Community case studies, taken apart

Any B2B community ongoing · Structure

Nine thousand members, and the team writes every answer

What was done

A company launches a Slack community, invites its whole customer list, and staffs it with two people from support who answer everything within minutes.

What was right

The response time is genuinely excellent and members say so. Nobody is being ignored, and the intent behind it is real care rather than neglect.

What went wrong

Because staff always answer first, members never get the chance. Member to member reply share sits near zero, so the room needs two full-time people forever and produces none of the compounding that justified building it.

What should have happened

Deliberately leave two questions a week unanswered by staff for a few hours. Name three members publicly as people worth asking. Report member to member share instead of total members.

What the right plan does to the economics

Our own arithmetic on a modelled team: two people at roughly 60,000 a year each is 120,000 spent to run a channel that behaves like a helpdesk. Moving member to member share from near zero to half typically lets one of those roles move to other work, and the community keeps functioning. That is about 60,000 a year, and the room gets better rather than worse, because members answering each other is the thing people stay for.

Our own modelled team and arithmetic, offered as a worked example rather than a named case

Any company with an advocacy tool ongoing · Distribution

Forty people reshared the launch. By week three it was eight.

What was done

A company buys an employee advocacy platform, loads company posts into it, and asks staff to reshare. Participation is tracked on a leaderboard.

What was right

The instinct is right. A post from a person does reach further than the same post from a brand page, on every major platform, and that is a real and documented asymmetry worth using.

What went wrong

Resharing sends identical content from smaller accounts to an audience that overlaps heavily with the company page. Unique reach is a fraction of the summed figure on the dashboard, and the leaderboard turns a favour into an obligation, which is why participation collapses.

What should have happened

Ask five people to write one post each about a problem they solved this week, in their own words, with no template and no approval queue. Write down what staff are allowed to say without approval, because that document is the actual programme.

What the right plan does to the economics

Our own arithmetic: a tool licence at roughly 8 dollars per user per month across 200 staff is about 19,200 a year. The reshare programme it enables produces heavily overlapping reach. The version that works costs an hour of five people’s time a month, roughly 3,000 a year in salary, and produces content nobody else could write. The expensive part was never the software.

Our own arithmetic on modelled costs · platform reach asymmetry is widely documented and observable on any account

Any ambassador programme ongoing · Retention

The channel went quiet, so somebody suggested better swag

What was done

Two hundred customers are recruited as ambassadors, given a discount code, a t-shirt and a private Slack channel. Three months later the channel is silent.

What was right

Recruiting was easy and the enthusiasm at signup was real. People genuinely wanted to be associated with the brand, which is the hardest part to manufacture and it was already there.

What went wrong

Nothing was asked of anybody. Merchandise is a token of status, not the substance of it, so once the novelty passed there was no reason to open the channel. Two hundred names also meant nobody had a relationship with anybody.

What should have happened

Cap intake at fifteen, speak to each personally, give one specific recurring task, and set a six month term with a clear end so leaving is not awkward.

What the right plan does to the economics

Our own arithmetic: 200 ambassadors at roughly 35 dollars of merchandise and shipping each is 7,000 spent once, against a channel that produced almost nothing. Fifteen people with a weekly task and genuine early access cost about 525 in merchandise and perhaps two hours a week of a community manager. The smaller programme is roughly a tenth of the cost and it is the one that still exists in month twelve.

Our own arithmetic on modelled programme costs, offered as a worked example

No figure on this page is attributed to a named company. Everything is our own arithmetic on modelled teams and is labelled as such. Teaching material, not advice for your business.

Three programmes, taken apart.

Nine thousand members, and the team writes every answer

Forty people reshared the launch. By week three it was eight.

The channel went quiet, so somebody suggested better swag

Tell us what your room looks like. We will say what is wrong, free.