Chief Nutrition
Meta Restructure — AU/NZ + US
1. Why rebuild — AU
- Learning starvation: 2,707 purchases/30d over 187 ad sets. Avg ~3.3/wk per ad set vs ~50/wk to exit learning. Zero exit.
- Junk drawer: "Chief | AU | Collagen" = 89 ad sets, 63% of spend, frequency 8.9. Creatine, whey, beef bars, retargeting, NZ, demo splits all inside.
- Duplicated testing: 4 standalone TOF test campaigns ($24.7k) testing the same products. No promotion path. Legacy [DS] campaigns at ROAS 1.5–1.8.
- No real TOF: "Cold Awareness" restricted to lookalikes, frequency 4.3 — warm re-hits, not reach.
- Bar under-invested: collagen bar = 22% of store revenue on 9% of spend, and its buyers are overwhelmingly returning customers — paid isn't acquiring for the flagship.
2. Product economics — AU (30d)
| Product | Spend | % | Purch. | CPP | ROAS | Purch/wk | DTC rev share* |
|---|---|---|---|---|---|---|---|
| Collagen Powder (incl. creator/PPA ads) | $76,452 | 64% | 1,872 | $41 | 2.75 | 437 | 32% |
| Beef Liver / Organs | $11,821 | 10% | 224 | $53 | 2.01 | 52 | 7% |
| Collagen Bar | $10,899 | 9% | 202 | $54 | 2.17 | 47 | 22% |
| Creatine | $5,720 | 5% | 125 | $46 | 2.66 | 29 | 5% |
| Whey | $4,982 | 4% | 131 | $38 | 2.57 | 31 | 15% |
| Beef Bars / Sticks / Biltong | $2,244 | 2% | 37 | $61 | 1.60 | 9 | 13% |
| Bundles / Packs | $2,691 | 2% | 49 | $55 | 1.80 | 11 | 2% |
| Brand / Multi (legacy rtg, organic reels, DPA — retires) | $4,585 | 4% | 67 | $68 | 1.52 | 16 | — |
- * Shopify AU store net sales, same window ($1.08M AU+NZ product lines). NZ = 2.4% of store revenue.
- Powder's 64%-vs-32% over-index is partly by design — powder is the introductory hero; cross-sells (whey, creatine) ride in its baskets. Don't index paid to DTC revenue directly.
- Whey/creatine DTC shares inflated by cross-sell (bought with collagen). Whey 15% of store ≠ 15% standalone demand.
- Bar is the exception the cross-sell caveat doesn't explain: 22% of store, 9% of spend, buyers mostly returning = genuine acquisition gap for the leading hero product.
- Bar CPP ($54) depressed by recent out-of-stocks — true CPP likely better.
3. New structure — AU/NZ
- Two CBOs per market split by role: Hero (planning-call hero list, protected floors) + Range (everything else).
- Ad sets = product-based everywhere. Creator/partnership ads sit inside their product's ad set as creative.
- 15% TOF layer per account, three objectives per region — LPV (concept seeding + qualified clicks), AWR (reach / ad recall), ENG (video engagement → warm pools). All sales campaigns Advantage+ audience, lowest cost.
- Targeting: PUR ad sets broad · TOF campaigns (LPV / AWR / ENG) net-new audiences (exclude customers + engagers). Monitor audience breakdown, adjust from the data.
- Budgets anchor to current spend distribution (launch config, low delivery disruption) — not DTC revenue. Month-2 rebalance planned.
Min floor = ad-set minimum spend inside CBO (learning-exit guardrail ≈ 50 purchases/wk). Monitor audience breakdown on Net New lanes, adjust from the data. Estimates from observed 30d CPPs; expect elevated CPP while each new ad set rebalances (fresh containers — winning ads carry over with existing post IDs, renamed to convention).
Budget logic
- Total = current run-rate ($119.4k/30d ≈ $4,000/d). No increase assumed.
- Anchor = current spend per product. Not DTC revenue (cross-sell distorts it).
- Adjustments: NZ carved at ~its revenue share ($200/d) · Bar + Liver rounded up to $400 floors · Brand/Multi retired, ~$150/d reinvested into Range · 15% TOF carve split LPV / AWR / ENG per region.
- Powder absorbs every carve — only unfloored AU line, and the over-indexed one.
- Purchases/wk = budget ÷ observed CPP × 7. Forecast ~525/wk vs current ~630 — the ~105/wk gap is the deliberate 15% TOF investment (spend moved off PUR), not a mix-shift bet. TOF halo should show in blended MER within 4–6 weeks; the carve is a rebalance lever if it doesn't.
Where testing lives
- 4 standalone test campaigns retire. New creative (4–5 net-new concepts/wk) enters the matching product ad set directly.
- Big-swing concepts seed in TOF LPV; cheap qualified clicks → promoted into product ad set. One home, one promotion path. AWR + ENG run proven brand/hero videos only — they are distribution lanes, not testing lanes.
- Retargeting folds into Advantage+ sales. Standalone rtg + [DS] BOFU DPA retire.
4. US account — same disease, worse economics
US Meta spend ($22,012 USD/30d) exceeds the entire US store's net revenue ($20,492 USD). Meta-only MER 107% vs 20% target. Sales ROAS 0.49 · CPP $162 vs ~$79 AOV · 112 purchases across 59 ad sets / 7 campaigns (~26/wk account-wide). At ~60% margin, break-even ROAS ≈ 1.67 — US Meta is deeply contribution-negative today.
- Mirror-image mis-index: bars = 64% of US store revenue but 22% of spend. Powder (AU's hero) takes 37% + most of the mixed test spend. US customers buy bars; the account sells powder.
- Fake TOF, worse: "Cold Awareness" = $4,059 (18% of spend) for 1 purchase, split across four state-level 1% lookalike ad sets (NY / CO / CA / FL).
- Same sprawl: 3 overlapping creative-test campaigns, demo-split ad sets, three naming eras.
- Healthier customer mix underneath: 40% of US store revenue is new-customer (vs 28% AU) — younger market, acquisition works organically; paid just isn't converting it.
| US product (30d, USD) | Spend | % | Purch. | CPP | ROAS | DTC rev share |
|---|---|---|---|---|---|---|
| Collagen Powder | $8,184 | 37% | 46 | $178 | 0.44 | 28% |
| Collagen Bars | $4,805 | 22% | 30 | $160 | 0.53 | 64% |
| Mixed test ad sets (both products) | $4,100 | 19% | 33 | $124 | 0.59 | — |
| Awareness (lookalike-restricted) | $4,059 | 18% | 1 | — | 0.02 | — |
| Beef Liver · Biltong | $864 | 4% | 2 | $432 | 0.30 | 8% |
US structure — one consolidated campaign, bars-led, gate the spend
US learning exit is not the goal at this scale (whole account = 26 purchases/wk). The goal is CPP/ROAS repair via consolidation + bar-led creative. Everything else in the US account (Cold Awareness $4.1k → 1 purchase, 3 overlapping test campaigns, demo splits) turns off in migration. Hard gate below.
US spend gate: sales ROAS ≥ 1.0 by day 60, ≥ 1.7 (break-even) by day 90 — else pull US Meta back to maintenance and revisit after AU playbook matures. Client decision to pre-agree at kickoff, not mid-flight.
5. Migration — inbox/outbox, per product, hard cutover
- Containers rebuild from scratch (no recycled campaigns or ad sets). Winning ads carry over with their existing post IDs — social proof retained — renamed to convention on duplication. Everything else turns off as its product is processed: zombies die by omission, no separate cleanup task.
- Hard cutover per product, not duplicate-and-scale: ad-set learning resets either way; overlap self-competes in auction, splits learning budget, muddies the read. Safety net = pause-don't-delete (30d rollback).
- No platform changes before 22 Jul (Sensic-managed). Owner: Aaron · Casey oversight · Justin & Kalle sign-off at kickoff.
Phases
| Phase | Window | What happens | Gate |
|---|---|---|---|
| 0 — Prep | Now – 21 Jul | Tracking QA both accounts (pixel + CAPI) · map 187 ad sets → product buckets · winner shortlist + rename map · present to client | Tracking verified · client sign-off incl. US gate |
| 1 — Shells + TOF live | 22 Jul | Build all shells · launch the nine TOF campaigns (LPV + AWR + ENG × AU / NZ / US) · PUR sits ready | Shells QA'd · TOF delivering clean |
| 2 — First cutover | 22 – 24 Jul | Beef Liver hard cut — validates the mechanic before the crown jewels move | Clean delivery + tracking after 24–48h |
| 3 — Rolling cutovers | 24 – 29 Jul | Bar → Powder (only after bar + liver clean) → Range · US parallel · NZ once AU winners locked · creative cadence starts (4–5 net-new/wk) | Each product rebalanced 24–48h before the next · red gate holds the sequence, not the plan |
| 4 — Sweep + systemise | 30 Jul – 1 Aug | Legacy off · budgets reconciled · naming audit · SOP | 100% of spend in new structure · day-30 criteria tracking |
Per-product cutover runbook
The repeatable loop behind Phases 2–3 — run per product, in order; US runs the same loop in parallel.
- Identify winners — pull the product's inbox bucket; winner bar ≥$500 spend/30d AND ROAS ≥ 2.3; cap per ad set (Powder ~15–20 · Bar/Liver ~6–8 · Range 3–5).
- Duplicate into the new shell — use the existing post ID (keeps social proof + engagement history); rename the ad to convention on duplication.
- QA each ad — landing page hits the correct live PDP (no redirects, not OOS) · ad copy current (price, offer, claims) · UTMs on convention and firing.
- Check the ad set (once all winners are in) — every intended ad present and named right · targeting broad (net-new on LPV) · Advantage+ placements · min-spend floor set where specified · standard attribution (7d click / 1d view).
- Cut over — 3-part move, same session — new ad set ON → old product ad sets OFF (pause, don't delete) → old CBO budget stepped down / new stepped up by the product's daily share. Skipping the budget step lets the old CBO redistribute into unmigrated products.
- Rebalance watch, 24–48h — delivery clean, purchases tracking, no spend leakage into remaining old ad sets. Green → next product. Red → hold the sequence, diagnose; rollback is one click.
Specialist runsheet
Action-by-action. Campaign
→ Ad set. Budget = change to apply in that campaign. Pause, never delete.
| # | Action | Campaign → ad set | Budget | Notes |
|---|---|---|---|---|
| Phase 1 — 22 Jul · shells + TOF live | ||||
| 1 | Build all campaign shells to spec (§3 + §4) | All 12 new campaigns | — | PUR stays OFF · QA naming, broad/net-new, Advantage+ placements, 7d click / 1d view |
| 2 | Launch AU TOF stack | EN | AU | LPV | TOF · EN | AU | AWR | TOF · EN | AU | ENG | TOF→ Brand | Net New each | $270 / $180 / $120/d | LPV = highest-CTR ads · AWR = reach/ad-recall on proven brand + hero videos · ENG = ThruPlay on best video content · all exclude customers + engagers |
| 3 | Launch NZ TOF stack | EN | NZ | LPV | TOF · EN | NZ | AWR | TOF · EN | NZ | ENG | TOF→ Brand | Net New each | $20 / $15 / $15/d | Same creative picks as AU · fold ENG into LPV if delivery starves |
| 4 | Launch US TOF stack | EN | US | LPV | TOF · EN | US | AWR | TOF · EN | US | ENG | TOF→ Brand | Net New each | $40 / $30 / $20/d | Highest-CTR US ads, bars-led · benchmark-only |
| Phase 2 — 22–24 Jul · first cutover: Beef Liver (validates the mechanic) | ||||
| 5 | Duplicate Beef Liver winners in | EN | AU | PUR | Hero→ Beef Liver | Broad | — | 4–8 high-spend high-ROAS creatives, mixed content pillars · keep post IDs · rename · QA PDP / copy / UTMs |
| 6 | Turn Liver ad set ON | EN | AU | PUR | Hero→ Beef Liver | Broad | +$400/d | Set $400 ad-set min spend |
| 7 | Turn OFF old liver ad sets | Old structure (Collagen CBO + Beef liver campaign) | — | Pause don't delete · step old CBO down by liver's share so budget can't re-spread |
| 8 | Rebalance watch | Both structures | — | 24–48h: delivery + tracking clean, no old-CBO leakage · green → next, red → hold |
| Phase 3 — 24–29 Jul · rolling cutovers (24–48h between each) | ||||
| 9 | Collagen Bar cutover | EN | AU | PUR | Hero→ Collagen Bar | Broad | +$400/d | 4–8 winners · check every PDP for stock (OOS history) · set $400 min spend |
| 10 | Collagen Powder cutover | EN | AU | PUR | Hero→ Collagen Powder | Broad | +$1,730/d | Only after bar + liver clean · 15–20 winners incl. creator/PPA — partnership permissions must carry · no floor |
| 11 | Range cutover (all 4 ad sets) | EN | AU | PUR | Range→ Creatine Whey Beef Products Catch-all | Broad | +$700/d | 3–5 winners per ad set · Beef Products +2–3 net-new creatives · old test campaigns off as winners migrate |
| 12 | NZ live | EN | NZ | PUR | Hero→ Consolidated | Broad | +$150/d | Dupes of locked AU winners, all products |
| 13 | US cutover (parallel, from day 2) | EN | US | PUR | Hero→ Bars Powder Catch-all | Broad | +$510/d | Bars-led · US winners + AU bar/powder import batch · kill Cold Awareness, 3 test campaigns, demo splits |
| Phase 4 — 30 Jul–1 Aug · sweep + reconcile | ||||
| 14 | AU legacy sweep | Cold Awareness · [DS] · BOFU DPA · emptied old campaigns | — | Pause everything left, never delete |
| 15 | Budget reconciliation | All new campaigns | — | Confirm zero spend outside new structure |
| 16 | Naming audit + reporting views + SOP | Account-wide | — | Ads renamed on convention · views per campaign role · SOP → Brand/SOPs/ |
Earliest schedule, anchored to 22 Jul day-one. If holding the 90-day arc (M1 low-hanging fruit → M2 restructure), shift Phases 1–4 right 2–4 weeks; logic unchanged.
6. Risks
| Severity | Risk | Mitigation |
|---|---|---|
| High | Ad-set learning reset — dip while each product rebalances | Softened by post-ID reuse (social proof carries over). Sequenced per-product cutovers (liver validates the mechanic first, powder only after it's proven); 24–48h rebalance gates; pause-don't-delete rollback net; judge nothing before 7 days. |
| High | US stays contribution-negative | Pre-agreed day-60/90 spend gate. Bars-led rebuild + AU-proven creative is the one structural lever not yet tried. |
| Med | Floors squeeze Powder | By design — Powder is over-indexed and funds the TOF carve too. Only ~$800/d of $2,530 is floored. |
| Med | 15% TOF carve cuts tracked purchases ~15% short-term (~525/wk vs ~630) | Deliberate investment, pre-agreed with client. TOF halo expected in blended MER within 4–6 wks; if absent at day 45, step TOF back toward 10% (carve returns to Powder — one budget move). |
| Med | Range ad sets never exit learning | Accepted for product readout (see §7 open question). CBO can concentrate a winner out organically; whey+creatine merge available. |
| Med | Tracking continuity | Phase-0 pixel/CAPI verification both accounts; identical UTM convention old→new; weekly reconcile vs client tracker (MER target 20–25%). |
7. Success criteria — day 30 after cutover
| Measure | Baseline | Target |
|---|---|---|
| AU blended sales ROAS (30d, excl. TOF) | 2.60 | ≥ 2.75 |
| AU Hero ad sets out of learning (≥50 purch/wk) | 0 of 187 | 3 of 3 |
| AU weekly purchases | ~630 | ≥ 520 (parity net of 15% TOF carve) → growth from month 2 |
| AU frequency on hero lines (30d) | 8.9 | ≤ 4 (4–5 net-new concepts/wk) |
| AU campaigns / ad sets carrying spend | 10 / 187 | 9 / 14 |
| TOF (15% layer: LPV / AWR / ENG) | none (fake awareness) | CPLPV · CPM/ad-recall · cost-per-ThruPlay benchmarked wks 1–2, then −20% |
| US sales ROAS | 0.49 | ≥ 1.0 day 60 · ≥ 1.7 day 90 (hard spend gate) |
| US Meta MER | 107% | Trending < 50% · campaigns 7 → 4 · ad sets 59 → 6 |
| Naming convention | 3 conventions | 100% EN | {region} | {objective} | {campaign} |
8. Open decisions
- Range grain — segment or consolidate? Segment per product (as drawn) = per-product reporting + budget control, accept ad sets living in learning. Consolidate (1–2 pooled ad sets) = ~105/wk pooled, exits learning, readout drops to ad level. Context for the call: creatine = 4.8% of spend vs 4.9% of DTC revenue (fairly indexed — a control lane mostly protects against another spend-doubling incident); beef products = 1.9% of spend vs 12.6% of DTC revenue (most under-fed line — a readout lane is how it proves promotion to Hero).
- TOF objective — decided 17 Jul: 15% layer, three objectives (LPV + AWR + ENG) in every region. Still open: ENG optimisation event (ThruPlay vs post engagement — pick at build from available creative) and the internal split, reviewed at day 30 against the wk 1–2 benchmarks.
- Hero list client confirmation (Justin & Kalle, Tue/Wed) — blocking final lock.
- US spend gate sign-off — pre-agree the day-60/90 thresholds and the pull-back consequence at kickoff.
- Month-2 rebalance trigger: shift $200–300/d Powder → Bar once out-of-stocks resolve.
Sources: Supermetrics Facebook Ads ad-level pulls — AU act_10153388789321214 (933 rows, reconciles to $119,394 / 2,707 / $301,431 AUD) and US act_1116470807358962 ($22,012 / 112 / $8,841 USD) — plus Shopify AU + US store net sales via Supermetrics, all 14 Jun – 13 Jul 2026. AU product classification: content-tracker ad→product map + name keywords; US: ad-set name classification. CPP-based forecasts are estimates from observed 30d CPPs and labelled as such. This plan makes no changes to any ad account — all platform changes applied by the team in-platform per EN safety policy.