Brand database · briefs/charm/BRAND.md · source of truth

Charm — Brand Database (BRAND.md)

Charm — Brand Database (BRAND.md)

The single source of truth for every durable fact about Charm. Any note David sends (inventory,

suppliers, accounts, contacts, decisions) is filed HERE the same turn it arrives — before any context

compaction. Memory files only POINT here. Never store secrets — reference env files. Hub renders this.

0 · Identity & entities

FieldValue
Legal entity / ownerCharm Enterprises LLC, Sheridan, WY 82801 (distributor on box). Purchasing entity on supplier PI: One Family LLC (PR phone +1-787-566-3729)
Key people & rolesJonathan Jesper — CEO, Jonathan@charmnutrition.com (never on camera per gate o1) · Jay Hunter — advisor (ex-CRO MaryRuth Organics) · Justin Cobb, Alex Bonilla, Craig Brocke — advisors · David — launch operator, sole creative verdicts
Support email / phonesupport@charmnutrition.com · 1-844-750-7272 (both printed on the 30ct box)
Brand relationshipPartner/client brand (Jonathan's company); David runs the DTC launch. David-only verdicts (2026-09-01)

1 · Products & SKUs

CodeProductSize / formatUPCFormula / spec sourceSupplier
CH-01PopFizz™ Vaginal Probiotics — "Fizzing Popping Probiotic Supplement", Berry Bliss30-stick box (30-day) · Net Wt 2.1 oz / 60g · 2g sticks · 40 boxes/carton8 60013 748?? 8 (on 30ct box art — confirm digits)clients/charm/PopFizz V-Probiotics/Formula/…CheerRock™ Pop Rocks-Intimate care_Berry flavor #2024040201_10Billion CFU (5).pdfHansford Biotech Co., Ltd. (Taichung, Taiwan) — Charles Hsieh (charleshsieh@hansfordbio.com), Chloe Wu (chloewu@hansfordbio.com, shipping/logistics); CheerRock™ pop-rocks base
CH-01-15PopFizz 15-stick box (15-day)80×40×145mm box · Net Wt 1.05 oz (30g) · 87 boxes/cartonseparate UPC (box art)same formulaHansford
CH-01-33ct sample pack (TBYB gift) — David 2026-09-013 sticks/set · 480 sets/carton · envelope dielines (Charm_dieline_envelope_100x155mm)—same sticksPRODUCED — 3,002 sets shipped (Hansford 2026-09-01)

Formula per 2g stick: HOWARU® Feminine Health = 10B CFU (2B Lacticaseibacillus rhamnosus HN001™ + 8B Lactobacillus acidophilus La-14®, IFF) · cranberry extract 90mg → 36mg Type-A PACs · D-mannose · vitamin C 100mg · 4.6 kcal · 0.9g sugar · vegan, GF, non-GMO, no artificial colors, dairy-free. Contains maltodextrin (never claim "no fillers"). Shelf life 24 months <26°C. Lab: Eurofins COAs (heavy metals, micro, LAB count, PACs) — 2025-08 lot, clients/charm/PopFizz V-Probiotics/COA/25-10/.

2 · Inventory & supply chain

SKUUnits on handLocation / 3PLLot / mfg dateShelf lifeCOGS (landed)MOQ / reorder lead timeUpdated
30ct box7,018 boxes SHIPPED (Hansford shipment notice 2026-09-01) · 40 boxes/carton · pallets 4–8 (175 full cartons + partial 18)Lansil Global (3PL) — receiving count at Lansil to confirmrun #1 (COA lot 2025-08, batch OD0536-07082025)24 mo from mfg$7.21 invoiced (Hansford final invoice, 7,027 boxes; PI 2025-04-14 said $7.00) · $0.240/stick · +$0.166/unit if the $3,000 storage fee is spreadTO FILL2026-09-15
15ct box7,993 boxes SHIPPED · 87 boxes/carton · pallets 1–3 (91 full cartons + partial 76)Lansil Global — receiving to confirmrun #1 (batch OD0534-07082025)24 mo$4.11 invoiced (8,000 boxes; PI said $3.90) · $0.274/stick — 14% dearer per stick than the 30ctTO FILL2026-09-15
3ct sample pack3,002 sets SHIPPED ✅ (exists — TBYB gift is real) · 480 sets/carton · pallet 9 (6 full cartons + partial 122)Lansil Global — receiving to confirmrun #1 (batch OD0582-10212025)24 mo$1.25/set invoiced (3,006 sets) · $0.417/stickTO FILL2026-09-15

Shipment math (verified): 15ct 91×87+76 = 7,993 ✓ · 30ct 175×40+18 = 7,018 ✓ · 3ct 6×480+122 = 3,002 ✓ · 9 pallets total · ≈ 339,441 sticks shipped (119,895 + 210,540 + 9,006).

Hansford FINAL INVOICE (David 2026-09-15, briefs/charm/docs/hansford-final-invoice-2026-09-15.png): OD0534 15ct 8,000 × $4.11 = $32,880.00 · OD0536 30ct 7,027 × $7.21 = $50,664.67 · OD0582 3ct 3,006 × $1.25 = $3,757.50 · storage fee $3,000.00 · gross $90,302.17 · less deposit received 2025-04-16 $38,900.00 · BALANCE DUE $51,402.17 (verified: lines sum). Invoiced quantities run 7 / 9 / 4 units above the shipment notice (8,000 vs 7,993 · 7,027 vs 7,018 · 3,006 vs 3,002) — reconcile against Lansil's receiving count. The invoice bills the shipped mix, not the PI's 10,000 × 30ct, so no 30ct balance is outstanding on it. Total product COGS $87,302.17 + $3,000 storage over 339,828 invoiced sticks ≈ $0.266/stick all-in.

Variance vs. the 2025-04-14 PI: PI ordered 2,000×15ct + 10,000×30ct; the shipment is 7,993×15ct + 7,018×30ct + 3,002×3ct — the mix shifted heavily toward 15ct and added the 3ct sampler. Treat the SHIPPED numbers as inventory truth; ask Jonathan/Hansford if a 30ct balance is outstanding.

Notes: source = Hansford shipment-quantity email, Chloe Wu → Jonathan, 2026-09-01 (filed by David 2026-09-03). Earlier compliance response said ≈14,000 units printed; label corrections planned for the NEXT print (statement of identity, SFP format, ingredient order, disclaimer box — CW review R1). Payment terms Hansford: T/T 50/50 or sight L/C, DDP. Shipping to customer: FREE, fulfilled by Lansil Global. → David's incoming inventory notes go in this section.

3 · Offer & pricing (locked — David 2026-09-01)

4 · Accounts & IDs (no secrets)

SystemID / handleNotes
Domain(s) / DNStrycharmnutrition.com (David-locked 2026-09-01) — still to register/attach; charmnutrition.com = printed support domain (resolves to a stub, controlled by Charm side)DNS → Netlify after registration
Netlify sitetrycharmnutrition · id 06e0ca9a-9afc-4e66-81e2-e170eac1a8c5 · team EcomScalers (teo-h3uxfci, Pro nf_team_pro) — moved 2026-09-28 from the personal free team (david-ecw5aq8) by POST /api/v1/sites/<id>/transfer at David's go ("move it to ecomscalers"); login david@davidsanghera.com; deploys via products/charm/deploy.sh (--skip-functions-cache)"charm" name globally taken; rename API needs body-wrapped form
29next store / campaignacm2.29next.store — the NailScience store, shared (David 2026-09-04). Campaign API key CAIlu40SJSami8E8OqhCx0pZGNOpkKtk087v9qTk (client-side, publishable) → creative/charm/29next-map.jsonCampaign 1877 'Charm Nutrition'. 9 PACKAGES LIVE (2026-09-07): 1 Mild $29.99 (was $24.99 → 2026-09-07) · 2 Moderate $39.99 · 3 Full $59.99 · 4 OTO Extra Box $19.99 · 5 Subscribe&Save $21.99/30d · 6 Try It First — 3 free sticks now, 30 sticks monthly: $0 → $21.99 on day 7 (7-day interval; webhook → 30d after renewal #1) · 11 Try It First — 3 free sticks now, 30 sticks on day 7: $0 → $29.99 on day 7 (webhook cancels after renewal #1) · 8 Shipping Protection $9.99 (order bump, store product 37/38) · 9 VIP 1 box $14.99 (was $12.49) + $9.99/30d · 10 Free Travel Sampler (3 sticks) — cart gift, $0.00 one-time on variant 112 (drawer unlock at $59; proven $0.00 on order 103598). (pkg 7, the membership-only VIP, was DELETED — superseded by 9.) WORLDWIDE: 244 countries in the checkout (ISO-3166 minus AQ + the four US comprehensive-sanctions destinations CU/IR/KP/SY). available_shipping_countries: [] means NO restriction, not US-only — proved with live GB/AU/DE orders (103539/103540/103541). Deleting showCountries from config.js kills the country field entirely; the list must be explicit. Order webhook 133 (order.created → https://trycharmnutrition.netlify.app/api/order-webhook) LIVE 2026-09-05 — the server-side Purchase; secret in Netlify env NEXT_WEBHOOK_SECRET, brand gate ORDER_SKU_PREFIX=CH- because webhooks are store-scoped and this store is shared with NailScience (whose own webhook 67 is still active — see asks). Shipping method default @ $0.00 attached to the campaign 2026-09-04 — without it the SDK throws "Shipping method 1 not found" and no order completes. Products CH-01 (109/var 110), CH-01-3 (111/112), CH-VIP (107/108) — all sellable. Ids → creative/charm/29next-map.json. Gateway = the NS store's — confirm it suits Charm.
Shopify store (checkout move, 2026-09-29)Charm · x9qttp-pe.myshopify.com · admin admin.shopify.com/store/x9qttp-pe · primary domain shop.trycharmnutrition.com · USD · Admin API 2026-07 · ships to 29 countries · fulfilment Manual (no 3PL app yet) · location "1309 Coffeen Ave"our app Charm Checkout (Dev Dashboard org 238222259, app 429827391489, client id 9c9be1bc2658b46dd16deb55678b7ff5, installed 2026-09-29) via the client credentials grant → 24 h Admin token; credentials in ~/.shopify/charm.env (outside the repo); builder products/charm/build_shopify.py → creative/charm/shopify-map.json. Also installed: 29next = 29next Shop Sync (products Shopify → 29next, 29next orders → Shopify; order-sync scopes only — not ours to widen), Facebook & Instagram, Google & YouTube, Messaging
Meta ad account · pixel (dataset)ad account 344424133382623 · pixel/dataset 3586199578185017 (David 2026-09-04)CAPI token stored as Netlify env FB_CAPI_ACCESS_TOKEN on the charm site + local backup ~/.charm.env (600). Never in git. Token lives on ASU system user.
GA4 · GTMGA4 LIVE (2026-09-04): property 552921247 'Charm' · measurement ID G-XQGMT5BKMY · web stream 15721727907 (defaultUri trycharmnutrition.com) — under the Nail Science account (accounts/391107531; GA4 accounts can't be created via API, and Charm already shares NS's 29next store — movable later if you want a standalone Charm account). Measurement-protocol secret created; both GA4_MEASUREMENT_ID + GA4_API_SECRET set as Netlify env. Bootstrapped best-in-class: 14-month retention, key events purchase/begin_checkout/lead_capture, 5 custom dimensions. GTM: IGNORE (David) — no container; pixel + CAPI fire directly from ch-attribution.jsSecrets in Netlify env + ~/.charm.env, never git
Klaviyo orgNONE YET — ask open
adscalerads client · Airtableclient 26 — Charm – PopFizz (CH-01 synthesized) (created 2026-09-07 via the adscalerads UI: brand kit plum/pink/coral + DM Sans, 8.8K-char synthesized brief, 10 profiles, 16 brand assets registered by URL, identity = the 30-stick box render) · client 27 — Charm – PopFizz (CH-01 PDP-brief) (2026-09-07 late: Brief C derived from the live PDP copy — tools/brand-hub/copy/charm/popfizz.json — same kit + 16 assets, its own 10 profiles) · raw full-brief client NOT made (killed angles — gate it first) · reuse portfolio Airtable base
Cookie / event prefix_ch_ (collision-checked 2026-08-30) · page code CH-01 · sessionStorage key ch_pdp_view_fired
Launch Console lanetools/launch-console/launches/charm/ (gitignored runtime) · hub workspace launch-center-hq.netlify.app/charm/

5 · Compliance ceiling

CL-00 (David 2026-10-06): this ceiling is the compliance officer's input on review, never a constraint on the first draft — the draft is written for conversion at full strength (copy-laws.md CL-00).

Structure/function only (DSHEA). Source = paid label review CW R1 (2025-10-21): no disease claims (yeast infections, BV, irritation, UTIs), "overgrowth" is a trigger word, "supports/maintains healthy X†" forms only; FDA disclaimer bold in a hairline box; statement of identity "Fizzing Popping Probiotic Supplement". Files: briefs/charm/compliance.md, forbidden-terms.txt (word-boundary the grep: "STI"→stick, "heal"→healthy false-positives), claims-cleared.md (approved bank + HOWARU® forms). Ad layer: no you-diagnosis, no intimate B/A imagery, no sexual-pleasure framing, institution names (Rutgers/IFF) in advertorial body only. HOWARU® = text mark only (no logo license). All IFF strain studies include lactoferrin → credit the STRAINS, never "PopFizz got these results".

6 · Decisions & taste log (dated — David's verdicts are ground truth)

7 · Asset map

📒 Full tagged index: [assets-index.md](assets-index.md) (76 entries — what each file is, where it can be

used, compliance status) · machine-readable twin assets-index.json · hub view /docs/charm/assets-index.html.

8 · Live surfaces

9 · Open questions / asks (console)

charm-identity-lock · charm-funnel-copy-review (Copy Studio, 5 live pages — replaces the withdrawn sales-letter/advertorial gate: David 2026-09-04 reviews copy only on live pages) · charm-tracking-ids · charm-29next-klaviyo · confirm Mild/Moderate/Full = 1/2/3 boxes · Lansil receiving counts vs. Hansford shipped (7,993 / 7,018 / 3,002)? · ~~is a 30ct balance vs. the PI's 10,000 still outstanding from Hansford?~~ (no — the final invoice bills the shipped 7,027) · pay Hansford balance $51,402.17 · Lansil receiving count vs. invoiced 8,000 / 7,027 / 3,006? · ~~3ct sample COGS?~~ ($1.25/set — invoice 2026-09-15) · couch-couple clip keep/kill · register trycharmnutrition.com · launch date + day-1 budget (deferred)

10 · Change log

2026-09-28 — trycharmnutrition moved to the EcomScalers Netlify team (David: "move it to ecomscalers"): one API call (POST /api/v1/sites/06e0ca9a…/transfer {account_id: 696867a8…}) from the personal free team (david-ecw5aq8, nf_team_dev) → EcomScalers (teo-h3uxfci, nf_team_pro); re-GET confirmed team + plan, the custom domain and env vars travel with the site; redeployed with --skip-functions-cache so the function bundle rebuilt under the new team. The 09-14 sweep had flagged the site as still-personal and §4 wrongly said EcomScalers until today.

2026-09-29 — Shopify store connected (read-only so far): products/charm/build_shopify.py report — 0 orders; one product created today, "PopFizz Vaginal Probiotics — 30 Sticks" (handle popfizz-vaginal-probiotics-30-sticks, 3 images, no description, on Online Store / POS / Shop) whose variants copy the RETIRED 29next tiers: One Month Supply — 1 Box $29.99 (popfizz_1) · Moderate — 2 Boxes $39.98 (popfizz_2) · Full — 3 Boxes $59.97 (popfizz_3); inventory tracked at 0 with CONTINUE, weight 0 lb. The live PDP sells 1 / 3 / 5 Boxes at $29.99 / $59.98 / $89.97 — the ladder question is David's (§9).

2026-09-29 — Shopify handoff built (dormant, SHOP.on:false): products/charm/tracking/ch-cart.js — the drawer's Secure Checkout (and the ?ob=c Buy One lane) builds a cart permalink on shop.trycharmnutrition.com (same root domain → Meta _fbp/_fbc + GA cookies carry over): tier → variant ids 1 Box 50550264397989 · 3 Boxes 50550264430757 · 5 Boxes 50550264463525 (product gid://shopify/Product/10794109304997, ids kept through the in-place align); UTMs as params + cart attributes, fbclid / gclid / ttclid / _fbp / _fbc / _landing as attributes, ref = utm_content (the AD### code) or utm_campaign; the trial lanes (plan=tbyb-*) keep 29next; ?co=29next forces the old path. While Shopify has no sampler variant the drawer drops the sampler promise. Smoke (WebKit phone, local files over the live domain, Shopify requests captured): 6/6 incl. one real permalink → Shopify checkout showing PopFizz $29.99 (no order; card payments + express checkout present).

2026-09-29 — Shopify apps (David: "go ahead and do 1,2 since its on my chrome logged in"): (1) the keys David pasted belong to the app "29next" = 29next's Shop Sync integration for acm2 (app URL + OAuth callback on shop-sync.apps-production.29next.com/acm2/; created by "Support Member" in the Charm Dev Dashboard org 238222259). Shop Sync syncs Shopify products → the 29next catalogue (it already made 29next products 136 + 138 from the Shopify PopFizz product) and 29next orders → Shopify for fulfilment; its config declares every scope but the store granted only an order-sync subset. David's call: a separate app, not a wider 29next grant → "Charm Checkout" created (app id 429827391489, client id 9c9be1bc2658b46dd16deb55678b7ff5, not embedded, App URL trycharmnutrition.com, scopes read/write products · publications · discounts · shipping, read orders · inventory · locations, unauthenticated read listings / read+write checkouts; version charm-checkout-1 released) and installed on x9qttp-pe (Settings → Apps: Charm Checkout · 29next · Messaging). Its client secret has to come from David (revealing it tripped the password-manager overlay). Every campaign-1877 package still points at the original 29next variants (110 / 112 / 108 / 38), so edits to the Shopify product never reach the live 29next checkout. (2) Facebook & Instagram (Meta) and Google & YouTube installed as sales channels; both account connections open the vendor's own sign-in popup → David finishes them: Meta → business owning ad account 344424133382623 · the EXISTING dataset 3586199578185017 (never a new pixel) · data sharing Maximum · the Charm Page; Google → the account holding GA4 552921247 (G-XQGMT5BKMY).

2026-09-29 (night) — Shopify store side ready for the ladder (products/charm/build_shopify.py, the Charm Checkout app): --align edited the PopFizz product IN PLACE (images / URL / variant ids kept) → option Package · 1 Box $29.99 · 3 Boxes $59.98 · 5 Boxes $89.97, compare-at $49.99 / $149.97 / $249.95, SKUs popfizz_1 / _3 / _5. --shipping: the store had been created with Philippine-peso defaults (Domestic = PH ₱165, free over ₱2,200; International incl. the US ₱1,200 ≈ $20 — every US order would have paid shipping) → both zones now Free Shipping $0 on any order ≥ $1 · Standard Shipping $6.99 on a $0 order (USD; the $1 floor keeps a discounted paid order free; only the sampler alone pays). --sampler: Free Travel Sampler (3 sticks) $0.00 (compare-at $7.50), SKU popfizz_sampler_3, variant 50550514843813, Online Store only — wired into the drawer (SHOP.gift). --storefront-token: public Storefront token 'charm-pdp' (kept in ~/.shopify/charm.env). Rate proof through the Storefront API (no checkout form touched): US 1 Box / 3 Boxes + sampler / 5 Boxes + sampler → Free Shipping; US sampler alone → $6.99; UK + Canada → Free Shipping. After David switched the store off the Philippines (USD, Los Angeles) a PH buyer can no longer buy (the line drops to 0) — only matters if PH is wanted. The API still reads the store time zone as Asia/Manila ("PST" = Philippine Standard Time) — check the Save. Handoff smoke 6/6 incl. a real permalink → Shopify checkout showing 3 Boxes $59.98 + the Free Travel Sampler FREE (no order). Drawer switch still OFF.

2026-09-29 (night) — Shopify checkout LIVE for the ladder: products/charm/tracking/ch-cart.js SHOP.on = true deployed (products/charm/deploy.sh, 1 file; /tracking/* is cached 1 h, so a visitor on the old drawer still lands on the working 29next checkout for up to an hour). Pre-deploy smoke 6/6; LIVE proof 4/4 on trycharmnutrition.com, no order — iPhone WebKit + desktop Chromium: PDP → 3 Boxes → Add To Cart → Secure Checkout → Shopify checkout with 3 Boxes $59.98 + the Free Travel Sampler FREE; ?ob=c Buy One → Shopify checkout 1 Box $29.99. Tracking: the checkout loads Meta's fbevents + signals/config/3586199578185017 (the right dataset); no event hits and no Google requests seen from our test browser (its IP geolocates to the Philippines — Shopify's consent handling may hold browser events) — confirm Purchase in Events Manager and GA4 Realtime on the first real orders. Revert = SHOP.on = false + deploy; per-visit fallback ?co=29next. Known: our drawer fires InitiateCheckout on the click and Shopify's Meta app may fire its own on the checkout (two ICs per checkout in Ads Manager until one is dropped).

2026-09-30 — Buy with Shop Pay on the PDP, the old checkout forwards the ladder, the full purchase-path matrix (LIVE 00:16 PT via products/charm/deploy.sh; David: "run the full qa matrix now with shop pay making sure all the combinations end up going to shop pay except for the trial" + "add this buy with shop button"):

2026-10-01 — Retro v1.33.0 + the Shopify checkout becomes the house default for PDPs (David: "launch-retro this and make the shopify checkout as part of our default setup on PDPs"). 11 learnings promoted (skills/product-launch/learnings.md); Phases 0 / 4 / 5 / 6 / 7 rewritten so a PDP's ladder checks out on Shopify by default (29next = trial / subscription lanes). The kit, built from Charm's code: tools/shopify_store.py (store side; Charm's spec creative/charm/shopify-spec.json; products/charm/build_shopify.py is now a shim onto it, same flags) · tools/shop_handoff.py + shared/shopify/shop-handoff.js (page side — Charm keeps its inline ch-cart.js; the module's links are byte-identical to it) · tools/qa/shop_handoff_check.py (36/36 on three engines, no checkout created) · tools/qa/shopify_matrix.py --brand charm (the generic live matrix — 16/16 on iPhone WebKit + desktop Chromium 2026-10-01, tools/qa/results/charm-shopify-20261001-1026/; Charm's 25-case runner imports its shared parts). Nothing deployed; Charm's live pages unchanged. The store tool's first read-only run found the shipping drift in §9 (one unconditional Free Shipping zone → the $0 sampler is orderable alone for $0; its product page answers 200, $0.00, available).

2026-10-01 — Subscriber cohort model on David's cohort table + Charm's first trial charges, and the analytics sync sees Charm again. David shared a cohort retention table (cohorts Nov 2025 – May 2026; averages 90.24 / 67.49 / 49.47 / 35.78 / 18.92 / 11.54% for months 1–6) — "use these … cohort averages and decays". Model: https://claude.ai/artifact/Doex81SMjd8a7Zqqp9zxfF (all four subscription brands; month 1 + the month-over-month decays from his table, months 7+ at 66.9% kept a month = the average of the month 2–6 decays, because his months 7–8 rest on one March subscriber). Charm per new subscriber, 12 months: 2.95 paid months · $61.55 revenue · $14.42 contribution ($4.89 a paid month: $21.99 less 5% refunds, $0.94 fees, $7.21 box, $7.85 fulfilment) · break-even CAC on renewals alone $13.38 by month 6. Today's house churn schedule (tools/proforma/assumptions/charm.json: 28.8 / 21.6 / 18% then 12.6%) gives 4.35 paid months · $90.90 — about 48% more than his table; the house files were NOT changed. 100 new subscribers a month → ~295 paying and $6,155 a month by month 12. Our subscribers so far (29next, synced 2026-10-01): 4 Try It First trials since Sep 22 — 628 and 859 cancelled before the day-7 charge; 661 and 694 were charged the first $21.99 box (orders 104687, 104753) and both charges were refunded the same day by a staff account (one subscription, 694, is still active: next charge Oct 31); 1 VIP membership active (662, $9.99, renews Oct 23). 0 of 2 charged first boxes kept (Sculptape: 0 of 2 too). Analytics fix: the sync had read 0 Charm orders since launch — on acm2 a line's SKU sits under stockrecord.partner_sku, not sku, and subscription list rows carry neither SKU nor orders. tools/analytics/analytics.py reads the SKU from either place and matches subscriptions through the brand orders that made them; Charm's lane rules (tools/proforma/assumptions/charm.json) now know the 'Try it For Free' and 'Buy 2 + Get 1 FREE' names (the old gift rule's 'Free' had been catching both).

2026-10-01 — The house retention assumption is David's cohort table, and Charm's subscriber LTV (David: "yes, use my table as the house assumption. So based on this, how much would the LTV be of an average charm subscriber? at 21.99 price point and at 24.99 pricepoint"). day-7 keep 70% → 90.24%, churn by cycle 28.8 / 21.6 / 18% → 25.21 / 26.70 / 27.67 / 47.12 / 39.01%, then 12.6% → 33.14% a month (the average of his month 2–6 decays); the VIP schedule is unchanged; the previous values stay in tools/proforma/assumptions/charm.json under previous_retention. The Launch Center's Analytics tab (Forecast + LTV) now runs on it; the Forecast also reads the schedule past its three sliders. Model: https://claude.ai/artifact/Doex81SMjd8a7Zqqp9zxfF. Charm LTV per new subscriber (a Try It First start, month 1 = 90.24%): about 2.97 paid months over a lifetime (2.95 within 12 months). At $21.99: $65.25 gross, $61.99 after 5% refunds, $14.52 contribution ($4.89 a paid month after the $7.21 box, $7.85 fulfilment and fees). At $24.99: $74.15 gross, $70.44 after refunds, $22.72 contribution ($7.66 a paid month). Per subscriber who pays month 1: 3.29 paid months, $72.31 gross at $21.99 · $82.17 at $24.99. Both prices assume the same retention; $24.99 still earns as much as $21.99 even if paid months fall 36% (to 1.90). Break-even CAC on renewals alone by month 6: $13.38 at $21.99 · $20.93 at $24.99. The model page has a price field for any other point.

2026-10-01 — Charm LTV with the real fulfilment rates (David: "$4 for shipping $1.20 for pick/pack ship"): a shipped box now costs $13.26 (was $15.06 on placeholders). Per new subscriber on the house retention (David's table, ~2.97 paid months): $21.99 → $65.25 gross, $19.86 contribution ($6.69 a paid month; break-even CAC on renewals by month 6 $18.30) · $24.99 → $74.15 gross, $28.06 contribution ($9.46 a paid month; $25.86 by month 6). Per subscriber who pays month 1: $22.01 / $31.09 contribution. Model page v3 + Launch Center analytics re-exported.

2026-10-01 — Packaging and inbound freight are included (David: "packaging is included and inbound freight is included"): both lines set to $0, so a shipped box costs $12.41. Charm LTV per new subscriber (~2.97 paid months on the house retention): $21.99 → $65.25 gross, $22.38 contribution ($7.54 a paid month; break-even CAC on renewals by month 6 $20.63) · $24.99 → $74.15 gross, $30.58 contribution ($10.31 a paid month; $28.18 by month 6). Chargebacks (0.5% × $25 = $0.12 a paid month) are not deducted; with them, $22.01 / $30.21. The Try It First day-0 order ($6.99 paid, $6.45 sampler + pick/pack + shipping, $0.50 fees) roughly breaks even and is not in the LTV. Model page v4 itemises every cost; Launch Center analytics re-exported.

2026-10-01 — Break-even at the CAC: at $60 a Charm subscriber never pays back (David: "so can we change price and cac on here to see when we would break even on that customer?" + "copy-chief this page so its easier for me to understand"). At the assumed $60 CAC (cac_assumed; CAC = ad spend per trial started), a new subscriber never pays back on subscription boxes alone: lifetime contribution is $22.38 at $21.99 ($37.62 short) and $30.58 at $24.99 ($29.42 short). Break-even CAC: $22.38 ever / $20.63 by month 6 at $21.99 · $30.58 / $28.18 at $24.99 (a $15 CAC pays back in month 3 at $21.99; $28 pays back in month 6 at $24.99). 100 new subscribers a month at a $60 CAC leaves the subscriptions $49,590 behind after 12 months ($41,381 at $24.99). Not counted: the trial order itself (about break-even), one-time ladder orders, VIP. Model page v5 (same link: https://claude.ai/artifact/Doex81SMjd8a7Zqqp9zxfF) opens on that verdict, with a CAC field next to the price, a payback chart and a "minus the CAC" column; copy-chief round 1 rewrote it answer-first in plain labels and dropped the "previous assumption" comparisons.

2026-10-01 — Refunds modelled at 2% (was 5%), and David's table checked against the market (David: "Is the cohort decay model too steep? what does a regular model look like for supplements?" + "Lets model the refunds at 2% instead of 5%"). assumed.refund_rate 0.05 → 0.02 in tools/proforma/assumptions/charm.json; a $21.99 box now leaves $8.20 after costs ($11.06 at $24.99). Charm LTV after costs per new subscriber: $24.34 at $21.99 · $32.80 at $24.99 (revenue after refunds $63.94 / $72.67). Break-even CAC: $24.34 ever / $22.43 by month 6 at $21.99 · $32.80 / $30.23 at $24.99. At the $60 CAC it still never pays back ($35.66 / $27.20 short); 100 new subscribers a month at $60 leaves the subscriptions about $47,630 behind after 12 months. Benchmark: Recharge's own platform data (getrecharge.com/blog/supplement-subscription-retention: ~1,800 supplement brands, first-time subscriber cohorts Jul 2025 – Jun 2026, order-sequence retention) has 86.6% making a 2nd order, 57.6% a 3rd, 33.8% a 4th, 9.8% a 7th and 1.4% a 13th, about 3.51 orders per paying subscriber. David's table, per subscriber who pays box 1: 74.8 / 54.8 / 39.6 / 8.5 / 0.8%, 3.29 boxes (6% fewer): a weaker first renewal, stronger months 3–4, a slightly steeper tail. So it is not too steep; on Recharge's curve Charm's LTV would be $25.95 / $34.97. Blog figures of 5–8% monthly churn measure a whole subscriber base, not a new cohort. The open risk is month 1: the model assumes 90.24% of trial starters pay for the first box (Recharge's 86.6% is people who already paid once), each 10 points off it costs $2.70 of LTV at $21.99 ($3.64 at $24.99), and the first 4 charged trials kept 0. Analytics re-exported; model page rebuilt (same link).

2026-10-01 — One-time orders (AOV) tab on the model page (David: "lets make this for AOV instead of Subscription as another tab next"). Same link, second tab: the ladder, the order mix, what an average order leaves after costs, and the share of buyers who must reorder to pay back the CAC. Charm, per order after 2% refunds, card fees, product ($7.21 a box), the free sampler on orders of $59+ ($1.25), $1.20 pick & pack and $4.00 shipping + $0.60 per extra box (an estimate; open check): 1 box $29.99 leaves $15.81 · 3 boxes $59.98 leaves $27.46 · 5 boxes $89.97 leaves $40.36. At the real mix so far (9 ladder first orders: 4 × 1 box, 5 × 3 boxes, 0 × 5 boxes; AOV $46.65) an average first order leaves $22.28, so at the $60 CAC no first order pays back (even 5 boxes is $19.64 short). A buyer has to place 2.7 orders' worth, which takes 63% of buyers ordering again after every order. Break-even CAC on the first order: $22.28 at today's mix, $40.36 if everyone took 5 boxes. Shopify so far (read-only aggregates, 2026-10-01): 4 orders, 1 cancelled; 3 paid ($149.95: 1 × 1 box, 2 × 3 boxes, AOV $49.98, nothing refunded). The Launch Center's analytics reads 29next only, so these Shopify ladder orders are not in it yet. Fixed: tools/proforma/assumptions/charm.json still priced the retired Mild/Moderate/Full ladder (ladder-2 $39.99, ladder-3 $59.99); now the BOGO ladder from §3 ($59.98 / $89.97, prices_note), so the Launch Center forecast prices ladder orders right. Analytics re-exported.

2026-10-01 — Three retention scenarios on the model page (David: "maybe we can have a different cohort decay model, this one can be Medium, an Optimisitic one, and a Pessimistic one"). Each is one row of his cohort table for months 1–6, then the shared 66.9% kept a month: Pessimistic = the May 2026 cohort (19 subscribers, the largest and latest: 89.5% pay box 1, 5.3% still paying in month 6) · Medium = the table's average (the house assumption) · Optimistic = the March 2026 cohort (7 subscribers, the best: 100% / 28.6%). Charm per new subscriber at $21.99: 1.90 / 2.97 / 4.43 boxes → LTV after costs $15.55 / $24.34 / $36.36 (break-even CAC by month 6 $14.68 / $22.43 / $31.64). At $24.99: $20.96 / $32.80 / $49.01 ($19.78 / $30.23 / $42.64). None pays back a $60 CAC: even optimistic at $24.99 is $10.99 short; on optimistic a $30 CAC pays back in month 6 at $21.99. The pessimistic row is the largest and most recent cohort. A Retention switch sits in the page's sticky bar, and the comparison table sits under the math strip (same link). Against the supplement market (David: "Which one of these is industry averages?"): Recharge's supplement curve gives about 3.51 orders per paying subscriber; Medium 3.29 (6% below, the closest to the industry average), Optimistic 4.43 (26% above: it matches the market for the first two renewals, 85.7% / 57.1% vs 86.6% / 57.6%, then holds far better), Pessimistic 2.12 (40% below). On the market curve, with Charm's 90.24% month 1, LTV after costs would be $25.95 at $21.99.

2026-10-01 — Medium = the supplement industry average (David: "make the industry average curve the Medium scenario"). The model page's Medium scenario is now Recharge's supplement retention (about 1,800 brands; 2nd order 86.6%, 3rd 57.6%, 4th 33.8%, 7th 9.8%, 13th 1.4%; orders 5–6 and 8–12 filled in geometrically). It starts from his table's 90.24% first box, because Recharge starts at a paid first order while Charm's first box follows a trial, and keeps 72.3% a month after the 13th (its 7th → 13th rate). Charm per new subscriber: 3.16 boxes; LTV after costs $25.95 at $21.99 · $34.97 at $24.99 (break-even CAC by month 6 $23.33 / $31.44); at the $60 CAC $34.05 / $25.03 short. Pessimistic (May cohort) and Optimistic (March cohort) are unchanged. His table's average ($24.34 at $21.99) stays on the page as a reference row: the Launch Center's house retention is still his table (not changed).

2026-10-01 — The economics gate (playbook v1.35.0, Phase 1c) on Charm: KILL at a $60 CPA (David: "add this to the launch-playbook as the second part after deepresearch and the angle studio, so we can quickly filter products"). Subscription: even optimistic retention never pays $60 back (LTV after costs $36.36). One-time: the average first order leaves $22.28, so 63% of buyers would have to reorder after every order. It would pass at a CPA under $23.33 (the industry curve by month 6). Page: https://launch-center-hq.netlify.app/charm/economics/ (rebuild: python3 tools/unit-economics/build.py --brand charm). Inputs added to tools/proforma/assumptions/charm.json: unit, the free sampler (gift_cost_per_order $1.25 on carts of $59+), and the $0.60 extra-box postage flagged as an estimate.

2026-10-01 — Break-even ROAS on a first ladder order: 2.09 (David: "put the breakeven ROAS as another box … AVERAGE ORDER PRICE / LEFT AFTER COSTS"): the $46.65 average order ÷ $22.28 left after costs. By tier: 1 box 1.90 · 3 boxes 2.18 · 5 boxes 2.23 (bigger orders carry more product cost, so they need a higher ROAS but leave more dollars). Below it, a first order loses money after ads. It's the fifth box on the model page's One-time orders tab.

Rendered 2026-11-16 from the git-tracked file. Edits happen in the repo (Claude files every fact David sends the same turn); this page is the read view.