Community

Why Your Community Fails in Month Three

I spent two years running a technical forum for open-source tooling. It hit about 4,000 active monthly users before it started dying. Not because the content was bad. Not because we lacked features. It died because we treated it like a product instead of an ecology. Here's what actually happens, and how to avoid the same mistakes. A community is a group of people who interact around a shared interest, goal, or identity, with enough repeated interaction that norms, relationships, and culture emerge. That's the textbook version. The real version is messier. A community isn't a forum. A community isn't a Discord server. Those are vehicles for community. You can have a community without either of those, and you can have both of those without a community. The distinction matters because it changes how you approach building, growing, and sustaining one. The vehicle determines friction. A forum thread takes effort to navigate. A Discord channel is instant but ephemeral. A subreddit is accessible but algorithmically hostile to niche topics. I've seen the same group of people thrive on one platform and wither on another. Pick the vehicle based on your members' behavior, not your preference.

Getting From Zero to Viable

The hardest phase is the first 100 people. Not because acquisition is impossible, it's because your community doesn't have enough internal density yet to sustain itself. You need enough people to respond to each other quickly enough that newcomers feel like they've arrived somewhere alive. This usually takes 3 to 6 months of consistent presence from you or your core team. I started by personally responding to every single post in the first four months. Not with elaborate answers, just acknowledgments, follow-up questions, and introductions. It took about 45 minutes a day. People noticed. They started responding to each other instead of waiting for me. That's the threshold. Once response chains form between regulars without your involvement, the community has crossed a self-sustaining line. The counter-intuitive part: don't optimize for member count at this stage. Optimize for interaction density. Fifty people who reply to each other daily are worth infinitely more than 500 who lurk. I watched a competing community hit 12,000 members in six months and dissolve within a year. Their ratio of posts to replies was about 8:1. Ours at 4,000 was roughly 1:2. We outlasted them by a factor of four.

Common Structural Failures

Over-moderation early on: When you're small, every rule feels necessary. It isn't. Strict content policies and aggressive moderation in the early phase suppress the very interaction patterns you're trying to grow. I had a co-founder who wanted threaded discussions only, no off-topic chat, and approved posts before they went live. We lost 60% of our first-month signups. Opening it up and letting things be slightly messy increased engagement threefold within two weeks. Platform dependency: Building on someone else's platform means you don't own the relationship. Reddit changed their API pricing in 2023 and several communities I knew of evaporated overnight because their entire infrastructure was tied to Reddit's internal systems. Always have a migration path. Keep an email list. Export your data quarterly. I learned this the hard way when our Discord community lost access to three days of archived conversations after a platform outage, and the knowledge loss forced us to rebuild from scratch. Monetization too soon: This is the most common mistake I see. A community that hasn't established trust and value gets charged before it's earned the right to ask. The transition from free to paid needs to happen after members have repeatedly benefited from the community's existence without paying. I've seen communities introduce paid tiers at 2,000 members and lose half their active base within 30 days. The same community at 15,000 members with a generous free tier kept 80% of their payers.

Measuring What Actually Matters

Most people track member count, post volume, and DAU/MAU ratios. These are vanity metrics for community health. They tell you size, not vitality. The metrics that matter are harder to capture but far more predictive: Response latency: How long does it take a newcomer to get a reply to their first post? Under 2 hours is healthy. Over 24 hours and you're watching them churn. Connection density: What percentage of members have interacted with at least three other members? Below 20% and you're running a bulletin board, not a community.

Organic emergence rate: How often do members create content, events, or sub-groups without prompting from you? If the answer is never, you're the community. That's not sustainable. Defection rate: When a member leaves, do others notice? Do they reach out? Communities with high defection awareness have genuine social bonds. Communities where people vanish without a trace have transactional relationships at best.

The Long Game

Communities peak and then plateau or decline. This is normal. The goal isn't to prevent decline, it's to manage it gracefully. Some communities evolve into something smaller but tighter. Some dissolve entirely. A few find a stable equilibrium where growth stops but the group remains valuable to those in it. All three outcomes are valid. The people who treat community building as a career tend to burn out within 18 to 24 months. The ones who survive treat it as a long-term practice, accept that their role shifts from builder to steward over time, and don't tie their identity to growth metrics. I stepped back from daily moderation after year two and shifted to quarterly strategy reviews. The community didn't collapse. It adapted. It wasn't perfect. Nothing sustainable ever is.