How We Fixed a "Leaky Bucket" Retention Problem and Added

The Challenge: The "Leaky Bucket" Problem

What They Do

Luminary Labs sells premium athleisure apparel targeting active women aged 25-40. Their bestseller (the "CloudFit Leggings") retails for $98 and has a 4.9-star rating with 12,000+ reviews.

The Problem: A Broken Retention System

Luminary Labs was great at acquiring customers—but terrible at keeping them. Their acquisition machine was working perfectly, but their retention bucket had multiple leaks:

What We Discovered in Week 1

• Broken email automation: Their Klaviyo flows had critical errors—63% of customers never received post-purchase emails
• No predictive timing: Win-back emails sent at random intervals (30, 60, 90 days) instead of based on actual purchase patterns
• Generic messaging: All customers got the same "We miss you" email, regardless of what they bought
• $127k opportunity: We calculated that fixing these issues could recover $127k in the first 60 days alone

Email Flow	Status	Impact
Post-Purchase Thank You	✓ Working	—
2nd Purchase Nudge (Day 30)	✗ BROKEN	$47k/month lost
Replenishment Reminder	✗ BROKEN	$38k/month lost
Win-Back Campaign (Day 90)	Partially Working	$42k/month missed

Our 3-Part Retention Fix

Fix #1: Repair the Broken Email Automation

The Problem: Due to a Klaviyo segment configuration error, 63% of customers were excluded from post-purchase email flows.

What We Did:

• Fixed the segment logic error (it was filtering out customers who had "placed 1 order" instead of "placed at least 1 order")
• Rebuilt all 4 core retention flows from scratch:
  • Flow 1: Post-purchase education (Day 3, 7, 14)
  • Flow 2: 2nd purchase nudge (Day 28-35, personalized by first purchase)
  • Flow 3: Replenishment reminder (based on product type)
  • Flow 4: Win-back campaign (60-90 days)
• Added A/B testing to every email (subject lines, timing, offers)

Fix #2: Predictive "Next Purchase Date" Models

The Problem: Luminary Labs sent win-back emails at fixed intervals (30, 60, 90 days) regardless of what customers bought. But leggings buyers reorder every 4-6 months, while sports bra buyers reorder every 8-10 months.

What We Did:

• Analyzed 18 months of purchase data and segmented by product category:
  • Leggings: Average repurchase cycle = 5.2 months
  • Sports Bras: Average repurchase cycle = 9.1 months
  • Tank Tops: Average repurchase cycle = 7.4 months
  • Accessories: Average repurchase cycle = 12.3 months
• Built dynamic email triggers based on product + order value:
  • Leggings buyer → Send replenishment email at Day 140 (4.7 months)
  • Sports bra buyer → Send replenishment email at Day 260 (8.7 months)
  • Multi-product buyer → Send at earliest predicted date with bundle offer

Fix #3: Automated Win-Back Campaigns

The Problem: Customers who didn't repurchase on schedule were lost forever—no follow-up campaigns existed.

What We Did:

• Created a 3-email win-back sequence triggered when customers pass their predicted repurchase date by 30 days:
  • Email 1: "Did your leggings wear out? Here's 20% off a replacement pair"
  • Email 2 (+7 days): Customer story + social proof ("Join 47,000 women who swear by CloudFit")
  • Email 3 (+14 days): Last chance urgency ("Your 20% off expires in 48 hours")
• Launched parallel Meta retargeting for customers who opened emails but didn't buy
• Created a "VIP comeback" offer for customers who spent $300+ in their first order

The Results: 19% LTV Increase

Metric	Before OFFSAGE	After 90 Days	Change
Customer LTV	$287	$342	+19%
Repeat Purchase Rate	22%	41%	+86%
Avg. Purchases per Customer	2.02	2.41	+19%
Email Revenue (per month)	$38,000	$94,000	+147%
Win-Back Campaign Revenue	$0	$127,000	+∞
Total Revenue (90 days)	$2.1M	$2.47M	+18%

What Changed, Specifically?

• Fixed the broken email automation → 63% more customers entered retention flows
• Built predictive repurchase models → Win-back emails sent at optimal timing (not random)
• Personalized by product category → Leggings buyers got different emails than sports bra buyers
• Created urgency-driven sequences → 3-email win-back campaigns (not one-off blasts)

Key Takeaways for DTC Brands

• Audit your email automation quarterly: Luminary Labs had a broken flow for 8 months and didn't notice—it cost them $376k
• Segment by product, not just demographics: A leggings buyer has a different repurchase cycle than an accessories buyer
• Win-back campaigns should be sequences, not one-offs: 3 emails convert 4.2x better than 1 email
• Retention is cheaper than acquisition: Recovering 1 churned customer costs $0. Acquiring 1 new customer costs $87-$210.

---