Foothold

Growth tips

My paid Slack community added 22 new members every month for eight months. Net growth: six per month. Here is the ceiling I did not know I was building.

23 October 2026 · 13 min read

Chioma Obi had been running all five growth levers for eight months: a referral program, a public weekly digest, three directory listings, two cross-community AMAs. She was adding 22 new members per month. Her community had stopped growing. A peer ran a single calculation on a whiteboard that explained why the growth had stalled and what was underneath it.

The eight-month picture

The B2B Growth Marketing Council is a paid Slack community for heads of growth, demand gen directors, and senior revenue marketers at B2B SaaS companies doing $15M–$200M ARR. $169 per month. Chioma Obi launched it in February 2024 — four months after leaving her role as VP of Growth at a Series B SaaS company where she had run demand gen from $8M to $47M ARR.

The community had real early momentum. Month two: 41 members. Month three: 67 members. Month four: 94 members. By month six (July 2024) she had 147 members. She was adding roughly 25 net members per month. At that rate she expected to cross 200 members by September and be at 300 by January 2025.

The growth trajectory in practice:

The deceleration was unmistakable in hindsight. She had added 153 members in the first six months (February through July). She had added 56 in the next six months (August through January). In the final three months of that period, she had added 11 members — about two per week. She was still adding 22 new members per month. Net growth had dropped to roughly 6 per month.

She had not been sitting still. From month three onward, she had been running what she thought was the full growth playbook.

The five levers she was running

Referral program. Set up in April 2024. Trigger: Stripe payment confirmation. When a new member’s first payment cleared, an automated Slack DM went out congratulating them on joining and asking them to invite a colleague who might benefit. The offer: one month free for each referred member who completed their 14-day trial. Average monthly referrals generated: four new trial starts.

Weekly digest. She published a public summary of that week’s best discussions on Beehiiv every Friday — 680 newsletter subscribers, split between community members and LinkedIn followers who had signed up from her posts. Open rate: 39%. The digest was genuinely good. She summarised the two or three sharpest threads from the week, extracted the consensus view, and wrote one paragraph of her own framing. People who were not yet members could read it and see what the community was actually discussing.

Directory listings. Three listings, set up in May 2024: community.ly (a community discovery platform), a Luma calendar with recurring community events, and a dedicated LinkedIn company page with trial CTA. Combined, these drove 6–8 new trial starts per month.

Cross-community AMAs. Two completed by the end of 2024: one with a paid Slack community for B2B content strategists (joint AMA, each operator ran the session in the other’s workspace) and one with a community for B2B sales leaders. Each AMA had generated 8–12 new trial starts in the two weeks after the session. She had two more in progress for early 2025.

Week-one activation: She had a welcome post in #intros, a channel guide pinned to #general, and a monthly “highlights” DM to all new members at day 30. She had not built a three-touch onboarding sequence. She thought about it but concluded it was a post-PMF optimization — something to automate once the community was larger. Her week-one activation rate, as she would calculate it for the first time in January 2025, was 31%.

The operator roundtable

In January 2025, she attended an online community-operator roundtable — eight operators in a Zoom room, each at different stages, comparing metrics informally. She described her situation: consistent acquisition, stalling net growth, all five growth levers running. One of the other operators, Omotola Adewale — who had been running a paid Slack community for UX designers at design-forward technology companies for three years — asked to share her screen.

Omotola opened a spreadsheet. She typed two cells: monthly acquisition rate (22) and monthly churn rate (empty). “What is your monthly churn rate?”

Chioma hesitated. She had been tracking her member count each month but had not framed it as a churn rate. She did the arithmetic quickly: 200 members at the end of November, 201 at the end of December, 203 at the end of January. She had been losing roughly 17–18 members per month and replacing them plus a few. “About 8 or 8.5%.”

Omotola typed 0.085 into the second cell and showed the formula in the third: =B1/B2. The result: 259.

“Your ceiling is 259 members. At your current acquisition and churn rates, you will never grow past 259, regardless of how many levers you add. Here’s the math.”

She walked through it: at any community size n, monthly churn volume is n × 0.085. When that volume equals 22 — your acquisition rate — net growth is zero. Solve for n: n = 22 / 0.085 = 259. Right now, at 203 members, churn volume is 203 × 0.085 = 17.3 members per month. Acquisition is 22. Net growth is 22 − 17.3 = +4.7 per month — roughly 5. As she grew toward 259, the net would approach zero. At exactly 259 members, she would add 22 and lose 22. Net growth: zero.

“You can add more directories, more partnership AMAs, paid acquisition, anything,” Omotola said. “All you are doing is raising the number in the acquisition cell. But the ceiling formula scales: a higher acquisition rate with the same churn rate produces a higher ceiling, but you still have a ceiling. The only way to escape it is to reduce churn.”

She called this the leaky-bucket ceiling. Every paid community without a fixed activation rate has one. The ceiling is derivable from just two numbers.

What was driving the 8.5% churn rate

Chioma pulled her Slack member activity export that evening. The spreadsheet she built classified every member who had ever joined — 310 in total, of whom 203 were still active and 107 had churned over 11 months — by one variable: had they posted in week one?

Week one was days 0 through 7 from the date they joined. A “post” was any message in any channel that was not an auto-generated Slack notification. Introduction posts in #intros counted. Questions, replies to threads, resource shares — anything that required the member to type a message and send it.

The breakdown:

Of the 213 non-activators:

Of the 97 activators:

The 8.5% monthly community-level churn rate was the weighted average of two structurally different populations. Non-activators were churning at nearly four times the rate of activators. Her overall churn rate was 8.5% because 69% of her members had never activated — not because she had a community-wide retention problem, but because she had a week-one problem concentrated in the 69%.

The implication for the ceiling: if she could raise week-one activation from 31% to 69% (a reasonable target given what other operators had achieved with a three-touch sequence), her expected monthly churn would drop from 8.5% to approximately 4%. The new ceiling: 22 / 0.04 = 550 members — more than double her previous ceiling, without changing her acquisition rate at all.

Why the referral program was not compounding

The referral trigger — billing event, day 14 — reached both activated and non-activated members. But the two groups responded differently.

She went back through her Stripe data and cross-referenced every trial-to-paid conversion with the member’s week-one activation status. Of the 312 referral prompt deliveries over eleven months, 41 had resulted in a referral (a member clicking through and inviting someone). Of those 41 referrals, 32 had come from the 97 activated members — a 33% referral rate from activators. The remaining 9 had come from the 215 non-activated members who had been paying long enough to receive the prompt — a 4% referral rate from non-activators.

She was running her referral program on a base where 69% of the recipients had no positive experience to share. The 4% referral rate from non-activators was generating noise rather than signal: those referrals converted at a lower rate and churned faster, because a referral from a member who had not actually engaged with the community was essentially a cold lead wearing a warm-lead label.

The second problem was the trigger timing. An activated member who receives a referral prompt at day 14 — two weeks into their membership — is not at peak motivation. They were at peak motivation the day they had their first meaningful exchange in the community: the day they posted an introduction and got four substantive replies, or the day they asked a question about budget allocation and got a thread with seven practitioners sharing their actual approaches. That day might have been day three or day five. A referral prompt that arrives at day 14 is arriving nine to eleven days late for the activated member and is arriving at a completely wrong moment for the non-activated member.

After she fixed activation and switched the referral trigger to the first activation event, monthly referrals went from 4 to 12 per month — the same offer, the same mechanics, fired at the right moment for the right members.

What the directories were doing

The directories were not the problem. The traffic they sent was real, in-ICP, and converting to trial starts at a reasonable rate: roughly 8% of directory visitors who clicked through to the landing page started a trial. That is a normal directory conversion rate for a $169/month paid community in a professional vertical.

The problem was what happened to that traffic after the trial start. Of the 7–8 trial starts per month coming from directories, roughly 31% activated in week one (matching her overall activation rate). The other 69% — 5 of 7 — went quiet after joining and churned by month three. The directories were sending her qualified traffic, and her community was losing 70% of it within 90 days.

There was a second effect she could not see directly: her community’s reputation in the directories was degrading in proportion to the activation problem. The conversion rate from directory referrals had dropped from 11% in month four (when the community was newer, the activity signals in her listings were fresher, and the 31% of new members who activated were more visible to browsers) to 8% in month ten. A community where most members have never posted looks different in a directory’s preview than a community where most members are active. She was slowly poisoning her own directory listings by sending inactive traffic through them.

She had been trying to address the stagnation by doing more — applying for two more directory listings, scheduling a third cross-community AMA. Each of those moves would temporarily push more traffic through the same leaky system. The directories did not need to change. The community they were pointing to needed to change.

Building the three-touch sequence

She spent three weeks building the onboarding infrastructure she had deferred since month one. The first thing she built was not the sequence itself — it was the intake form. Without intake data, the Day-0 DM could not be personalised; without personalisation, it would work like a generic welcome email, which she already had and which was producing 31% activation.

She added four questions to her Stripe checkout (via a Typeform integration): current role title, company ARR range, primary growth marketing challenge right now (open text, prompted with “be specific — not ‘driving more pipeline’ but the specific constraint you are trying to solve”), and what they wanted to get from the community in the first three months. She tested the form herself and found it took 90 seconds to complete.

The three-touch sequence:

Day 0: Within one hour of a new member joining the workspace, Chioma (or, eventually, Foothold) sent a personal DM. The message named the stated challenge from the intake form, linked to one specific current thread in the most relevant channel, and gave one action: “Given what you described about [challenge], there is a thread in #[channel] from this week that I think you would have something to add to: [link]. Worth a read and maybe a reply — [thread author] is running demand gen at a company in a similar ARR range to yours.”

Day 3: A conditional check. The Day-3 message only went to members who had not yet posted anything — not a generic reminder for everyone, which activated members would find condescending. For non-posters, the Day-3 message lowered the bar: instead of asking them to post a reply to a thread (the Day-0 ask), it asked them to do something with less social exposure: “Quick check-in — I noticed you haven’t had a chance to drop into the workspace yet. The lowest-bar way to get started is just to reply to the #intros thread from [specific recent member who joined this week]. They asked a question about [topic from their intro] and you might have a useful take.”

Day 7: The operator scorecard. Not a member-facing message — a summary DM to Chioma herself listing every member who had joined in the last seven days, their activation status (posted / not posted), their stated challenge from intake, and days since join. The three most at-risk members (joined 6–7 days ago, not yet posted) were flagged at the top. For each at-risk member, she sent a manual personal follow-up: a direct question that assumed engagement (“I know [thread from intake-matched channel] was live this week — did you see [specific practitioner’s] take on [topic]? Given your challenge with [their stated problem], I thought it might be directly relevant.”).

She had estimated this would take 30–45 minutes per week for the Day-0 DMs plus 15–20 minutes for the Day-7 follow-ups on at-risk members. The actual time was higher in the first month — she was writing each DM from scratch while she tested what phrasing worked — but settled at 40 minutes per week once she had templates for each of her four channel routes and three intake-form patterns.

The first three months after deployment

She deployed the sequence in February 2025, exactly one year after launching the community.

Week-one activation in the first month of the sequence: 48% (up from 31%). The first month was partial — she did not have intake form data for existing members, only for new joins after the Typeform integration went live, so the personalisation was limited for the first few weeks. Month two: 61%. Month three: 69%. She stabilised between 67% and 72% over the following six months.

Monthly churn tracked the activation improvement but with a delay: the non-activating members who were still paying continued their pre-sequence churn pattern for several months while the new activating cohorts built up. Community-level monthly churn:

The lag was expected: the non-activated members who had been paying for several months continued their pre-sequence churn pattern until they churned or (occasionally) self-activated from community content. The improvement in community-level churn rate reflected the shift in the composition of the membership — an increasing proportion of new members who had activated and an decreasing proportion of long-tenured non-activators who had not yet churned.

Referral acquisition after switching the trigger from billing to first activation event:

Total monthly acquisition went from 22 to 31 — not because she added a new acquisition channel, but because the referral program started functioning as a lever rather than as a poorly-timed email to non-activated members.

The compounding effect

The first six months of the sequence produced growth that looked different from any of the previous eight months:

The same 22-per-month gross acquisition she had been running since month three, now with a 31/month gross (referrals improving), and a churn rate that had gone from 8.5% to 3.8%. The new leaky-bucket ceiling at those numbers: 31 / 0.038 = 816 members — more than three times the ceiling she had been approaching a year earlier.

She was not at 816 members. She was at 413. But the growth trajectory had changed structurally: the community was compounding rather than approaching stasis. The same directory listings that had been sending 7–8 trial starts per month to a below-31%-activation community were now sending the same traffic to a 69%-activation community. The conversion rate from directory referrals had improved to 11% — back to where it had been in month four, when the community was small and fresh. The activity signals in her listings had improved as the proportion of active members increased.

MRR:

The cross-community partnerships she had been scheduling also started producing different results. The two AMAs she ran in early 2025, with activation above 60%, brought in 12–15 trial starts each (up from 8–12) and converted at higher trial-to-paid rates. The partner operators were noticing: both had mentioned that the members they sent had reported the community being more active than they expected. One had reached out to propose a third AMA unprompted.

What she would have done differently

Chioma summarised the error in one sentence: she had treated the five growth levers as a parallel list, not an ordered dependency chain.

The reference guide covering all five levers explains the implementation order clearly: activation first, then referrals wired to activation events, then content cadence, then directories, then partnerships. She had read similar advice and understood it abstractly as “fix retention before scaling acquisition.” What she had not understood was that this was not a preference or a best practice — it was a mathematical constraint. Each lever is a multiplier on whatever base it operates from. A referral program multiplies the referral behaviour of the membership base it runs on. At 31% activation, it multiplied 31% of her base by a high referral rate and 69% by a low one, producing a blended rate that felt like the referral program was “working but not taking off.” At 69% activation, the same program was multiplying a base where most members had something to share.

The directories and partnerships were not wrong to run. They were producing real, in-ICP traffic. But they were producing it to a community that was failing 70% of the traffic in week one, degrading its own reputation in the directories it had listed in, and setting up partnership introductions that would disappoint the referred members and the partner operators who had sent them. Running those levers before activation was fixed was not neutral — it was actively counterproductive.

The three-touch sequence she eventually built took three weeks. She had known roughly what it needed to look like for the previous eight months and had kept pushing it to “after the community is a bit bigger.” The activation audit showed that during those eight months, 157 non-activating members had churned — members who, at 69% activation, would likely have stayed through month six. At $169/month each, for an average tenancy of three additional months, that was roughly $79,000 in foregone revenue deferred in favor of an optimization she “wasn’t ready for yet.”

She did not frame it as a mistake. She framed it as the difference between understanding a growth framework abstractly and understanding what it means for your specific community when your churn rate meets your acquisition rate at a derivable number you never calculated.

Three results worth tracking

Looking back at the eight months before the sequence and the twelve months after, three metrics changed structurally — not as one-time improvements that faded, but as changes to the baseline:

Week-one activation rate: 31% → 69%. The sequence was the mechanism. The intake form was what made personalisation at scale possible. Without the four-question intake form, the Day-0 DM would have been a generic welcome message that achieved roughly the same 31% as the welcome post in #intros. For the full sequence architecture — the Day-0 DM structure, the conditional Day-3 logic, and the Day-7 scorecard format — see why paid Slack communities lose members in week one and the Day-0/3/7 sequence walkthrough.

Monthly churn rate: 8.5% → 3.8%. The 4.7-point reduction was almost entirely structural: as the composition of the membership shifted from predominantly non-activating to predominantly activating, the weighted-average churn rate shifted accordingly. No individual intervention changed the experience for members after week one — the community’s content, events, and peer density did not change. What changed was the proportion of members who arrived in week one with a specific thread to read, a specific person to reply to, and a specific action that felt lower-stakes than posting a cold introduction to a workspace they had been in for two hours.

Net member growth: +5/month → +21/month. The leaky-bucket ceiling went from 259 members to 816 members. The community was no longer building toward an asymptote; it was building toward a ceiling that kept receding as the activation rate and referral multiplier improved together. For the mechanics of how retention compounds into net member growth over a twelve-month horizon — the specific math of why a 4-point churn reduction at 200 members is worth more than a 10-unit acquisition increase — see the growth and retention compounding case study.

Related: Slack community growth tips — the five levers and the right implementation order · Why paid Slack communities lose members in week one · How growth and retention compound over twelve months · The Day-0/3/7 sequence: what to write before the first invite goes out.