Before You "Build a Community," Decide: Library or Coffee Shop?
We analysed traffic across thousands of Discourse communities and found something intriguing: the internet has been using the word “community” to describe two very different things.
We spent the last month examining traffic data across all the communities we host, and one statistic stands out: 38 communities generate 31% of all traffic on Discourse, while the other 5,000+ communities share just 11% among them.
That's a 382x difference in scale between the typical high-volume community and the typical standard one. And what really caught our attention was that these two groups seem to be doing fundamentally different things.
- The big communities look like libraries.
- Forum A recorded 349 million pageviews over six months in 2025.
- Forum B, 228 million. Forum C, 409 million.
These are massive knowledge repositories, and most users arrive via Google or agentic search, find an answer to a specific technical problem, and leave. In 2019, someone asked, "Why won't my servo motor respond to PWM signals below 20Hz?" and received a detailed answer. Six years later, that topic gets found by dozens of people every week who have exactly the same problem. The ratio of readers to writers in these communities is extreme. Most visitors never post anything. They don’t even sign up. They're there because Google sent them to the one page that solves their problem.
The smaller communities look like coffee shops. They average 154,000 pageviews over the same six-month period. That sounds tiny compared to the giants, but it’s because of one fundamental difference: the people showing up are there to actually talk. They are the minority who have decided to invest their time in having discussions. They're checking in regularly to see what's new. They might be a group of people interested in mechanical keyboards arguing over switch types, a book club working through Proust, or parents swapping potty-training strategies. These communities exist for connection, rather than information retrieval. When someone asks a question, they often get a keen back-and-forth dialogue rather than a single definitive answer. The same usernames keep appearing. There's an actual community in the social sense.
The internet has been using the word "community" to describe two completely different things, and that confusion has probably led to many bad decisions about how to build and manage these spaces. When an organisation says they're "building a community," they might mean they're hoping to create a comprehensive knowledge base that will help thousands of people find information on how to use their product via search. Or they might mean they're trying to cultivate a group of people who will actually put in the effort to have a conversation and learn directly from one another. Those are distinct goals that require distinct engagement strategies, moderation approaches, and possibly different software features.
The library model seems to work best for topics with clear right and wrong answers. Programming languages behave in specific ways. Hardware has specifications. Software has bugs that can be reproduced and fixed. When someone encounters an error message, they don't need discussion; they need the answer. And once that answer exists somewhere credible on the internet in a well-indexed, searchable format, it can serve thousands of people who have the exact same problem. The value compounds over time. Every old topic remains useful indefinitely. The Straight Dope Message Board, which we noticed ranks ninth in total traffic despite its founding columnist retiring in 2018, demonstrates this perfectly. Ancient discussions about scientific questions or word etymologies continue to attract attention because their answers remain valid. For the vast majority of users, there is no motivation to join the community because it has already served its purpose.
The coffee shop model operates differently: the value derives from the social interaction itself. Someone might ask, "What should I read next if I loved A Little Life?" and get twenty different recommendations from twenty different people, each reflecting that person's taste and personality. There's no single correct answer. The point is the exchange of perspectives. These communities need active, engaged members who check in frequently and feel invested in the space. The motivation to join and engage is intrinsic. When someone posts, they're satisfying their need to belong to a group, or their need for influence or for recognition. They continue to return as long as that need is being met.
What strikes us at Discourse as genuinely interesting is that both models work; they just work for completely different reasons, and their value should be gauged by completely different metrics. But here’s the most fascinating thing – they don’t have to be mutually exclusive.
If you're running Forum A, you should measure success by SEO performance, pageviews per topic, and low read time – indicating that people can find solutions quickly. If you're running a small community focused on urban planning, bread baking, or whatever, you should track more traditional engagement metrics like posts per active user, high read time, questions answered, and DAU/MAU (or stickiness).
The 382-fold difference in scale makes sense once you see the distinction. A library can serve unlimited readers without degrading; in fact, more readers make it more valuable because more search traffic means better SEO, which means even more readers. A coffee shop has natural size limits. Once you get past Dunbar's number (somewhere around 150 people who can maintain stable social relationships), the character changes. People stop recognizing names. Conversations fragment into smaller groups. The social cohesion that made the space valuable starts to break down. Maybe you can scale to a few thousand if you intentionally manage growth and fragmentation, but you're never getting to 349 million pageviews with a discussion-focused forum unless you've also built a library.
But the best libraries contain coffee shops.
Nurturing a core group of people within the library who do keep coming back to develop expertise by having conversations with each other means they are around to share that expertise when people ask questions that haven't been answered yet. They also shore up the library's future by keeping its searchable content up to date and evergreen. The relationship becomes symbiotic, with the library providing reach and the coffee shop ensuring the value.
Before beginning work on a community strategy, organisations need to decide whether their target audience needs a library or a coffee shop at that point in time. By recognising the different dynamics, community builders can develop targeted strategies that maximise knowledge retrieval, the power of human connection, or, ideally, both.