×
From a historical tool landscape
to a modern data stack
Level 1 · The Company
Meet Pets Deli
Premium pet food, Berlin
Tailored meal plans for dogs & cats
Transparent ingredients
Convenient subscriptions
Direct-to-consumer brand
Level 2 · The Problem
A stack that could not scale
Every data source had to squeeze through the same old pipe.
Long processing timesReports waited on the pipeline
No data validationErrors surfaced downstream, if at all
Architecture not built to scaleEvery new source strained the stack
Level 3 · Feeding Time
Feeding time
Click a data source and watch the bowl fill.
The feeding stage
Live pipeline
Load
Transform
Test
Shop & Orders
Web Analytics
Marketing
Transform runs in dbt, every row is tested
Historical run
Figures illustrative.
0 / 3 fed
Bowl full - pipeline validated end to end
Feed all three sources to unlock the results
Level 4 · The Results
Insights, served 75% faster
The Outcome
How we got there
75%
faster time-to-insights
Snowflake
scalable warehouse, migrated from Postgres
dbt-tested
every run validated
Modern data warehouse
Postgres to Snowflake, room for GA4-scale workloads
Transformation layer in dbt
Modular, incremental, maintainable models
Tests on every run
dbt unit tests, trusted numbers downstream
Data sources
Postgres shop database
The modern stack
Snowflake
Metabase
Metabase
More case studiesAuxeumNatsanaAll case studies
© Gemma Analytics · Case Study Pets Deli