Overview

Available-to-promise: on-hand in ShipBob + confirmed inbound from Fulfil − already allocated − safety stock. Quarantine stock in Fulfil doesn't count — if QC hasn't released it, it doesn't exist for allocation.

Run per SKU, per channel, per location. DTC and subscription (Skio) share a pool. Amazon FBA is its own pool. Retail is committed against specific POs in SPS Commerce.

Retail first (chargebacks > DTC stockout cost), Amazon second (Buy Box recovery is expensive), DTC last (most flexible — we control the site, can show back-in-stock dates, can pause ads via the Media Buyer).

• DTC: 14-day cover based on Shopify/Skio trailing 30-day velocity
• Amazon: FBA minimum + 7-day buffer from Seller Central
• Retail: reorder point based on PO frequency from SPS Commerce history

Never promise un-released inventory. Never drop below a safety floor without human approval. When demand shifts — big PO, viral moment, Skio subscription spike — rerun allocation across all channels and flag anything below floor.

Build weekly forecasts by SKU, by format, by channel using Cogsy as the demand model.

For established SKUs (8+ weeks history): pull sales velocity from Shopify, Skio subscription base, Amazon Seller Central, and retail sell-through from SPS Commerce. Adjust for seasonality, the promo calendar, and planned ad-spend changes from Triple Whale.

For new SKUs (< 8 weeks): use the analog method. Find the closest existing SKU's launch curve in Cogsy, adjust for planned marketing spend. Flag wide confidence bands — new SKU forecasts are guesses.

Weeks 1–4 ≥ 95%, Weeks 5–9 80–85%, Weeks 10–13 65–75%.

Track forecast-vs-actual weekly in BigQuery. Variance > 15% for 2 consecutive weeks: flag and rerun. Big misses usually mean a promo overperformed, a competitor launched, an influencer post went viral, or an off-cycle retail PO hit.

Feed output to: inventory-allocation (reorder points), cash-forecasting (revenue projections), and production planning (run sizing via Fulfil).

Direct-method weekly cash flow: opening balance + receipts − disbursements = ending balance. Roll forward every week.

• DTC/Skio: same-day via Shopify Payments
• Amazon: 14-day disbursement cycle from Seller Central
• Costco: net-30 from invoice date
• Walmart: net-30 to net-60

Pull disbursements from Fulfil (PO payments), Ramp (card spend, ad invoices), and Rippling (payroll). Model PO payment cadence as 40% deposit / 60% pre-ship balance.

Key output: the cash-trough week — lowest ending balance in the 13-week window. Every PO approval and budget increase gets scored against what it does to the trough.

Run three scenarios: base, conservative (−15% receipts, +10% out), worst (−25% receipts, +20% out). If worst case breaches the floor, flag it even if base is clean.

• Any week below minimum operating cash in base scenario
• Trough drops > 20% week-over-week
• A PO or budget request compresses buffer below $100K above floor

Every batch starts in quarantine in Fulfil. Stays there until COA is reviewed and disposition is set. This is 21 CFR Part 111 — QC has sole release authority.

When a COA arrives (from Cohera Nutra QC lab, stored in Google Drive), check every analyte against vrün's internal spec — not the supplier's spec, ours.

• Actives: each ≥ label claim. HPLC for vitamins, ICP-MS for minerals, UV-Vis for standardized extracts.
• Heavy metals: lead, arsenic, cadmium, mercury — below the tighter of USP or Prop 65.
• Pesticide screen (99 analytes): all below LOQ via LC-MS/MS.
• Micro: total aerobic, yeast/mold within limits. Salmonella and E. coli not detected.
• Identity: confirmed via HPTLC.

Any single out-of-spec result fails the batch. No partial passes. Failed → reject disposition in Fulfil. Cannot be released without formal investigation and re-test.

For incoming ingredient COAs from suppliers (VitaSource, PectaCo, etc.): validate against vrün's ingredient spec. If a supplier COA says "conforms" without quantitative results, flag for independent testing before the ingredient goes into production at Cohera.

Track NSF status: GMP is facility-level (current/not). Certified for Sport is lot-level (290-substance screen). A lot can be COA-passed but CfS-pending — can ship but label accordingly.

• Any analyte out of spec
• Missing analytes on a COA
• NSF lot test pending with ship date approaching
• Supplier COA lacks quantitative results

A launch has dependencies. Track all of them. Nothing ships until every gate clears.

1. COA passed (via coa-compliance, Fulfil lot status = released)
2. Third-party validation confirmed (Eurofins results in Google Drive)
3. NSF status documented
4. Inventory received at ShipBob, QC released in Fulfil
5. Channel allocation locked (via inventory-allocation)
6. Shopify PDP live (hidden until go), Skio subscription configured
7. Amazon listing retail-ready in Seller Central (A+, Brand Story, FBA received)
8. Klaviyo flows built, tested, live
9. Postscript SMS sequence approved and scheduled
10. Creative assets approved (all passed claim review via creative-brief)

Default channel split: DTC 40% / Subscription 20% / Amazon 15% / Retail 7% / Reserve 13%. Adjust based on pre-order signal and retail PO commitments.

Pre-order window: 10–14 days on Shopify before broad push. Seeds reviews via Junip before paid media scales. If pre-order volume exceeds DTC allocation within 72 hours, trigger a reforecast in Cogsy and flag production planning.

• GO: all green.
• GO WITH CONDITIONS: 1–2 non-critical gates amber (CfS pending, creative not all approved). Document.
• NO-GO: any critical gate red. Delay.

The go/no-go decision is always human. This skill assembles the picture — it doesn't decide.

• MER floor: minimum blended marketing efficiency (total revenue ÷ total ad spend from Triple Whale)
• New-customer CAC ceiling: maximum per-new-customer cost
• Daily budget bands per channel in Meta Ads, TikTok Ads, Google Ads
• Pacing: ±10% vs weekly plan tracked against Ramp

1. Check guardrails across all active ad sets.
2. Healthy (ROAS above target, CAC below ceiling, freq < 2.5): scale budget up to 20% per step. Higher re-triggers learning phase.
3. Declining (ROAS dropping, CAC rising, freq climbing): reduce or pause. Check for creative fatigue.
4. Zombie (spent > 1x AOV, zero conversions): kill immediately.
5. Breached (CAC above ceiling or MER below floor): cap spend at last-safe level. Do not raise. Escalate.

Creative fatigue signals — watch in this order: CTR drops first, then frequency rises above 2.5, then CPM rises, then CAC spikes. By the time CAC moves, you've already wasted spend. When CTR drops 3–4 days straight with rising frequency, flag the creative. Check if ugc-sourcing has replacement cuts scored and ready.

After a creative swap, suppress scaling for 48 hours. Don't kill something that just re-entered learning. When a channel is capped or paused, redistribute budget to healthy channels with headroom. Keep total spend within ±10%.

• CAC ceiling breached 2+ consecutive hours
• MER floor breached blended
• No replacement creative available for a fatigued top performer

1. COA gates whether the product can be marketed (is the batch released in Fulfil?).
2. The approved claims library gates what we can say (has regulatory signed off on this claim?). The claims library lives in Google Drive, maintained by regulatory.

Every structure/function claim in any ad, Klaviyo email, Postscript SMS, Shopify PDP, or creator brief must map to the approved claims library. If it's not in the library, it can't be used.

• No disease claims. "Supports immune health" = structure/function (OK if substantiated). "Prevents colds" = disease claim (never).
• No comparative claims without substantiation on file.
• Required disclaimers on every asset.
• UGC creators are held to the same rules — a creator saying something we can't say is our liability.

• Check every claim against the library
• Flag anything not in the library — don't approve, don't assume
• Route flagged claims to regulatory

Score concepts: hook strength, clarity, claim safety, brand voice. Reject anything with an unsubstantiated claim regardless of creative quality.

Pipeline: find → vet → brief → receive → score → route or reject.

• Hook (first 3 seconds): stops the scroll?
• Clarity: value prop clear within 15 seconds?
• Claim safety: anything we can't substantiate? (check against creative-brief rules)
• Brand fit: matches voice, aesthetic, audience?
• Production quality: watchable? (doesn't need polish, needs to be clear)

Above threshold → tag by concept and funnel stage, route to the Media Buyer for testing in Meta/TikTok. Below threshold → send feedback notes to creator.

If a video has a potential health claim without substantiation: hold. Don't route to paid. Flag to Head of Brand. An unsupported claim in a paid ad is an FTC/FDA risk.

Keep the pipeline ahead of demand. Paid-pacing flags creative fatigue on winners — when it does, there should be 3–5 scored replacement cuts already waiting. Coordinate with paid-pacing to know what concepts are needed next.

Creator contracts must include paid-media usage rights (whitelisting, Spark Ads). A great video without usage rights can't be routed to the Media Buyer — flag for contract amendment first.

• Net revenue (after discounts) from Shopify, Amazon Seller Central, wholesale
• COGS landed from Fulfil (ingredients, co-man, packaging, freight) = CM1 (gross margin)
• Fulfillment from ShipBob, shipping, payment processing, returns = CM2
• CAC from Triple Whale (blended or attributed by channel) = CM3, the number that matters

When CM3 drops 3+ points below target for any SKU, flag it. Attribute the gap — pull data from Fulfil (COGS), ShipBob (fulfillment), Ramp (spend), Triple Whale (CAC), Shopify (discounts). How much of the gap is COGS, how much fulfillment, how much CAC, how much discounting, how much channel mix.

• Pricing: raise price, adjust subscription tiers in Shopify/Skio
• Bundling: multi-pack bundles amortize fulfillment + shipping
• Discount discipline: cap promo stacking, tighten welcome offer
• CAC compression: tighten MER floor, shift spend to Klaviyo/Postscript retention
• Co-man renegotiation: volume-tier pricing at Cohera
• Fulfillment: negotiate ShipBob rates, reduce dimensional weight
• Shipping: adjust free-ship threshold, zone optimization
• Channel mix: push toward subscription, DTC over wholesale

Model 2–3 scenarios: CM3 impact, volume risk, payback period.

This skill does not touch the product. Formulation, ingredients, and product design are not financial levers and are not in scope.

• CM3 below target 2+ weeks sustained
• Pricing changes proposed (always need human approval)
• Co-man renegotiation needed (involves procurement)