Every tracked "close" in the marketing stack is a sales close. This is the other ledger — each student's progress toward their own first commercial acquisition, the event the "we work free until you close" guarantee hangs on. The guarantee clock now runs on each student's own contract length (6–18 months, per Ramphis's system), not a flat 6 months, and Ramphis's RAG health signal is the primary per-student indicator. Wiki source of truth: projects/cashflow-code/references/student-outcomes-tracker.md
| Student | RAG | ICP | Contract | Guarantee | Success | Last act. | Documented stage | Proof | Notes |
|---|
has_churned=true); churn dates and per-student refund status are not in the API and need confirmation.| Student | Start | Contract | Outcome | Churn date | Notes |
|---|---|---|---|---|---|
| Daniel Taghavi | 2025-04-30 | 12mo (→2026-04-29) | Churned | not recorded | Was guarantee-active before the 7/04 reclassification. 0/40 modules; stalest last-activity of that group. Refund status not recorded — confirm. |
| Abelardo Martinez | 2025-04-30 | 12mo (→2026-04-29) | Churned | not recorded | Churned 7/04 — was guarantee-active. Also an LP→active-investor proof example (see Proof panel), now churned. Refund status not recorded — confirm. |
| Ting Li | 2025-10-17 | 14mo (→2026-12-16) | Churned | not recorded | Experienced (3+ prior deals); 0/40 modules. Never came back. Refund status not recorded — confirm. |
| John & Lisa Lopez | 2025-12-18 | 6mo (→2026-06-18) | Churned | not recorded | Couple (one membership). Refund status not recorded — confirm. |
| Marcos Soares | 2025-12-29 | 6mo (→2026-06-29) | Churned | not recorded | “Marcus & Priscilla” in the system (couple); 0/40 modules. Refund status not recorded — confirm. |
| Joe Mocquant | 2026-02-08 | 12mo (→2027-02-08) | Churned | not recorded | Churned per Georgy 7/04 (no contact). Had been corrected to a 12-mo contract earlier the same day before the churn call. Refund status not recorded — confirm. |
| Giselle Marzan | 2026-03-07 | 6mo (→2026-09-06) | Recovered → ChurnedDispute | not recorded | Recovered once, then churned. Did not ask for a refund — filed a payment dispute / chargeback (Georgy 7/04). |
| Starria Haigood | 2026-03-16 | 6mo (→2026-09-15) | Recovered → Churned | not recorded | Requested a refund ~3/23 — Georgy closed her back in (a recorded save) — then ultimately churned. Owns 10 doors; had a 28-unit target near a VA university. Whether the final exit included a refund is not recorded — confirm. |
| Carl Gonzales | 2026-03-21 | 6mo (→2026-09-20) | Churned | not recorded | Was in-window before the 7/04 reclassification. Refund status not recorded — confirm. |
| Umar Farooq | 2026-04-01 | 6mo (→2026-10-01) | Churned | not recorded (blank in API despite has_churned=true) | API-only — never on the CSV/roster before the 7/02 sync. Paid $10,800 in full (2-call close); no-debt constraint (religious reasons) per Georgy's 4/01 close call. The only churn the API structurally flags — no refund. |
| Karla Altamirano | ~2025-04-01 | 12mo (→~2026-04-01) | Churned | not recorded | 12-mo plan; started ~2025-04-01 (Georgy 7/04, approximate — “a long time ago”), contract now elapsed. Never in Ramphis's system. Never came back. Refund status not recorded — confirm. |
Alan D. (Case 1, $1.7M PR 7-unit) · Adrian ($150K raised, 66-unit) · STR→MF (Case 5)
Ismael (GC, $300K commitments — pre-success) · Pedro (electrical, pre-success sale — review in moderation)
Ismael — LP turned active investor, has a published Trustpilot (carries this angle). Abelardo made the same LP→active transition but has since churned — weaker as public proof. Plus Sean Hamilton (actual-success example, not a tracked student): 3 commercial closes in 4 months (TX/DR/PR) + Equis LP — his coaching testimonial still isn't public.
Pedro Ramirez — active investor who became a limited-partner investor after (Elevated / Jarabacoa villa buyer). His own Trustpilot is in moderation (appeal in progress).
Named win: Lesdier saved $14,000 in taxes since joining — clears the $10,000+ bar (lowered from $25K). A second named case with permission would strengthen it.
has_churned was true for only 1 of 11 real churns — the other 10 were added by hand. Give the CSM a churn flag with a required churn_date (right now churn_date is blank even for Umar, whose has_churned=true).exit_type enum {none, refund, dispute, …}.first_deal_closed is false for all 50 students — including Pedro, who has a documented sale. Model the two tiers we actually use: pre-success (sale / raise / LOI / tax win) vs actual acquisition, each with a deal reference + date. This is what discharges the guarantee.has_trustpilot_review / has_testimonial can't express "in moderation / appeal / hidden" — exactly the state Heitor's and Pedro's Trustpilot reviews are in. Model each proof asset as {type, status: published / in-review / hidden, url}.rag_status value. Two students (Luay, Zach) carry "orange", outside the documented red/yellow/green enum. Either document orange or normalize it — right now it can't be trusted to aggregate.start_date ≠ purchase date.start_date/end_date and stop deriving it two ways.churn_date/refund status for each. Ramphis follow-ups: onboard Sean (proof-only, no roster change) / Tania / Mateo into the system, set Pedro's first_deal_closed flag if/when the actual acquisition lands, confirm the non-standard rag_status="orange" value (Luay, Zach), and confirm per-student refund status for the churned list. On every ACTUAL success, fire the testimonial ask immediately.