Why Is My Online Community So Quiet?

You open your community, see three hundred members and no new posts since Tuesday, and start wondering whether anybody is there. Most of them are there. They are reading, and the research on online communities has said for twenty years that this is what most members do.

Yes. In 2006 Jakob Nielsen, the usability researcher, named it participation inequality and gave it a rule of thumb that is still quoted everywhere: in most online communities, about 90 percent of people read and never contribute, 9 percent contribute now and then, and 1 percent account for almost all of the activity.

Ten years later, The Community Roundtable, a research firm that studies community management, pushed back on that rule. In its 2016 research it found participation closer to 55, 25 and 20 once you count only the members who logged in that month. Discourse, which makes community software, reads those numbers as about 20 percent creating content, 25 percent responding to it, and 55 percent reading without taking part.

The two sets of numbers disagree on how much people take part, but they agree on the shape. Even in a well run community, the largest group of active members is reading. A member who reads every post and never comments is still getting what they came for, and plenty of them will buy, book or refer somebody without ever leaving a reply.

A community feels dead when the few people who would post do not, and the most common reason is that the first thing they are asked to do is too big.

Think about what a typical welcome asks a new member to do. Introduce yourself, tell everyone about your business and share what you are working on, in a room full of people you have never met, where the last post is four days old. That is three requests at once, all of them written from scratch, and all of them public. Most people will decide to come back later, and later rarely comes.

Nielsen’s own answer to how you overcome participation inequality was “You can’t.” What he did suggest was reducing it, and two of his suggestions speak directly to that welcome post: make it easier to contribute, and have people modify something rather than create it from scratch. A first post that asks somebody to write a short biography breaks both.

A quiet community is usually not a dead one. Its first step is too big.

It does, and the people who build community software say so plainly. Jeff Atwood, a cofounder of Discourse, wrote in 2014 that “the worst thing you can do is open with a blank site.” His advice was to have real conversations already running before people arrive, to write the kinds of posts you want to see, and to have the owner and staff visibly answering and welcoming people.

The same thing applies after launch. When a new member lands and the newest post is the owner’s announcement from last week, nothing on the screen shows them what a normal reply looks like or that anybody will answer theirs. Nobody wants to go first in an empty room.

Make the first step small enough that a person can take it in under a minute without thinking about how it will look. In practice that means a few changes.

• Ask a question that can be answered in one line. “What are you working on this week?” gets more answers than “Tell us about yourself.”
• Give people something to finish rather than something to start. A fill in the blank prompt, or a short list of choices, is easier than a blank box.
• Count the small actions. A reaction or a one word reply is a first step, and it is the step that comes before a post.
• Go first. The owner answers the welcome question before anyone else does, so the newest member can see what an answer looks like.
• Answer every first post, fast and by name. The person who gets a reply the same day knows somebody is reading.
• Reward the first post, but keep the reward small. Nielsen’s advice was to reward participants without over rewarding them, because people who post only for the prize stop when the prize stops.
• Measure the people who showed up, not the people who signed up. The Roundtable’s numbers came from counting members who logged in that month, and a list of every person who ever joined will always make the room look emptier than it is.

Central Pathway, the free community HubArchitect™ runs, is built around this. Every new member is pointed to one pinned post in Start Here that asks three short questions: what you are building, what feels disconnected right now, and what you want to make stronger first. The welcome email points to that post, and four days later a short follow up email points to it again for anybody who has not gotten there yet.

The first post carries a small points reward, comments and reactions earn points too, and those points move members up through the community’s levels. Gerald also sends every new member a direct message when they join, so the first reply a member gets comes from a person, not an automation.

None of this is a promise about numbers. Central Pathway is a young community, and most of its members read more than they post, which is exactly what the research says to expect. The point of the setup is that nobody joining has to wonder what to do first or whether anybody will notice when they do it.

Quiet becomes a problem when people stop showing up, not when they stop posting. If members are still logging in, opening posts and coming to live sessions, the community is being used, and the work is to make the first post easier. If the people who used to log in have stopped coming back at all, that is a different question about whether the community still gives them a reason to visit, and no welcome prompt will fix it.

Looking at the first thing your community asks a new member to do, could a stranger answer it in under a minute?

The 90, 9 and 1 figures are Jakob Nielsen’s rule of thumb from 2006. The 55, 25 and 20 figures are from The Community Roundtable’s research, which counts members who logged in during a given month, and the breakdown of what each group does is how Discourse describes that research. Neither is a law, and your own community will have its own numbers.