← All case studies
▦Data · ADAData & AI 2022–Present

One data pipeline for 20+ global brands

Fourteen platforms, fourteen definitions of the same number. Here’s how I made them agree.

2,000+pipelines run daily
500+stores covered
14platforms, one definition
ShopeeGMVTikToknet salesLazadarevenue?CoupanggrossHKTVmallordersMomopaid amtFoodPandaGMV-ishShopifysalesAmazonnet revZaloraGSVRobinsonsturnoverGrowsaritotalRakutensettledBliblivalue?same number, 14 definitionsBRONZEraw exports,14 platformsSILVERone standard schema→ Analytics teamGOLDcombined report tablesdashboards& clients
14 platforms, each with its own idea of “sales”
!The problem

Same number, different meanings.

We wanted brands to see all their eCommerce data at a glance, across 14 platforms. But each platform shares different data, and even where they overlap, each defines its metrics its own way.

Without a shared definition you’re comparing numbers that don’t mean the same thing.

Sales
Traffic
Media
Content
Affiliate
Ops
Shopee
TikTok
Lazada
Coupang
HKTVmall
Momo
FoodPanda
Shopify
Amazon
Zalora
Robinsons
Growsari
Rakuten
Blibli
What each platform gives us
▦What I did

Agree on the meaning, then build.

I worked out what each number really meant, from the data and from how people use each platform. Then I normalized it into a bronze → silver → gold pipeline.

Bronze→raw reports, untouched
Silver→one standard database for Analytics
Gold→combined view for dashboards and clients
✓What happened

Dashboards on one foundation.

2,000+ pipelines now run every day for 500+ stores. A few dashboards I’m proudest of:

Reckitt & L’OréalReal-time campaign dashboardAct within minutes on live campaign days.
Nestlé, Mars & ColgateBuyer & platform behaviorConsumer behavior, platform by platform.
Mars WrigleyDigital shelf performanceBetter positioning through image and content.
Built with
SQLPythonETL pipelinesDashboards