Distribution Secrets: What Actually Works
Published a month ago
Hey {{ subscriber.first_name | strip | default: "there" }}, After a year of building out my app portfolio, I’ve become convinced of one thing: building apps is easy now. With AI, anyone can ship something decent. The real game, the thing that actually separates the people who make it from the people who don’t, is distribution. In this email I want to walk you through how I think about it, the channels worth testing, and the exact system I use to run distribution experiments across my apps. Stick with me to the end and I’ll share the real numbers from what’s working for me right now (and what isn’t). Sponsored by Amazon Developer Your mobile skills already work on the big screenSpeaking of underserved channels: Fire TV is a platform built for the frameworks you already use, and almost none of us have shipped to it yet. The friction used to be the TV-specific glue work, D-pad navigation, focus management, and performance debugging. Amazon’s new Devices Builder Tools for AI handles that layer for you. Run one command and your coding agent gets Fire TV expertise built in, so it knows the best practices before you write a line of code. Alright, back to it. Here’s the trap I see devs fall into all the time: they spend months building something great, polish every screen, ship it… and then nothing happens. Nobody shows up. Not because the app is bad, but because nobody knows it exists. There are tons of genuinely good apps that die quietly for exactly this reason. So I want to flip the order for you. Start with distributionBefore building, ask yourself: where are my users going to come from? Make a few guesses, then test them. And here’s the key part, you don’t need an app to test this. All you’re checking is whether you can grab the attention of the right people. That attention can look like engagement on social media, newsletter signups, blog readers, freebie downloads, a waitlist, whatever. If you can’t get attention without an app, building one won’t magically fix that. This way you validate the idea in weeks instead of burning months on something nobody wanted. This is actually the big lesson from Traction (Gabriel Weinberg & Justin Mares), one of the books that shaped how I think about this. Their argument is blunt: most startups don’t fail because the product is bad, they fail because they neglected distribution. So they suggest spending about half your time on traction from day one, running in parallel with building, not after. The channels worth knowingPaid is the fastest to test:
Free takes longer but compounds:
Programmatic, distribution through code itself:
How I actually run experimentsThis is the part that changed everything for me, and it comes straight out of Traction. The authors call it “Bullseye”: the idea that at any moment, one single channel tends to drive most of your growth, and your job is to find it. I run it as three columns: What’s Possible, What’s Probable, What’s Working. What’s Possible, brain-dump every distribution idea you can think of. Don’t filter. Don’t let “that’s too expensive” or “I don’t know how to do that” stop you. Traction makes the point that founders dismiss channels too early based on gut feeling, when the data often says otherwise. So write down everything. What’s Probable, score them and move your best 2-3 bets here. Pick things you’re already good at, or things you’ve seen working for competitors. These are the ones you’ll actually run cheap, quick tests on. Now design those experiments. Time-box them. Some channels are slow to start (SEO especially), so anywhere from a couple weeks to a couple months is fair, and don’t spend more than $500 testing one. Track the numbers: impressions, conversion rate (impressions to installs), and the big one, CAC, how much each install actually costs you. What’s Working, if something clearly beats the rest, move it here and double down. Milk it. Improve conversions, push more output, raise the quality. There’s always more room. Too many people spread themselves thin right when they’ve found something that works. Traction’s whole thesis is to go all-in on that one channel once you’ve found it. And if a channel isn’t working? Kill it. Even if it works for everyone else. Knowing what to cut is just as important as knowing what to double down on. You can always come back to it later. Where I’m at right nowFull honesty, here’s my real scorecard. TikTok organic is the most promising thing in my Probable column. We’ve generated over 10M views with organic content and grown a couple of accounts to 2k followers in less than 3 months. What we’re working on now is the conversion from those views into installs, which honestly isn’t great yet. That’s the current focus. I’m also testing Apple Search Ads, and this is where the math gets real. The results look okay-ish, but the unit economics of most of my apps just aren’t there yet. Here’s what I mean: the LTV (what an average user spends after installing) is low for most of my apps, around $0.10 to $0.75. Meanwhile CAC on Apple ads runs $3-10 per install. You can’t pay $5 to acquire a user worth $0.50. So I’m working both fronts at once: pushing LTV up, and experimenting to bring CAC down. I do have one app with a better LTV, around $2, which makes it a real candidate for paid ads. I’m starting that experiment as I write this. Stay tuned for updates. That’s the whole game. Keep building, sure. But spend your real energy figuring out how people find you. Talk soon, Vadim P.S. Hit reply and let me know what have you experimented with and what’s working for you. |