Data Engineering
From legacy data stack to modern data stack
How Watchfinder moved from a traditional data stack to a modern one on Google Cloud, reducing time to insights by 80% through BigQuery, dbt and Terraform.



Watchfinder & Co. is an online retailer of second-hand watches.
Website
Industry
Industrial Goods & Manufacturing
Location
United Kingdom
Stack
GKE · BigQuery · dbt · Terraform
Astrafy's role was to design and implement from scratch a new data stack for Watchfinder on Google Cloud and with leading data technologies. Astrafy also took care of knowledge transfer and training sessions once the solution was implemented.
The challenge
Slow time-to-insight and unreliable data quality
Watchfinder had a traditional data stack where there was a long time to insights from data generation to exploiting this data in various downstream applications. A significant amount of time was spent on infrastructure and fine-tuning applications instead of working on the data and getting valuable insights from it. Data quality was also often problematic and there was no automation whatsoever.
The solution
A greenfield modern data stack on Google Cloud
The solution is a greenfield Modern Data Stack project that involved different parts. We started with infrastructure and security foundations in order to build the data stack on solid foundations. Then we deployed the data applications on a GKE cluster and used BigQuery with dbt as the center for the transformations. Automation was a key part of the solution, with use of Terraform and a DataOps tool.
Results
Faster insights and fully automated infrastructure
Results were really transformative for Watchfinder and could be summarized as follows:
Reduced time to insights by 80%
Full visibility on data quality with test automation
Onboarding of new engineers in a matter of hours and convenience of dev environment to iterate quickly
100% automation of infrastructure via Terraform
IAM bindings with "least privileged access" and "segregation of duty"
Latest case studies


