Koen Van Lysebetten

Cases · two chapters

Work that shipped.
Numbers that hold.

Regulated AI now — and the growth systems that trained the judgment. Same person. One arc.

168%GIMBER MRR / 6 mo
25%→8%VTM GO churn
6k+/dayNajm cases
$10M+India AI/AR/VR
Trusted by brands across industries
GIMBERVTM GOBelgium RetailerJamalon MacmillanSMALSDevGapBlueprint Labs NMML / IndiaUS Restaurant Intel

Case detail

Challenge → approach → receipts. Same transferable logic across industries.

E-commerce / DTC beverages

168% growth through systematic revenue leak detection

GIMBER

Challenge

Underperforming conversion rates and high churn. No systematic approach for identifying and fixing revenue leaks. Flying blind without attribution systems.

Approach

CRO-led growth loops with data-driven optimization. Systematic identification of leakage points through cart abandonment recovery, email flow optimization, pricing strategy refinement, and checkout friction elimination.

Revenue growth
168%
Timeframe
6 months
MRR
+168% / 6 mo
Team
7 growth marketers
CROEmail automationDynamic pricingAttribution

Full-funnel growth seat: find leaks, fix the system, measure MRR.

Streaming / Media

VTM GO: churn from 25% to 8%

VTM GO (DPG Media)

Challenge

VTM GO wanted to reverse engineer Netflix’s market domination for Belgian streaming. Focus on retention through full-funnel lifecycle optimization. High churn rate at 25%.

Approach

Complete Netflix strategy reverse engineering: multi-platform customer journey mapping, sophisticated segmentation on Snapchat/TikTok/Facebook, and retention-focused lifecycle automation across all funnel stages.

ROI
2,380%
Churn
25% → 8%
Timeframe
1 year
Reach
Full-funnel
Lifecycle automationMulti-platform retargetingBehavioral segmentationNetflix tactics

Lifecycle system for Belgian streaming — retention as the product.

Retail / FMCG

Omnichannel orchestration: CAC reduced 35%

Belgium’s biggest retailer

Challenge

Fragmented customer journeys with multiple touchpoints. No integrated communication strategy. High CAC at $120+ with poor retention and engagement.

Approach

Full-funnel Customer.io implementation with behavioral triggers, advanced segmentation, and automated lifecycle flows for food discovery and retention.

CAC
$120 → $78
Engagement
+40%
Conversion
+32%
Journey
Unified
Customer.ioBehavioral triggersLifecycle automationSegmentation

Omnichannel orchestration principles that travel across industries.

E-commerce / Publishing

3× revenue across 22 MENA countries

Jamalon

Challenge

International expansion across 22 diverse markets with different languages, cultures, and customer behaviors. Scaling complexity through traditional methods was unsustainable.

Approach

Systematic testing framework adapted to local context while maintaining core methodology. PRIOR → FIRE → WIN cycles replicated the methodology, not tactics — proving frameworks transcend geography.

Revenue
CAC
−35%
Repeat rate
+40%
Markets
22
Universal frameworkPRIOR-FIRE-WINGeographic scalingLocalization

Framework transcends cultural and linguistic barriers.

Publishing / Media

From intuition to data-driven growth engine

Macmillan Publishing

Challenge

Publishing house wanted to shift from ad-hoc marketing to systematic growth. Decisions based on intuition, no structured approach for experimentation or validation.

Approach

Validation Loop methodology with GIF Challenge experiment design. Systematic hypothesis formation, rapid testing, and data-driven decision-making through structured PRIOR → FIRE → WIN cycles.

Decision velocity
10× faster
Experiment cycles
2-week sprints
Framework
Standardized
Impact
Measurable
Validation loopsExperiment designGIF ChallengeData-driven decisions

Structured experimentation became their standard operating procedure.

Government / Public sector

4-layer customer success lifecycle for government

SMALS (Belgian government)

Challenge

SMALS had no structured customer success lifecycle for B2C citizen interactions. Lack of systematic approach for citizen journey optimization led to inefficient service delivery.

Approach

4-Layer Framework with GROWS methodology: Customer Success Phases, Strategic Goals, Tactical Actions, and Measurable Metrics for each phase of the citizen lifecycle.

Sector
Gov / citizen services
Process
Streamlined
Engagement
Improved
Service quality
Measurable
4-layer frameworkGROWSLifecycle mappingGovernment B2C

If systematic frameworks work in government bureaucracy, they work anywhere.

Government / Cultural heritage

$10M+ AI/AR/VR museum digital transformation

DevGap × Nehru Memorial Museum & Library

Challenge

$10M+ government contract for complete digital transformation of India’s premier memorial museum. Complex scope: AI chatbots, AR/VR experiences, LED installations, cultural preservation technology.

Approach

Enterprise-level project management with large cross-border teams (DevGap UK + India, Tagbin). AI-powered storytelling with interactive conversations with historical figures. AR reconstructions of Indus Valley Civilization.

Contract value
$10M+ USD
Technology
AI / AR / VR
Impact
Digital preservation
Team
Cross-border
AI chatbotsAR/VRLED systemsCultural preservation

Enterprise government delivery with $10M+ budget discipline.

Food tech / Data intelligence

Three agents, human gates: ops cycle in minutes

U.S. Restaurant Intelligence Platform

Challenge

A leading restaurant intelligence company relied on 200+ human operators to extract menu data from thousands of restaurants monthly. Manual scraping created bottlenecks, month-long processing cycles, and ballooning costs.

Approach

Designed and deployed a fully autonomous GenAI pipeline powered by three specialized agents: Web Scraper (headless browsing + NLP), PDF Parser (multimodal LLMs), and ML Uploader (schema-aware validation) — with human-in-the-loop operating model for non-technical teams.

Processing
1 mo → 10 min
Team size
200 → 3 agents
Cost
−90%
Data quality
High reliability
Anthropic ClaudePythonHeadless browsersServerless

Reusable blueprint for enterprise GenAI automation initiatives.

Developer tools / AI

Strategic market analysis: $5B+ opportunity sizing

Stealth AI platform

Challenge

AI coding tools had solved speed-to-prototype but none addressed the “iteration cliff” — NPS drops 20+ points when users modify generated code. That architectural limit caps TAM at $17B vs $85B potential.

Approach

Comprehensive competitive intelligence across Lovable, Cursor, Devin, Windsurf, and Bolt. Mapped threat matrix from Microsoft Copilot bundling, Google Opal, and Claude reasoning capabilities. Identified strategic windows.

TAM identified
$17B → $85B
Market window
6–9 months
NPS delta mapped
15–60 pts
Threat timeline
2026 Q1–Q3
Competitive intelligenceMarket sizingThreat analysisStrategy

First-mover framework for capturing the iteration workflow gap.

Venture building / DTC

Healthy Herbs validation: 3.48× ROAS in 14 days

Blueprint Labs

Challenge

New venture in herbal supplements needed rapid demand validation before committing to inventory and brand development. Traditional market research would take months.

Approach

MicroTest™ Validation Engine with systematic A/B testing, unit economics modeling, and real-time optimization. 14-day sprint to prove product-market fit with minimal capital.

ROAS
3.48×
CTR
4%
Gross margin
68%
CPC
€1.01
MicroTest™Unit economicsA/B testingMeta ads

Validated at 3.48× ROAS before any inventory investment.

Why this works across industries

Universal principles that transcend boundaries — proven across e-commerce, government, SaaS, media and publishing.

01 · Frame Hypotheses, baseline metrics, success criteria, experiment design.
02 · Ship Fast implementation, real monitoring, early signal detection.
03 · Decide Win or kill. Keep only what compounds. Drop the rest.

Prioritise the biggest leaks, validate quickly, scale only what works.

Need this on your stack?

Book a 20-minute call Story