Fashion D2C • Velvet & Co

$50k to $500k MRR

Velvet & Co, a sustainable fashion brand, was stuck at $50k MRR with rising ad costs. We deployed our AI Performance Engine to automate creative testing, predictive bidding, and omnichannel retargeting. Result: $500k MRR in 6 months.

ROAS

Revenue

CAC Reduction

Creative Velocity

The Plateau

Sarah Chen, founder of Velvet & Co, was frustrated. They'd been stuck at $50k MRR for 8 months. Ad costs were rising (CPA up 45% YoY), creatives fatigued within 72 hours, and they couldn't test fast enough to find winners. Manual ad buying was killing them.

CPA increased 45% year-over-year
Creative fatigue within 72 hours
Testing only 2-3 creatives per week
No unified attribution across channels

AI Performance Engine

We deployed our **AI Performance Engine**: automated creative testing (50 variants/day), ML-based LTV bidding, cross-channel retargeting, and unified attribution. The system connected to Shopify, Meta, Google, TikTok, and Klaviyo to optimize in real-time.

50 creative variants tested daily
Predictive LTV bidding
Omnichannel retargeting

What We Did

Data Integration & Audit

Connected Shopify, Meta Ads Manager, Google Analytics 4, TikTok Ads, and Klaviyo into a unified dashboard. Analyzed 6 months of data and discovered 60% of ad spend was going to low-LTV customer segments (AOV <$45). Identified top-performing creatives and audience segments.

Tools: Shopify API, Meta Conversions API, GA4, Triple Whale, Northbeam

AI Creative Engine

Built a custom AI system to generate and test 50 ad variations daily. Used GPT-4 for copy generation (headlines, body text, CTAs) and Midjourney + Runway for visual assets. Auto-launched winning variants on Meta with $50 test budgets. Reduced creative production time from 5 days to 4 hours.

Tools: GPT-4 API, Midjourney, Runway ML, Canva API, Meta Creative Hub

Predictive LTV Bidding

Implemented ML model to predict customer lifetime value based on first purchase behavior (product category, AOV, traffic source, time of day). Trained on 18 months of historical data. Shifted bidding strategy to optimize for predicted LTV instead of ROAS, allowing us to bid 3x higher on high-value customers.

Tools: Python, TensorFlow, BigQuery, Meta Advantage+ Shopping

Omnichannel Retargeting

Set up synchronized retargeting across Meta, Google, TikTok, and Email. Cart abandoners received 7-touch sequences within 48 hours: Email (15 min) → Meta ad (2 hours) → SMS (6 hours) → TikTok ad (12 hours) → Email (24 hours) → Google Display (36 hours) → Final email (48 hours). Recovered 34% of abandoned carts.

Tools: Klaviyo, Meta Pixel, Google Ads, TikTok Pixel, Postscript SMS

Dynamic Product Ads

Implemented catalog-based dynamic ads showing personalized product recommendations based on browsing history. Created 12 audience segments (by product category, price point, browsing behavior) with tailored messaging. Example: 'Still thinking about that dress?' vs 'Complete your sustainable wardrobe.'

Tools: Meta Catalog Manager, Google Merchant Center, Dynamic remarketing

Post-Purchase Upsell Flow

Built automated post-purchase email sequences to increase repeat purchase rate. Day 1: Thank you + care instructions. Day 7: Style guide. Day 14: Complementary product recommendations. Day 30: Exclusive discount. Increased repeat purchase rate from 18% to 67%.

Tools: Klaviyo, Shopify Flow, ReCharge (subscriptions)

6-Month Growth Journey

Month 1-2

Foundation

Data integration, audit, AI creative engine setup. Tested 300 creative variants. MRR: $50k → $85k.

Month 3-4

Optimization

Launched predictive LTV bidding, omnichannel retargeting, dynamic product ads. MRR: $85k → $250k.

Month 5-6

Scale

Implemented post-purchase flows, expanded to TikTok. System fully automated. MRR: $250k → $500k.

The Results

Return on Ad Spend

Industry avg: 2.5x

Monthly Revenue

From $50k in 6 months