The community I built by accident

Priya's office in Pune is on the third floor of a building that used to be a textile warehouse, above a chai stall whose owner she's known since she moved in. The stall opens at 6:30am. By 7 she's usually downstairs with a small steel cup, watching the auto-rickshaw drivers argue about cricket, and then she's back upstairs at her desk with the window open because the AC is broken and has been for six months and she's stopped caring.

On a Tuesday in February she opened her old Discord server for the first time in almost two years. The name of it was still there in the sidebar, slightly faded, like a friend who'd moved away. 'Maker Mornings.' She'd named it that in 2024 when she launched her first product, a small Pomodoro app for designers, and she'd been so sure the community would matter that she'd built the Discord before she'd built half the features.

She scrolled to the most recent message. October 14, 2024. Someone named 'rohan_2294' had said 'lol same'. Below that, no replies. Above it, six weeks of her own messages — the welcome pack she'd written, the kickoff call announcement, the pinned agenda, the Loom intro she'd recorded in her bedroom with a ring light she'd bought specifically for it. She watched 20 seconds of it before closing the tab. The Priya in that video was earnest in a way that made present-day Priya wince.

She didn't reopen the server that day. She didn't reopen it for another month. But three months later, when her second product — a much smaller, much weirder thing for indie devs who wanted to deploy from a phone — hit something like 600 paying users almost by surprise, she thought about it again. Some of them had started emailing her with questions that had nothing to do with the product. They wanted to know what other people were using. They wanted to know if she knew anyone hiring. One of them, a guy in Lagos, just wanted to say hi.

On a Wednesday in March she opened a new Discord. She named it after the product. She made one channel, called it #lounge, posted one message — 'hey, welcome, no agenda, talk about whatever' — and put the invite link in the footer of her next product update email. Then she closed the tab and didn't open it for two months.

What killed Maker Mornings

The first Discord died of organization. Not from neglect, not from spam, not from a Reddit-style mod war — from sheer earnest scaffolding. Priya had read three threads about how to build a community and she'd taken them all literally. There was a #introductions channel where you were supposed to post a Loom of yourself, presumably so 40 strangers could watch you say your name. There was a Monday kickoff call at 8am IST that two people attended twice, including her. There was a weekly accountability thread that she pinned and unpinned and re-pinned. There was a 1,200-word welcome message that nobody read.

The structural problem wasn't the structure. It was that none of the structure came from anyone wanting it. It came from Priya wanting the community to exist before the community existed. She had built the foyer, the coatrack, the guest book, the little dish of mints by the door, and then she'd opened the front door and nobody came in because there was no party — just her, standing in her own foyer, asking every passer-by if they wanted to introduce themselves.

There's a particular feeling when you run a Discord like that. You check it every 20 minutes. You see the same three usernames in the sidebar. You write a 'hope everyone's having a great week!' message in #general because the silence is unbearable. The message gets one wave emoji, from someone you suspect added the server by accident. You feel, very briefly, that you might be the protagonist of a small tragedy.

Maker Mornings died at six weeks. Priya didn't formally close it. She just stopped posting in it, and then everyone else, who had been politely posting in it because she was, also stopped. By month three the server was a graveyard. By month six it was a graveyard with a Loom intro pinned to the wall.

Why she didn't open the new one for two months

Partly she was busy. The new product was small but it was growing in a way that her first one hadn't, and most of her days were spent in support emails and a deploy queue and a long argument with Stripe about KYC. But partly — and she's only said this to me once, and only after I'd asked her three times — she was scared.

Scared of what, exactly. Scared that if she opened it she'd start managing it. She knew herself well enough to know that. She'd watch the new-message count climb to 8, then 12, then sit at 12 for an hour, and she'd panic and write a thoughtful message that would kill whatever was about to happen. She'd done it before. She didn't want to do it again.

So she opened the Discord, posted the welcome line, put the invite in the email footer, and then physically removed the Discord app from her phone. She kept it on her laptop, but she muted every channel. The mute icon next to #lounge looked, she said later, like a small promise to herself.

For the first two weeks, she said, she got the urge to check it maybe forty times. By week three the urge was down to once a day. By week six she'd more or less forgotten the server existed. The product was growing. The support queue was full. The chai stall guy had started letting her use his stool when his back hurt. She didn't think about the Discord.

What she found when she went back

She opened it on a Sunday morning in late May, mostly out of mild curiosity, the way you might check on a plant you'd left at a friend's house. The server had 84 members. There were six channels she hadn't created. There was a pinned FAQ in one of them that ran to about 900 words and was, as far as she could tell, entirely accurate.

The channels: #lounge, which she'd made, plus #showcase, #help, #jobs, #lagos-meetup, and one called #cricket that she initially assumed was a joke and which turned out, on inspection, to be a serious channel for people who were mostly Indian devs and who wanted somewhere to talk about cricket without being annoying in #lounge. The FAQ was in #help and had been written by someone named 'mara' who Priya had never spoken to, and it answered the most common questions about her product better than her own docs did, because Mara was a user who'd actually been confused by all the same things and had taken notes.

There was an onboarding ritual. New people, when they joined, posted in #lounge with a one-line bio — not because anyone had told them to, but because the existing members had been doing it for weeks and new arrivals matched the pattern. Nobody had pinned anything titled 'how to introduce yourself.' It had just happened. The bios were short and funny and weirdly specific. 'arpit, bangalore, deploy from the train.' 'noor, beirut, two laptops because one of them has a key stuck.'

Priya sat at her desk for about an hour, scrolling. She didn't post anything. She made herself a second cup of chai. She came back and scrolled some more. Around noon she wrote one message in #lounge that said, 'this is incredible, I've been away, thank you for being here,' and went to lunch.

The exact things she didn't do

Looking back at the two months between opening the server and finding 84 people in it, the absences are the interesting part. Here's what Priya specifically did not do during that window, none of which were strategic decisions — most of them were the result of not opening the app.

• No scheduled AMA. She never sat down with the founder-as-oracle hat on and announced a Wednesday 4pm Q&A. The questions people had got asked in #help, mostly to each other, and the answers compounded because they stayed visible.
• No daily check-in thread. No 'what are you shipping today?' prompt. The people who wanted to share what they were shipping just shared it, in #showcase, when they had something. The people who didn't, didn't.
• No leaderboard. No XP, no roles for activity, no 'top contributor of the month.' No one game-ified anything. People posted because they had something to say or because they wanted help, not because they were earning points.
• No forced introductions. No #introductions channel with a template. The bios in #lounge happened because the first three people did it and it spread. The pattern was contagious. The pattern would have died if there was a template.
• No Loom from the founder. No welcome video. No 'hey I'm Priya, here's the vibe.' The vibe got set by whoever showed up. Mostly it got set by Mara, who is now, by acclamation, an unofficial second mod.

The temptation, every one of these, is to read them as wisdom. They weren't. They were the byproduct of an absent founder. The Discord worked because Priya wasn't there to make it not work. That's not a flattering story, and Priya, to her credit, is the first to say so.

Community-around-the-product versus community-of-people

There's a distinction worth naming here, because it explains why Maker Mornings died and why the new server didn't. A community-around-the-product is one where the product is the point. The conversations are about the product. The reason people show up is to talk about the product. The Slack groups that some SaaS companies run for 'power users' are like this — useful, sometimes, but they have the conversational range of an HR onboarding meeting.

A community-of-people-who-happen-to-use-the-product is different. The product is the entrance — it's what got everyone in the same room — but it's not the topic. The topic is whatever the people themselves bring. Cricket. Jobs. The Lagos meetup that 11 of them organized last month at a co-working space and at which exactly one person mentioned Priya's product, and only in passing.

The second kind is more valuable, by a wide margin, and almost impossible to plan. You can't decide your community is going to be about more than the product. The members decide that, or they don't, and they only decide it if there's enough oxygen in the room for it to happen. Founders who post 'reminder: keep things on-topic' kill the second kind dead. Founders who let a #cricket channel exist, even if they don't watch cricket, are the ones who end up stewards of something bigger than their product.

Maker Mornings was trying to be the first kind. The new server became the second kind, because nobody enforced the first.

The three things she did do that mattered

It's not that Priya did nothing. The story sounds better if she did, but it's not quite true. There were three things she did, after she came back from her two-month absence, that mattered enough to name. They are not strategies. They are habits she stuck to even when they were inconvenient.

1. 1.She replied within 12 hours to every direct question. For the first three months after she came back, if someone @-ed her with a real question, she answered within half a day. Not with a long answer — sometimes one sentence. But she answered. The signal this sent was not 'I'm here always.' It was 'if you actually need me, I exist.' That's the only message a community founder needs to send.
2. 2.She never deleted a thread that drifted off-topic. The cricket channel started as a tangent in #lounge that ran for 200 messages over a weekend. A lesser founder would have pinned a 'please keep #lounge on topic' message. Priya watched it happen, noticed it was making people happy, and a week later created #cricket herself with a one-line message: 'making this its own channel because you all clearly want it.' The community had voted with its feet. She followed.
3. 3.She posted product updates with the why, not just the what. When she shipped, she didn't post a changelog. She posted a paragraph about why she'd made the choice — what tradeoff she'd been wrestling with, what she'd almost done instead, what she was nervous about. The updates got replies because there was something to reply to. People felt like they were inside a head, not reading a press release.

None of these scale, in the strict sense. They're the kind of things a founder does when there are 80 to maybe 800 people in a room. After that you need help, and the help comes from people like Mara, who emerge when there's room to emerge and don't emerge when the founder is everywhere at once.

Where to put a community, and what each surface is actually for

Priya's the wrong person to ask about platforms in the abstract — she'd say the platform doesn't matter, which is half-true. But she's been in enough of them, as a user and a maker, that she has opinions. Here's a rough map of the surfaces and what each is genuinely good at, stripped of the marketing copy each platform writes about itself.

• Discord. Real-time, low-stakes, casual. Good for the second kind of community — the one where people stick around for the room, not the product. Bad for anything that needs to feel professional or that involves customers paying you a lot of money. Threads disappear. That's a feature, mostly, because most threads should disappear.
• Slack. Work-feels, professional-ish, slightly higher stakes. Good for B2B communities where the members are using your product inside their company and want to ask questions in a context that looks like the rest of their work. Bad for vibes. Slack communities for designers and indie devs almost always feel slightly off, like a team meeting that won't end.
• Circle, Skool, and the paid-community platforms. Mid-formality. Good for memberships, courses, cohorts — anything where there's a transaction at the door. Bad as a free hangout, because the platform itself signals that there's supposed to be a curriculum. People show up looking for the lesson.
• Reddit. Drive-by traffic, basically no ongoing relationship. Good for top-of-funnel — your product gets mentioned in a thread, a few people click through. Bad as the place your most engaged users live, because Reddit's incentives don't reward being a regular. Everyone is a stranger.
• Email lists. Asymmetric, low intimacy. Good as a broadcast channel. Bad as a community. The reply rate is decent — better than people think — but the replies stay between you and the sender. Nobody builds an onboarding ritual in your inbox.
• In-app comments and feedback widgets. Highest signal, lowest reach. The people who comment on a feature inside your app are the people who care most. The ten of them who do are worth more than the 800 in the Discord, for the purposes of product decisions. They're just not a community — they don't talk to each other.
• An in-person dinner once a year. More useful than all of the above for the top 10 users. Priya did this last year, paid for dinner at a restaurant in Mumbai for the nine people who could make it. Three of them have become near-cofounders in feel — they DM her about strategy, she sends them early builds, two of them are now informally advising her about pricing. The dinner cost less than one month of a community manager's salary and produced more lasting closeness than any Discord ever has.

The mistake most founders make is choosing a platform first, deciding the community will live there, and then trying to bend the audience to fit. The platform is downstream of who the people are and what they want from being in a room with each other. Designers want different surfaces than DevOps people. Vibe coders want different surfaces than CFOs.

What it feels like to be a steward instead of an organizer

Priya talks about her role now in oddly humble language. She doesn't say she runs the Discord. She says she 'lives there.' She means that fairly literally — she's in it most days, she scrolls, she reacts to things, she answers questions when she's the right person to answer, she stays quiet when she isn't. She moderates rarely. She's banned exactly one person in three months, a spammer who was selling a crypto thing.

When she meets other founders and they ask her how to start a community, she's gotten visibly uncomfortable with the question. She doesn't have a method. The method is: don't have a method. Open the door, leave the lights on, don't stand in the foyer, come back in two months. Which is, as advice, almost useless — most founders can't do it, because most founders, by temperament, are people who do things.

The doing instinct is the problem. A founder who is good at doing will, in week two of a community, do things to it. Schedule a kickoff. Pin an agenda. Record a Loom. Each thing they do reduces the space for someone else to do something, and someone else doing something is the actual mechanism by which a community becomes a community. The first FAQ written by a user is the moment the room becomes theirs. If you write it for them, the room stays yours, and a room that's yours is not a community — it's an audience watching you tidy.

Priya's role, as she understands it now, is steward. The community was already happening before she showed up. She just owned the URL. Her job is to keep the URL working, reply when called, host the dinner once a year, and otherwise stay out of the way of a thing that was, at some point in March or April, deciding for itself what it was going to be.

What I'd say to the founder who's about to start a Discord

I'd say wait. Wait until you have users who are emailing each other already, even if you haven't introduced them. Wait until the support emails have started turning into 'do you know anyone who's done X' questions, because that's the signal that your users are looking for each other and just don't have a room yet. Open the room then. Not before.

When you open it, open less of it. One channel, not seven. No rules, no welcome message longer than two sentences, no Loom. Put the invite somewhere quiet, like the footer of an email or the bottom of a settings page. Don't tweet about it. Don't put a giant banner in your app. The people who'll make the community what it should be will find the link and come in, and the people who would have made it noisy won't bother.

Then ignore it. Genuinely ignore it. Not 'check it once a day' ignore it — actually mute it, take the app off your phone, let it sit. Come back in six to eight weeks. If there's nothing there, you opened it too early; close it without ceremony, no announcement. If there's something there, follow whatever pattern has emerged, name the channels that the conversations are already implying, and start replying to direct questions within 12 hours.

That's the whole playbook, such as it is. The hard part isn't any of the individual moves. The hard part is staying out of the way of something whose growth doesn't depend on you, when every instinct you have as a founder is to grab it by the stem and pull.

Before there was a community to land in

I asked Priya, last time we talked, where her users had come from before there was a Discord. Not where the Discord had come from — where the users had. She thought about it for a while. Some had come from Twitter, from a tiny thread she'd posted about a deploy bug. Some from a Reddit post she hadn't even written, just been mentioned in. A few from a podcast appearance, a few from a Stack Overflow answer she'd written three years ago that had quietly become canonical. Most, when she traced it, had come from other people's communities — Discords she wasn't in, Slacks she'd never heard of, group chats that surfaced her product organically in conversation.

Those communities had been her community before she had one. They were where her users lived. She's spent the last six months mapping them, partly out of curiosity, partly to know where to send the next version of herself — the version that hasn't opened her own server yet, who's still in the foyer phase, who needs to know where her people already are.

Her map of those communities lives on the communities directory now. It's not exhaustive. It's her curated list — the rooms where she found users before she had a room of her own, the places she'd send a new founder who hasn't earned an audience yet and needs to be among other people who'll recognize what they're doing. It's the inverse of starting your own Discord, and for most people, most of the time, it's the better move. You don't need a community. You need to be in one.

The chai stall downstairs opens at 6:30. Priya's usually there by 7. The auto-rickshaw drivers still argue about cricket. The window upstairs is still open. The AC is still broken. The Discord, when she checks it now, has 137 people and three more channels than it did in May, none of which she made. Mara wrote another FAQ last week. Someone in Manila started a thread about a deploy bug and 14 people piled in to help before Priya even saw it. By the time she did, the bug had been solved, the thread had pivoted to a discussion of monorepos, and she just reacted with a thumbs up and went back to her chai.