Two membership campaigns run daily by Claude. Research changed this plan substantially from my first draft, including the order of operations and where the money most likely comes from. Every number is labelled with its source, and where I am guessing I say so.
Built 2026-07-24 · Window 2026-07-25 to 2026-08-23 · Target measured as MRR, not cash
Stripe, verified. The $97 kit has sold zero units.
Plus 18 founding members. The most valuable group in the system.
Of 154 total, 45 paid. Fewer real prospects than it looks.
Corrected 25 July from 885, which counted tag records rather than people. Mailable, real members, never campaigned. Plus 745 imported leads held separately.
Member counts from Circle Admin v2, re-read 2026-07-25. Revenue from Stripe, read 2026-07-24. AB reads 154 rather than the 156 in the first draft because Lucia and Logan were deleted on 24 July, not because members left. Paid counts are unchanged at 45 and 17.
Four existing FFL members upgrading from Orchestrator to Architect is $1,000 a month. No new leads required.
Orchestrator is $98/mo and Architect is $348/mo, a delta of $250 per upgrade. Four upgrades is $1,000 of new MRR from people who already pay you, already trust you, and already have a card on file. I had been aiming the entire engine at acquisition and ignoring this completely.
Expansion revenue converts far better than cold acquisition because payment friction and trust are already solved. Against 17 existing payers, landing 2 to 3 upgrades is a realistic month and 4 is a good one. It is now week one of the plan.
| Route | What it takes | New MRR | My read |
|---|---|---|---|
| FFL upgrades, Orchestrator to Architect | 4 of 17 existing payers | $1,000 | Best ratio available. Run it first and hard. |
| FFL founding members convert to paid (route removed) | Wrong on 24 July. 17 of the 18 founding members already pay. Verified 25 July. | ||
| FFL Orchestrator, new, from the 129 non-paying members | 10 new members | $980 | Best untouched list. Real members with logins who never got the campaign. |
| AB Pro at $99/mo, from the 104 non-paying members | 11 new members | $1,089 | Email 1 went out 24 July. Emails 2 and 3 are written and waiting on Send. |
| FFL Architect, new | 3 new members | $1,044 | Most revenue per conversion, hardest single ask. |
| Cold outreach to the 745 imported leads | consent decision first | unknown | Blocked, not sized. Bulk-imported 27 May, never engaged. See the correction below. |
Route table corrected 2026-07-25. Segmenting the owned lists showed the 24 July version oversold two routes. The FFL founding-member route is gone: 17 of the 18 founding members already pay, so the pool is one person. And "the 885" was never 885 people in the community. That figure counted tag records, and 760 of them carry no community member id at all, because they are imported leads rather than members. The real in-community non-member count is 126, of which 129 across all non-paying members are mailable after the audience guard. The 745 mailable leads are a separate, unqualified list.
Prices from the live agencybuilders.com and futurefocused.ai pricing tables, read
2026-07-24, and cross-checked against _docs/membership-levels.md, which agrees.
Conversion counts are my arithmetic on the rate bands sourced at the foot of this page.
Two corrections to my own earlier numbers.
One. I treated the 111 non-paying AB members as 111 prospects. Circle's platform data puts monthly actives at about 38% for communities of 51 to 500 (Circle). Against 156 people that is roughly 59 actives, and if most of the 45 payers are among them, the genuinely active free pool may be nearer 14 people than 111. The rest are dormant and convert at win-back rates of 2 to 5%, not community rates. Day one is therefore a segmentation job, not a campaign.
Two. The goal must be gross, not net. At category churn of roughly 6.5% a month for education and digital media (Recurly, own platform data, sample size undisclosed), a 35-member base loses 2 members a month. Gross new revenue has to clear about $1,300 to net $1,000. Without this rule the plan could report success while MRR sits flat.
What this card said on 24 July, and why it was wrong. It cited the June 21
audit at FFL-PAID-TIER-SEEDING-PLAN.md line 11, "Zero engagement in paid tiers
(Orchestrator, Architect, Insider all ghost towns)," and called it the top risk in the plan.
Re-checked against the live API rather than the June audit: Orchestrator and Architect both hold real content and are safe to sell. Sideline to MVP has 12 posts, the advanced Claude Code space has 6, Member Discussion has 5, and Monthly Group Calls holds recordings from 15 June and 20 July. Two Architect calls are on the calendar, 17 August and 14 September, so the upgrade ask can name a real date instead of a manufactured deadline.
The remaining gap is Insider at $1,498/mo (live futurefocused.ai pricing table, read 2026-07-24): VIP Lounge and Mastermind hold one post each and are genuinely empty. Insider is a conversation sale rather than an email one, so it does not block this plan. Run no Insider push until those two spaces have real content. The original point still stands where it applies: selling into an empty room produces churn at the first renewal.
Section AI charges $62.50/mo billed annually for unlimited live and on-demand courses, certificate programs, office hours and AI coaching (sectionai.com/pricing, checked today). That is a known brand offering more than Orchestrator for 36% less. Most paid communities sit at $26 to $50/mo.
No framing makes "three courses and a community" worth $98 against that. What does justify it is access to you rather than access to content, and the specificity of the room. I would reframe Orchestrator around a response guarantee and the member directory, and anchor the price against executive AI training budgets rather than against other courses. Separately, the $98 to $348 gap is 3.55x with no rung between. A tier near $198 would make the goal reachable from a much larger willing pool.
36 of 62 paid members across both brands have not logged in for 14 days or more.
Agency Builders is 29 of 45, Future Focused is 7 of 17. Read from last_seen_at on the
Circle Admin API, so this is genuine last-login data rather than inference. Nobody has cancelled
yet, which is the point: this is churn forming, not churn that has landed.
This was invisible when the plan was written on 24 July, because the plan measured member counts and not member behaviour. It matters more than any single campaign in this document for a simple reason: a monthly target compounds only if members stay. At roughly 6.5% category churn on a 35-member base you lose about 2 members a month, and gross has to clear about $1,300 to net $1,000. That arithmetic gets worse, not better, if two thirds of the base is already drifting.
The running order changes. Week 2 as written turns on the Agency Builders acquisition engine. Pouring new members into a brand where 29 of 45 payers have stopped showing up is the wrong sequence: it buys members who arrive into a quiet room and leave at the first renewal. Retention work at D27 is far too late to be useful. Recommend moving the AB retention audit ahead of the AB acquisition push, keeping the Future Focused upgrade path first either way since it is unaffected and remains the fastest route to the number. That is a sequencing call for Jay, not a change I have made to the calendar below.
| Capability | Status | What it unblocks |
|---|---|---|
| Circle Admin v2, both communities | Working | Reading members and tags, and posting into spaces. Verified today by publishing three real posts. This is how the daily in-community work runs. |
| Circle member DMs | Unverified | The 1:1 sequences are among the highest-return tactics. I have confirmed posting but not DM sending. Day one check. |
| Circle email broadcast | No API | Confirmed 404 across four endpoint variants. Community email is UI-only, so every broadcast needs your hand on Send. |
| Cloudflare Pages and D1 | Working | Deploys plus a database, so I can build funnel analytics myself with no new connector. |
| Funnel analytics | Missing, mine to fix | Neither landing page has any analytics at all. I have been working blind on traffic and conversion. Built on D1 in week one. |
| Gmail send | Working | Right for the 23-lead drip. Wrong for hundreds, so I will not use it that way. |
| HubSpot, portal 485253 | Effectively unusable | 15,111 contacts, but only 14 are sendable today. Detail below. |
| Social publishing | Needs reauth | Top-of-funnel and workshop promotion. Costs you one reconnect. |
| Stripe | Read working | Revenue truth. Memberships bill through Circle, so paid-member counts lead and Stripe confirms. |
| ManyChat to Instagram | Disconnected | Comment-to-DM. Needs you to refresh Instagram permissions in ManyChat. |
| Slack | Needs auth | My daily written update reaching you without you coming to look for it. |
What I need, in priority order.
On using Business Builders HubSpot for the FFL promotion. I checked it and it
will not work as it stands. Of 15,111 contacts, 535 are designated marketing contacts and 521 of
those have unsubscribed from all email, leaving 14 sendable addresses. A real
pool sits underneath, 12,348 with an email and no opt-out, but reaching them means paying HubSpot
to designate them first, and most large lists in there carry 2019 and 2020 dates. Worth noting on
its own: you appear to be paying for 535 marketing contacts of whom 521 can never be emailed.
Recommendation is to prove the offer on the 885 FFL Circle contacts, where consent matches the
offer and the cost is nothing, then decide on HubSpot spend against a known conversion rate.
Detail in campaigns/ffl-membership/send-plan-hubspot.md.
The approval batch. Every draft that needs your yes, in one message, at a fixed time. Upgrade asks, DMs, social posts, broadcasts. This goes first because the tactics that actually produce revenue all sit behind it, and if it slips to the end of the day the plan degrades into autonomous busywork that earns the least.
log.md and redeploy the dashboardOne hard rule on the ask. I am forbidden from making an upgrade ask to any member below an activation threshold. A member who has never posted, never opened a lesson and never attended a call has no basis on which to price a $98 tier, so they price it against nothing and correctly decline. For the dormant group the 30-day job is getting them to act once. Any revenue from them is a bonus, not the plan.
By brand and tier. An annual prepay at two months free is $1,188 in cash but $99 of MRR, and it counts at $99. Gross must clear about $1,300 to net $1,000.
Share of members who complete a first action, then conversion by segment. Activation predicts conversion, which is why the ask is gated on behaviour.
Eleven members won while the sending domain burns is a loss. First-year members renew at 75% against 84% overall, so onboarding runs from day one.
Honest probability. Research puts this at roughly a 55% chance of $1,000 net new MRR in 30 days if the existing-payer upgrades and founding-member conversions run first and hard, and under 20% if the plan leads with cold-list reactivation and content production. My first draft led with cold-list reactivation. That is the single biggest change research made, and it is why week one now points at 35 people who already know you rather than 885 who barely do.
Defensible. ChartMogul SaaS Conversion Report (real product data, card-required vs opt-in trial bands) · Marketing General 2025 Membership Benchmarking (17th annual survey, first-year renewal 75% vs 84%) · ON24 2025 (registration to attendance 57%, vendor platform data) · Recurly (category churn, sample size undisclosed) · Circle (38% MAU band, Circle is self-interested) · Section AI pricing (checked directly today) · Baremetrics (annual vs monthly retention).
Rejected as unsourced. Research surfaced and I am deliberately excluding: "$1 in community returns $6.40", "communities grow revenue 2.1x faster", "annual plans reduce churn by 51%", "founding member groups reach 90% conversion", "targeted messaging reactivates up to 45% of inactives", and several conversion figures from creator-economy content sites that cite each other in a circle with no primary methodology. None of these appear anywhere in this plan, and none should go into member-facing copy.
My own inference. All segment sizes pending day-one segmentation, all conversion counts, and the gross-to-net calculation. Treated as estimates, not forecasts.