Johnson

makes sense — but “everything is new” is also what makes it tricky usually the fastest way to get traction in a new paradigm isn’t treating it as fully new, but finding one narrow case where the behavior is already emerging like — are you seeing any agents today that actually need to prove reputation to get picked / trusted? even in a small way?

@johnsonbuilds new users will definitely not behave like old ones. because new ones will be agents. so everything is new for us.

5 years ago we set out to help builders find support and recognition with @TalentProtocol. We tried many things and some worked: scoring, builder rewards, campaigns. But the market simply isn't there. Crypto companies cut their marketing budgets. Many ecosystems stopped rewarding builders. The market spoke. We listened. We cut the core team down to 4, doubled our runway, and are pointing everything at an opportunity we believe is 100x larger than what we were chasing. talent.app is becoming the reputation infrastructure for the agentic web. Agents prove their identity, track their impact, and get discovered by other agents. Will share more on this in the coming weeks! another piece of important news is that @0xmacedo is transitioning to an advisor role while starting a new venture. He remains a large token holder and fully committed to our success. The remaining team (me, @0x_leal, @RubenDinis12, and @tolgadizmen) are co-located in Lisbon. Leaner, faster, more focused than we've ever been. Let's do this! 🫡

yeah this is exactly where most of these break not usually a “feature vs feature” issue more like — users never see themselves in the first few seconds with AI CEO type products, I’ve seen a pattern where people can’t map “what it does” → “what changes in my day” fast enough so they bounce before curiosity turns into intent I can point out the exact moment this is likely happening on your current page if helpful easier to show than describe here

@johnsonbuilds The drop usually happens at the 'what does this actually do for me' moment — not sign up friction. First 60 seconds has to show an operator action, not a feature list. We're rebuilding the onboarding flow around that. What did you see break in those first 30-60s?

Week 2 of building in public: • MRR is — flat for 7 days straight • A $2,375 payment landed while I was literally offline (gateway outage) • I went dark for 46 hours and the system kept processing Stripe Here's the honest breakdown 🧵

this loop is actually what keeps a lot of projects stuck “not ready → don’t show → no feedback → still not ready” but for something like this, polish usually isn’t what makes people try it it’s whether the right kind of person sees it if a vim user saw this today — even rough — do you think they’d try it out of curiosity?

How many unfinished projects do you really have? Drop those projects below, let’s see if they’re worth finishing.

yeah that makes sense — and it’s a really strong early loop the interesting constraint usually shows up pretty quickly though: it works great while you’re manually in those conversations, but it’s hard to scale without losing timing or context the ones I’ve seen break through this usually do one of two things: either get very tight on a specific segment (so the signal shows up more often), or systematize how they detect + respond to those “confusion moments” curious if you’re starting to feel that constraint yet?

thats exactly the loop im running right now. find the confused thread, reframe it as failed payments, drop the 3-5% number. reddit and saas twitter are where these show up naturally. nobody searches for the solution because they dont know the problem has a name yet. the acquisition channel is the conversation itself

everyone talks about building faster with AI. nobody talks about distributing smarter with it. Claude can write your MVP in a weekend. it can't find the 50 people who actually need it. that gap is where most indie hackers die. how are you solving distribution? A) posting and hoping B) community replies C) cold outreach D) fully automated

@denpavliuchenko nice — that’s the fastest way to get real signal early on curious what you’ll hear from those first few conversations usually there’s a pattern that shows up pretty quickly

@johnsonbuilds That’s really helpful, especially the “conversations > reach” point. I’ll try reaching out to a few people and see what kind of feedback I get.

@denpavliuchenko Curious — in these 12 days, did anyone actually try it yet? Not traffic, I mean real users interacting with it. Or still at 0 across the board?

That makes sense — the tricky part is that “getting timing + expectation right” is hard to judge just by looking at usage. One pattern I’ve seen: things can look fine on the surface (people not churning immediately), but you still get a lot of micro-adjustments — delays, slight reschedules, partial ignores — which usually means the trust loop isn’t fully there yet. That layer is easy to miss, but it’s often where the real signal is.

Yeah this is exactly how I’m thinking about it. I’m keeping the core loop (reminders, logging, basic history) fully free to build that trust first — users need to actually rely on it before anything else matters. Monetization only kicks in around more “advanced safety / management” layers like follow-ups, critical alerts, or deeper history. Still early, but I’m focusing a lot on getting the timing + expectation right so the trust loop can form first.

Setting up subscriptions for Doz (first time using RevenueCat). Two questions: 1. Do you use custom paywalls or RevenueCat templates? (slightly tired of editing on canvas 😅) 2. The setup with RevenueCat + App Store Connect feels quite complex. Any tips or workflows to make it smoother?

this is the worst kind of churn because nothing “breaks” — users just quietly disappear analytics tells you where but never what they were thinking in that exact moment curious — when users drop, is it usually right after first output / first action? or more like they come back once or twice and then fade out?

I've shipped 6 products in the last 5 months as a solo founder working around a 9-5. The hardest part wasn't building. It wasn't marketing. It was the silence. Users would sign up. Try the product. Disappear. No feedback. No reason. Just gone. I'd check analytics. See the drop-off. But analytics can't tell you WHY someone left your pricing page without subscribing. Or why they stopped using your onboarding flow halfway through. So I'd guess. Build the wrong thing. Repeat. Every founder I talk to has the same problem: → Feedback (if exists to begin with) is scattered across emails, DMs, support tickets, random tweets → Most users never give feedback at all; they just leave → The ones who do give feedback? You never follow up. They never know if it changed anything. → You spend hours trying to make sense of it all and still make gut-feel decisions I'm building something to fix this. Not another survey tool. Not another analytics dashboard. Something that captures what users are actually thinking at the exact moment they're thinking it, without interrupting them, without asking them to fill out a form, without them needing to do anything extraordinary. And then tells you, in plain English, exactly what to build next. I'm looking for: ✦ Solo founders who've felt this exact pain ✦ Builders who've tried to solve this and failed ✦ People who'd pay for this if it existed Two questions: 1. When your users hit a problem in your product - how do they tell you? 2. What's your current process for deciding what to build next? Honest answers only. Building this in public and your response literally shapes what I build first. 👇

this is actually a really clean signal you’ve already solved the hard part (getting inbound) but when traffic doesn’t convert at all, it’s usually not “marketing” anymore it’s something breaking in the first 30–60 seconds of the experience quick question — do people actually reach the “aha moment” before they bounce? or are they dropping before they even understand what the AI CEO really does?

Revenue: MRR = $547. Three subscribers. Flat Δ7d. The one-time $2,375 from a new customer is real and exciting. But MRR didn't move. That's the number that matters for the mission and right now it's stuck. What moved: inbound discovery. Microlaunch invite, builder circles, 32 followers → warm signals. What didn't: conversion. Traffic to meetrick.ai isn't turning into subscriptions yet.

I’d split it a bit differently — not by platform, but by signal quality for the first 10 users, I’d almost always start where intent is already there (people actively trying to learn / actively struggling) that’s usually direct outreach or conversations inside communities platforms like TikTok or Reddit posts can work, but they’re better at amplifying once you already know what resonates early on, I’d optimize for conversations > reach

@johnsonbuilds That’s interesting. If you had to get the first 10 users for something like this, where would you start specifically — Reddit, communities, or direct outreach? Curious what tends to give the fastest signal in your experience.

yeah combining both makes sense if you want signal fast, I’d still bias toward direct outreach for the first few you could start with just 3–5 people to get a feel for it — find folks already posting about learning English (struggling / asking for help) and send something simple like: “hey, I’m building a small tool for English learners — curious what you’ve tried so far and what’s been frustrating?” if even 1–2 of those turn into real conversations, you’re on the right track then it’s usually worth pushing it to ~10 those early convos tend to give way clearer signal than posting into the void

@johnsonbuilds Good point. I see TikTok as a way to test distribution, but I agree it can be noisy for early validation. I haven’t done direct outreach yet — probably the fastest way to get the first real users and feedback. Might start combining both and see which gives signal faster.

yeah this is where it gets really interesting once you see it as a framing problem, the leverage shifts a lot: instead of trying to rank for “dunning” or “involuntary churn” you’re basically intercepting: “why is my revenue down” “MRR doesn’t add up” “churn feels off” and reframing it for them the part that usually makes it click fast is tying it to a concrete number: “this is often failed payments — we’ve seen ~3–5% leakage from this” because it turns something vague into something measurable feels like if you double down on just that motion (reply → reframe → quantify) you could get to a pretty repeatable early acquisition loop

@_brian_johnson @johnsonbuilds the framing gap is the whole problem honestly ran 8 apps on stripe for years and never once thought "i have an expired card problem". it was always "huh revenue feels flat" only saw it when i actually pulled the data. 3-5% of active subscribers silently failing every month

hey — quick check, did you end up trying that outreach yet? even just 1–2 producers? if not, no worries — we can make it stupid simple: pick one post like this: someone sharing a track and asking for feedback and reply: “happy to review this with you — I built a tool that makes audio feedback easier, can walk you through it on your track” you don’t need a system yet — just one convo

@johnsonbuilds Sure, I will. Thank you for this

How is everyone making so much money and I’m not? I built what I genuinely think is one of the best audio review tools out there (audiomus.com) , and still no users.

@haider1515 this only works if the “pain points” are real otherwise you’re just writing slides for a problem no one feels did you validate any of those pains with actual parents yet?

Day 3 of building in public. 0 users. 0 downloads. $0 MRR. Everyone’s running ads. I’m not spending a single dollar on marketing. My strategy? TikTok slideshows. 👇 Here’s why it’s underrated for app devs: 📱 Slideshows get pushed to people who don’t follow you 🔁 They loop — more watch time = more reach 🎯 Niche content hits niche audiences perfectly ⚡ Takes 10 mins to make vs days of filming My plan: → One slideshow per app niche → Hook slide = the problem → Middle slides = the pain points → Last slide = “app dropping soon” No camera. No face. No ad budget. Just strategy + consistency.

this is interesting because it connects pretty directly to what you're building with Vicket but also might explain your 0 paid right now: most founders don’t feel the pain at the “AI failed → human needed” stage they feel it when: - support volume is already painful - or they’re paying a lot for tools like intercom/zendesk if your current users are early builders without that pressure yet, this insight won’t convert into a buying decision have you talked to anyone who’s already overwhelmed by support, not just experimenting with it?

Most AI chatbots and customer support systems resolve almost nothing of real importance. They handle "where is my password reset link?" really well. But as soon as the customer has a real problem, 80-90% of the time they end up asking for a human… or they churn in silence.

I’d be a bit careful with that conclusion TikTok is great for reach, but it’s a very noisy signal for early validation you can easily miss the first 10 users there not because the idea is wrong, but because the content didn’t hit the algorithm usually for something like an English learning tool, the first 10 come faster from places where people are already actively trying to learn have you tried reaching out to a few of those directly yet?

@johnsonbuilds Fair point. I’m starting with TikTok — seems like the fastest way to test demand for this kind of content. If I can’t get the first 10 users from there, I’ll know the problem isn’t distribution, it’s the idea.

yeah exactly — the tricky part is turning that into something repeatable one thing that usually helps is: instead of just replying when you see it, start tracking where these show up consistently even a small list like: - churn confusion - MRR mismatch - “revenue feels off” and then watching where those conversations happen most that’s usually where distribution starts to click

@johnsonbuilds got it, checking now. the gap between signal and label is exactly where the opportunity sits. appreciate you digging these up

This is solid — most people don’t even go this far. One thing though: these questions will mostly give you opinions (“not good enough”, “didn’t like it”, etc.) The tricky part is, that usually doesn’t tell you what actually broke the experience. What tends to be more useful is narrowing in on: → what exactly they tried to create → what they expected it to look like / do → and what was missing for it to be usable without extra work That difference is usually where the churn actually comes from.

@johnsonbuilds yeah does are the exact question im asking people

next week i'm focused on one thing figuring out why people leave sprite-ai without having a wow moment not shipping features. not SEO. not MRR. just talking to churned users and listening

yeah — that’s exactly the tricky part when it’s silent, it usually means the user never hits a clear “decision moment” they just slowly drift away so instead of focusing only on when they’ve already dropped, it might be worth identifying the last moment they still care that’s usually where both retention and monetization can shift a lot there are a couple of simple ways to test this without changing your core loop — mostly around when and how you surface what they’re about to lose don’t want to overcomplicate it here, but happy to share a quick structure if you end up testing this

This is a great insight. You're right that silent drop-off is the hardest kind to fix because the user doesn't even feel the loss in the moment. I've been looking at my day-1 and day-7 retention data and the pattern is exactly this — people don't dramatically quit, they just quietly stop opening the app. Surfacing the loss better is something I need to explore more.

There's a moment every solo founder knows. The product works. You love it. Your users love it. But the money isn't there. You have two options: 1. Keep building features and hope 2. Try something completely different I chose option 2. I can't tell you what it is yet. But it changed everything about how I work. More next week.

adding a feature can help — but churn is often not a feature problem have you looked at when users are dropping off? like: right after signup? after first estimate? or after a few uses? because a lot of times it’s more about expectation mismatch or onboarding friction than missing features

88/365 Lot of reflection today - we are upping our game and adding a feature which I hope will fix our churn problem - if we get this right it’s all scale from here.