One month in. Still no ads. Still no growth campaign. Just a daily question, a song, and people who keep coming back.
I want to write this one differently.
Not because the story has changed, but because we have something we didn't have a month ago: evidence. Real numbers from real people making real decisions about whether this is worth their time. And I think the honest thing to do with evidence is to say it plainly. What's working, what isn't, and what it means for where we're going.
So here's the month, measured.
The Signal
521 posts across 25 days. No dead days. Not one.
That's roughly one post per active user per day, sustained across an entire month with no incentives, no streaks dangling in front of people, no push notification campaigns nudging anyone back. People are just showing up. Every day since launch, someone opened the app, read the question, and decided it was worth answering honestly.
That's the first thing I want to say about this month. Not the number. The pattern behind it. A spike is easy to explain. A floor that holds for 25 consecutive days is something else.
Week-over-week retention has held between 79% and 90%. Week 1 to Week 2: 90%. Week 2 to Week 3: 79%. Week 3 to Week 4: 85%.
Retention is the metric social products can't fake. Growth can come from a lucky mention, a well-timed post, a moment of visibility you didn't earn. Retention comes from the product being worth returning to. It comes from someone sitting down on a Friday night, without any prompt from us, and deciding this is where they want to spend a few minutes of their honesty.
Two users have posted every single day since launch. Not a streak they're protecting. They were already posting every day in beta. One of them has a 34-day streak. That's not a power user. That's a ritual that has become identity. I don't take that lightly.
What's Growing Underneath
2,373 likes given. Not accumulated. Given. This community likes each other's posts, genuinely and often. But what I care about more than the total is the direction: average likes per post grew from 2.5 in beta to 6.5 last week. That's a 2.6x increase in the quality of engagement per post, as the community has gotten to know itself.
People aren't scrolling. They're responding.
440 comments across 521 posts. Nearly one comment per post. The top posts this week pulled 8 to 11 comments. Real conversations, not reactions. People are reading each other's answers and writing something back. That's what community actually looks like before it gets loud.
The community poll has now run for two weeks. 21 questions submitted. 67 votes cast. Members aren't just answering what we ask. They're writing the questions themselves, voting on what gets asked next, shaping what this space sounds like. That's a different kind of engagement. That's ownership starting to form.
Where It Breaks
I'm going to be direct about this because it's the most important part.
37% activation rate. Of 70 people who signed up in March, 44 never posted once. Not once.
That's the hardest number in this post to write, and it's also the most honest. Circle retains almost everyone who posts. The product does its job once someone is in. But the gap between downloading the app and making that first post is wider than it should be. People arrive, they read the question, and something stops them.
I think I know what it is. The questions are real. They ask for something real in return. Vulnerability isn't always accessible on a Wednesday when you're tired and haven't thought about your day yet. We designed a space for depth and then didn't make the entrance easy enough for everyone. That's on us.
We're not going to make the questions shallower. But we can make the door wider. That's the work.
The question is the product. I say this because the data says it first.
Mid-month, March 15 through 22, we averaged 12.4 posts per day. The week before and the week after: 18+ per day. The difference wasn't the community. It was the questions. Abstract phrasing, questions that left too much open and required too much work before the feeling could land, pulled as low as 10 responses in a day.
The format that works is specific and visual. "If your [X] had a [Y], what would it be?" Concrete. A clear image to step into, then follow wherever it leads. The experience becomes finding the right song, not figuring out what the question means. When the gate is open, people walk through it.
This is the most direct product insight we have from the month. The question is the gate. Our job is to keep it open.
Chapters are underperforming. 14 created, 78 views, 10 unique viewers.
I want to say it clearly because the instinct is to bury numbers like this when everything else is pointing in a good direction. Chapters aren't doing what we hoped yet. The concept is right. A personal room, a collection of moments, something that doesn't look like anyone else's Chapter. All 14 that exist feel genuinely distinct. But almost nobody is finding them.
That's probably a discoverability problem. It might also be a timing problem. A feature that needs more people in the room before it has the audience it deserves. I'm not sure yet. But I'm not going to pretend it's working when it isn't.
What We're Building Next
Every feature in the next cycle is a direct response to something the data showed us.
Question archive. 521 posts across 25 questions. These aren't disposable moments. They're a record of how people moved through a month. We're building a way to surface that history, to let users return to what they wrote two weeks ago, to feel the distance between who they were and who they are now. The "save a moment to revisit a year from now" prompt this week proved the community is already thinking about looking back. We should make that possible.
Comment editing. 440 comments, top posts hitting 8 to 11. People are thinking before they write here. That deserves better tools. The ability to finish an in-flight thought, to correct something said too quickly, to express without constraint after the moment has passed.
Inner Circles. The poll already proved this: people want to write the questions, not just answer them. Inner Circles is the natural next step. Small groups hosting their own daily ritual. Their own questions, their own community, their own definition of what this space is for. There's a version of Circle where every question comes from the community. Shaped by it, molded by it, not from one voice.
Activation. We need more first posts. Not more signups. The community is alive. The experience, once someone is in it, speaks for itself. The work now is making sure people actually get in.
One Person
I want to say something about how this is being built, because I think it matters.
This is one person. Limited time, real constraints, every decision made intentionally because there's no room for anything that isn't tied to what we're actually seeing.
Nothing is added lightly. This space doesn't feel like mine anymore. It feels like a duty. The people who show up every night to answer a question honestly deserve a product that earns that honesty. Every feature is built in that direction. Every metric is read with that in mind.
Why It Will Grow
I'm not going to say Circle will grow because the market is there, or because social is broken, or because the timing is right. I'll say it because of what the numbers already show.
High retention. Increasing depth of engagement. A community that is actively shaping its own experience. A ritual behavior that has, for at least two people, become part of who they are.
Products like this don't grow because they're pushed. They grow because people come back, and eventually, they bring someone with them.
We're not trying to fix social media. We're not in competition with it. We're building something outside of it. A quieter internet, with a little more integrity, for people who are ready for that.
One by one, they're finding their way here. And when they do — they stay.
One question. One song. One honest corner.