Digital was judged on last-click ecommerce revenue while most purchases happened in store — so the channel that drove footfall looked unprofitable.
← Back to the universe
RetailEcommerceMarketingWebsites
Omnichannel Retail System
Online-to-store measurement and a lifecycle engine connecting 34 physical stores to digital demand.
Multi-brand retail group · 2021 — 2023
4.6x
Blended ROAS
22%
Owned revenue
34
Stores connected
A geo holdout test across matched store clusters measured the true incremental effect of paid social on in-store revenue.
Measure incrementality, then fund it. Build a lifecycle layer that owns the customer between purchases.
POS-to-CRM identity resolution, offline conversion imports, geo-based incrementality testing.
Quarterly incrementality re-tests, catchment-level budget allocation, and creative localization per region.
When measurement is wrong, the correct decision looks like the wrong one. Fixing the lens changed the entire budget.
- 01Ran geo holdout experiments to establish incremental lift baselines.
- 02Connected POS data to the customer data layer for offline conversion imports.
- 03Rebuilt product feeds and local inventory ads per store catchment.
- 04Launched lifecycle flows across email, SMS and WhatsApp.
- Blended ROAS from 2.8x to 4.6x
- 22% of revenue from owned channels
- Local inventory ads driving measurable footfall in 34 stores
Geo holdout design
Local inventory ads
ShopifyKlaviyoMeta AdsGoogle Merchant CenterBigQuery
Performance lead and measurement architect.
Digital was reclassified from cost center to the group's primary footfall driver.
Continue exploring
AI
TwinAI Growth OS
An end-to-end growth operating system: server-side tracking, automated reporting and a suite of AI agents running acquisition operations.
HealthcarePatient Acquisition Engine
A HIPAA-conscious acquisition and booking system for a clinic network across four cities.
Real EstateReal Estate Demand Engine
A cross-border demand system for off-plan property launches with broker-grade lead routing.