for a revenue share, review sites and blogs with affiliate links, and creators on long-term contracts with a base fee plus commission.
Partnerships means working with companies that share your target audience. Upstream/downstream partnerships. User-exchange deals. Conference co-hosting. Integration partnerships with complementary products.
That’s the landscape. Every growth strategy you’ve ever heard of fits somewhere in there. When someone says “growth hacking,” they’re usually talking about finding creative ways to use these channels. When a founder goes on a podcast, that’s demand generation. When you see a viral tweet about a product, that’s either organic content or influencer marketing. It all maps back to this list.
The one concept that changed everything
But here’s the thing that actually matters: most of these channels won’t work for your product. Not because they’re bad channels, but because different products need different channels. This is the concept that rewired my brain: product-channel fit.
I always thought about it as: build the product, then find the right channel to distribute it. Like channel selection is a downstream decision you make after the product exists. It’s not. The channel shapes the product.
If your product is cheap and self-serve, you want PLG channels: SEO, community, referral, influencer. Users can pull out a credit card and just buy it. If your product costs $100K+ per deal, you need founder-led sales, SDR teams, ABM, partnerships. Those buyers don’t discover tools through Reddit posts. They find them through sales conversations, partner referrals, and conference demos.
Take Statsig, an enterprise analytics tool. They tried influencer marketing, Product Hunt, Reddit, SEO. All the PLG playbooks. None of them worked. Not because they’re bad channels, but because Statsig sells $100K+ contracts to engineering teams. The founder came from Facebook and Microsoft, both consumer product companies. The whole team had a consumer product mindset. For the first eight months they didn’t make a single dollar. Every three months the team would say “we need to do more PLG.” It never worked because the product was too complex, too expensive, and sat too deep in the development stack for self-serve discovery. Only after they hired a sales team did things start to move.
On the other side, consider Otter.ai. Instead of buying the same Google ads as every competitor, they used their own transcription technology to live-caption conferences at TechCrunch and WebSummit. Attendees could follow along with real-time subtitles on big screens, and all that transcribed content became SEO pages with backlinks. They invented a growth channel that only their product could power. No competitor could copy it because the channel was built on their core technology.
The lesson: before you pick a channel, you need to know what kind of product you are. The answers to “what’s my price point?” and “where does my user spend time?” eliminate 80% of the channel list. What’s left is what you should focus on. And if you can find a channel that only your product can own, that’s the real unlock.
The traps
Once I understood the map and product-channel fit, the next thing I started seeing everywhere were the mistakes. Not just other people’s mistakes. Mistakes I would absolutely have made. Some of them I already had.
”Just get some influencers and press coverage”
This was my default assumption. If I could just get a popular creator to mention my product, or get featured in a big publication, that would be the breakthrough. It sounds reasonable. It’s also mostly wrong.
Start with influencers. The pricing ranges wildly: a Twitter post might cost $50 to $100, a dedicated YouTube video from a major creator can cost over $100,000. Most AI companies pay somewhere in between. A 60-90 second integration in a YouTube video runs in the tens of thousands. In hot niches like vibe coding, the best creators are raising prices every quarter. Some have doubled their rates in a single quarter. They have waitlists. They add surcharges for crowded categories.
But the real surprise is the math. The viral rate on influencer posts is about 10%. Nine out of ten posts don’t blow up. That’s normal. The whole economic model depends on the 10% that go viral carrying the ROI for the entire campaign. If you test with 3 influencers, see no results, and conclude that influencer marketing doesn’t work, you haven’t run enough volume for the math to kick in.
And the posts that actually go viral are almost never hard ads. They’re stories. They ride a trending topic. They feel organic. One example: a creator pitched an idea to tie their sponsored post to a recent AWS outage. The brand’s team thought it was risky. They went with it anyway. It blew up that same day. If they’d insisted on a scripted product demo instead, it would have gotten a fraction of the reach. The lesson is counterintuitive: give creators the product, explain what it does, and let them figure out how to make it interesting. When you force a script on them, you’re replacing their judgment with yours, and yours is worse for their platform.
Also, don’t copy your competitors’ playbook. Their creator was half the price six months ago. The topic was hotter then. The audience hadn’t seen ten other products pitch the same thing yet. The same creator, same format might give you 1/10th the ROI today. The window closed.
Tracking is messy too. Even on YouTube, you can only track about 60% of conversions through links. Twitter is about 50%. TikTok and Instagram are even worse. A simple “how did you hear about us?” question during onboarding goes a long way. Some brands refuse to let creators use tracking links at all because a UTM link instantly signals “this is an ad” to the audience.
Now press coverage. I used to think getting featured in Forbes or the Wall Street Journal would be a game changer. It’s not. Even top publications drive surprisingly little traffic compared to viral social content. Press is useful as a credibility signal. “As seen in Forbes” looks good on your landing page. But as an actual user acquisition channel, it’s limited. Some media you can buy your way into (Business Insider accepts paid placements), but the publications that actually move the needle, like TechCrunch, can’t be bought. The most reliable way to get TechCrunch coverage is to have a funding round. If you’re an early founder counting on PR to drive growth, you can basically let that go.
”Build for everyone, figure out the audience later”
This is the trap that scares me the most, because it’s the one I would fall into naturally.
There was a company that built an AI version of Zapier, where you could describe a workflow in plain language and it would build the whole automation for you. This was in 2023, and they were one of the first to do it. The marketing worked incredibly well. They ran viral influencer campaigns, some posts getting millions of impressions for under $100. In less than six months they had close to 200K users.
But the product was too general. Marketing teams could use it. Sales teams could use it. Operations teams could use it. Everyone could use it, but nobody had a great experience, because the product was trying to serve every use case at once. Users didn’t know what to optimize for. The engineering team got pulled in every direction with conflicting feature requests from completely different types of users. They couldn’t prioritize anything.
The growth numbers looked great on paper. Active users, new signups, impressions. But when you talked to actual users, the experience wasn’t good. The company eventually narrowed down to focus on just sales and marketing teams, and everything improved: user experience, engineering focus, conversion, retention. But they’d lost time and momentum.
This was a product problem, not a marketing problem. The marketing was excellent. The product hadn’t figured out who it was for. This is the kind of mistake a builder like me would absolutely make: build something general, market it to everyone, and wonder why nobody sticks around.
”PLG is always the right call”
This one hit home. Because if I’m being honest, one reason PLG appeals to me is that it means I don’t have to do sales. Just make the product good enough that it sells itself. Let users sign up, try it, love it, pay for it. No cold calls, no sales team, no awkward pitch meetings.
But that’s not a strategy. That’s avoidance.
If your product is genuinely PLG-able, it needs a fast aha moment. A developer should be able to connect an API or install an SDK and see value quickly. DataDog is expensive but PLG-able because developers can start using it in minutes. If your product requires meetings, setup help, and procurement approval before anyone gets value, that’s a sales product. Be honest about which one you are.
And even when PLG works, it has a ceiling. This was counterintuitive to me. PLG products targeting prosumers and small businesses have high churn rates. Individual users cancel, downgrade, switch to competitors. In the early days this doesn’t matter because new signups vastly outpace cancellations. But as your revenue grows, the absolute number of churning users grows too. At tens of millions in ARR, the churn pressure becomes real. You’re running harder and harder just to stay in place.
That’s why Notion added enterprise sales. Slack added enterprise sales. Gamma recently launched B2B pricing. These aren’t failures of PLG. It’s the natural evolution. B2B contracts (one to three year deals) have much higher revenue quality. The growth is slower but stickier. The sales cycles are long and exhausting, but once a company signs, they rarely leave. PLG is a great starting strategy for the right product, but it’s not the end state. At some point, almost every company needs to layer on some form of sales to keep growing.
Designing growth in, not bolting it on
After learning about all the traps, I started thinking about what you can actually do proactively. And the most powerful idea I came across was the growth loop.
The idea is simple: your product should have a built-in mechanism where existing users naturally bring in new users. Not through a referral program you slap on at the end, but through the core product experience.
Gamma, the AI presentation tool, is a great example. People create presentations with Gamma and share them with colleagues and clients. Those people see the Gamma branding, experience the product in action, and some of them sign up. The product is the distribution. No ad spend required for that loop to work.
When evaluating a company, the first question to ask is: what’s your growth loop? What are the specific numbers at each step? How many existing users bring in new users? Which users share and which ones don’t? If you can’t answer these questions, there’s probably a ton of untapped potential. If you haven’t even thought about it, the product simply wasn’t designed for growth. Reforge is apparently the OG of this concept and goes deep into the design and measurement of growth loops.
One nuance for AI products specifically: the aha moment tends to be faster because many AI products are conversational. Users don’t need to click through onboarding flows or set up configurations. They type something, get a result, and immediately understand the value. But the core metrics (retention, conversion, referral) haven’t changed. People still need to write, make videos, collaborate, sell. AI just changes how they interact with the tool, not what they’re trying to accomplish.
And then there’s timing. Some of the biggest growth stories simply can’t be replicated, even if you copy every single thing they did. Lovable exploded in the vibe coding space. But if you launched an identical product today with an identical playbook, you wouldn’t get the same result. The market was less crowded then. The topic was fresher. The creators who promoted it were cheaper and had more novelty to offer their audiences. Same with Gamma. There are plenty of AI presentation tools now. None of them can recreate Gamma’s growth trajectory even with the same tactics.
There’s always an element of timing, of being in the right place at the right moment with the right product. That’s not something you can engineer. But what you can engineer is being ready when your moment comes: having the right channels identified, the right product-channel fit figured out, and the right growth loops designed. So when the window opens, you’re not scrambling to figure out the basics.
I want to be clear: I’m not a growth expert now. I talked to a few people, watched some interviews, and started piecing things together. But even this level of understanding has already changed how I approach my next project. Before I write a single line of code, I’m going to ask myself:
Who is this for? Where do they spend time? Is this PLG or sales-led? Which channels actually make sense? How does this product naturally spread from one user to the next?
These aren’t marketing questions. They’re product questions. And I wish someone had told me that two years ago.
I still love building. That hasn’t changed. But I used to think the product was the whole thing. Build something great and the rest takes care of itself. Now I think the product is maybe half. The other half is the path to the product. The best builders don’t treat those as separate problems. They design both at the same time.
I wrote this for anyone who makes things and has no idea why nobody’s using them. The whole world of growth felt overwhelming to me until I realized it’s not magic. It’s a map. There are channels, there are rules for which channels fit which products, and there’s a science to making them work. Once you see the map, it stops being scary and starts being a problem you can solve.
The product isn’t done when it works. It’s done when people can find it.