First installment of the recurring monthly case study series. One anonymized engagement per month — diagnostic, intervention, outcome. Names anonymized; numbers and timelines real.

The brand at intake

$14.2M D2C wellness/supplements. 22% YoY (down from 67%). 78% Meta-heavy paid mix. Reported CAC $94. AOV $74. Repeat rate 18%. Cash runway 11 months.

Stated problem: "Meta is broken. We need to pivot to TikTok." Actual problem: unit economics broken in ways no channel pivot could fix.

The Diagnostic (Days 1-30)

90-Day Audit — 36 checks across 6 zones:

Z1 Data & Attribution: 2/6 (P0). No CAPI deduplication. Meta overclaiming 100% vs back-end. Last-click. 14 overlapping events. No source-of-truth dashboard.

Z2 Acquisition: 5/8. Meta ASC at defaults (50% existing-customer cap, 8 creatives vs benchmark 15-50). AI search visibility 3/20 commercial queries vs competitor avg 11/20.

Z3 Creative Pipeline: 3/6. 14 variants in 90 days (benchmark 100-150). Zero AI tools. Avg creative age 87 days.

Z4 Conversion: 4/6. Mobile LCP 3.4s. 7 form fields. Mobile converting at 52% of desktop.

Z5 Retention: 1/5 (P0). Email at 6% of revenue (benchmark 25-30%). 2 active flows. No RFM. DKIM/DMARC missing. Inbox placement 63%.

Z6 Operating Model: 3/5. No marketing ops role. No four-metric dashboard. No CFO review. No test reserve.

True-cost CAC math: reported $94 + creative $8 + platform overhead $11 + first-order CS $15 = $128. Contribution per first purchase: $29.60. Net after first purchase: −$98.40. Only 18% bought again. Cohort broke down at month 3.

The Intervention (Days 31-90)

Rebuilt against binding constraints, in order:

Wks 1-2 · Data foundation. Stape sGTM across all paid channels. Reconciled Shopify to platform attribution via Triple Whale. Migrated to MMM + MTA + geo-holdout incrementality. Cleaned 14 events to 5. Meta's reported revenue dropped $4.2M → $2.4M (matching back-end). Every subsequent decision became correct.

Wks 3-4 · Retention engine. Klaviyo rebuild. Fixed DKIM/DMARC (inbox 63% → 91%). Built 5 core flows (welcome, abandoned cart + SMS, browse, post-purchase + replenishment, win-back). Deployed RFM + Klaviyo AI send-time. Email contribution 6% → 14% by end of month 1.

Wks 5-6 · Acquisition restructure. Meta ASC: existing-customer cap to 25%, Incremental Attribution on, audience signals from CRM, creative volume to 15-25 active per ad set. PMax split into 3 asset groups by margin tier. AEO foundational work — entity hygiene, schema, FAQPage on top condition content.

Wks 7-12 · Creative + conversion + ops. Deployed AI creative stack (Veo, Arcads, AdCreative.ai). Variants 14/90d → 38/30d. PDP LCP 3.4s → 2.1s. Form fields 7 → 3. Shop Pay promoted (mobile express checkout 14% → 42%). Migrated to four-metric dashboard + three reporting templates. Test reserve carved out at 5% of budget.

The Outcome (Day 90)

True-cost CAC: $128 → $76 (−40.6%) AOV: $74 → $89 (+20.3%) Repeat purchase rate (90-day): 18% → 31% (+13 pts) Email contribution to revenue: 6% → 22% (+267%) Contribution per first purchase: −$11 → +$34 (crossed zero) Mobile conversion rate: 1.4% → 2.6% (+85.7%) AI search citation share: 15% → 38% (+153%) Cash runway: 11 months → 31 months (+20 months)

Contribution margin flipped negative to positive per customer — and compounded across the customer base, extending cash runway from 11 to 31 months without raising capital. The board approved a Q3 spend increase to $1.4M monthly paid budget.

Three patterns worth internalizing

1. The CMO's instinct about the binding constraint is usually wrong. Most blame paid channels for unit economics problems. The actual constraint is almost always one or two zones deeper — data foundation, retention, creative pipeline, or operating model. Zone 2 gets blamed because it's the visible noise. The audit fixes the diagnosis.

2. Compounding across pillars is non-linear. One zone fixed: 10-20% improvement. Two: 30-50%. Four in parallel: 50-80%+ within 90 days. This brand didn't get to 40% CAC reduction by optimizing one thing — they rebuilt four interlocking systems simultaneously.

3. Reported metrics lie until reconciliation is built. Reported CAC was 37% lower than true-cost. Meta ROAS 2x higher than incremental. Every budget decision built on those numbers was wrong. The data foundation rebuild was the highest-ROI intervention — not because data is interesting, but because every subsequent decision became correct rather than systematically wrong.

When this kind of engagement makes sense

If your operation has any combination of: CAC trending up without obvious channel cause, contribution margin at or below zero, repeat purchase rate below benchmark, email under 15% of revenue, or leadership disagreeing on what the marketing numbers are — you're likely in the same pattern.

Start with the 90-Day Audit. Rebuild the zones scoring lowest. Many find execution easier with an outside team running operating-model installation alongside the in-house team. If that's the work you're considering, Praxxii Global runs this exact pattern across D2C, B2B SaaS, fintech, healthcare, and B2B services. Free 60-minute diagnostic call before any commercial commitment.