Should you build a community or a membership?

The pitch is a good one. Recurring revenue instead of launches. An audience that talks to itself. Something competitors cannot copy, because they cannot copy your people.

Now picture month seven. The weekly call has four people on it. Two of them are your team. Somebody posted a question on Tuesday and nobody answered it, so you answered it, again.

Both rooms are ordinary outcomes. Which one you get is mostly settled before launch. It has little to do with the platform you pick.

The members are the product

Look at what the durable paid communities sell. Vistage has been running CEO peer groups since 1957. Chief leads with the network and the peer group before it mentions anything else. Hampton puts founders into small groups that meet monthly, and it turns away roughly 92% of applicants.

None of those is selling access to a founder's brain. They are selling access to each other.

That is the fork. If the value in your room is the other members, you have a community. If the value is you, you have a coaching program with a chat window attached. It will run on your presence for as long as it runs.

Both are legitimate businesses. They ask different things of you. Only one of them keeps working in a week when you are ill.

Participation is lopsided, and the arithmetic starts small

Jakob Nielsen named the pattern in 2006, and it has held up.

In most online communities, 90% of members read without contributing, 9% contribute occasionally, and 1% write most of it.

Small private professional groups do better than that. Higher Logic's own community research puts the split closer to 70-20-10 in smaller communities. Members know each other there, and the topic is narrow. The curve flattens. It does not invert.

So run the numbers on your own launch. Forty paying members is a good first month for most experts. At the friendly end of that range, eight of them post sometimes and four carry the room.

Four people. That is your entire conversation in week one, and two of them will be traveling.

This is the cold start, and it is why so many communities feel thin in the first quarter. The room needs enough people to make a conversation likely on a random Tuesday.

A quiet room costs more than no room

Here is the part the platform demos leave out.

A member pays you, opens the space, and finds three posts from last month. They do not conclude that the community is young. They conclude something about you.

Then they do the thing most people do in a silent room. They say nothing, because being the only person talking is worse than lurking. They read for a few weeks, and they cancel quietly.

Those people were your best buyers. They liked you enough to pay for recurring access before they knew what it was. Spending that goodwill on an empty room is expensive in a way the churn report does not show.

No community at all costs you nothing. A quiet one costs you the people most likely to buy the next thing.

Four things to see before you build

None of this is an argument against building one. It is an argument for reading the room you already have first.

• People are already asking to meet each other. They reply to your newsletter asking whether you know anyone else dealing with this. If nobody has asked, you would be creating the appetite for belonging. That is the hardest kind of demand to manufacture.
• Your members can help each other. Same stage, same problem, and not competing for the same clients. A room of people who want something from you and nothing from each other stays quiet. Better software does not change that.
• You can hold a cadence on your worst month. One live call a week, or one thread you start every Monday. Pick the version you would still run in a bad quarter, then commit to it for two years.
• You are willing to turn people away. Curation is what keeps the conversation worth reading. A community that takes all comers becomes a room where few people recognize anyone.

Three out of four is a signal to wait. Four out of four means the thing already half exists, and you are formalizing it.

What AI changed about this decision

Plenty of older communities ran on questions and answers. Somebody asked, somebody knowledgeable replied, and that exchange was the value.

Chatbots answer those questions now, instantly, without anyone having to be awake. A community built as a help desk has already been replaced.

What survived is the part a model cannot supply. Forrester's June 2026 survey of B2B community programs found that peer-to-peer connection is what members value most.

AI does help you run the room. It drafts prompts, flags spam, surfaces the thread worth your reply, and shows you who has gone quiet. It cannot make people care whether they show up, and caring is the whole asset.

The smaller version worth running first

A community is one of several recurring formats. Which one fits your business follows from what your audience already buys, a separate question worth its own answer.

If the four signals are close but not there, run the cheap test. A six-week cohort. A quarterly roundtable for 12 people. A monthly call attached to something you already sell.

Each one has an end date, which is the point. You find out whether these people want to be in a room together. You also find out what hosting them does to your week, before you promise it forever.

Watch two things while it runs. Whether members talk to each other without you starting it. And whether you look forward to the call on the morning of the call.

The second one matters more than it sounds. Hosting is the part that quietly turns you into the constraint on your own business. It arrives as a standing appointment you cannot reschedule.

Every other product you sell works when you make it well. A community works when other people show up for each other on a quiet Tuesday, in a slow week.

That part sits outside your control, and production quality does not buy it.

So the question worth answering first is a small one. Does the room already want to exist?