Case Study: Papier streamlines ETL and builds an attribution model with Fivetran

Papier Builds Customer Attribution Model, Replaces Time-Consuming ETL

Papier, a fast‑growing design and personalization e‑commerce expanding into the US, Australia and Europe, was hamstrung by fragile, in‑house ETL scripts that made centralising ad, clickstream and transactional data into Redshift slow and error‑prone — the CTO spent a full working day each week fixing ETL and data was only synced once per day. To address this, Papier adopted Fivetran’s automated ETL connectors (integrating Facebook, Bing, Google and other sources) alongside Amazon Redshift and Looker.

Fivetran replaced Papier’s brittle pipelines, doubling the number of usable data sources, increasing sync cadence (production DB every five minutes, Facebook hourly), and freeing roughly 20% of the CTO’s time; with data centralised in Redshift and surfaced in Looker, Papier built its own attribution model to measure ROAS/LTV and grew a data‑driven culture (about 60% of employees now use Looker dashboards). The result was faster, more accurate reporting and better product and acquisition decisions powered by Fivetran.

Papier

Joe Robertson

Chief Technology Officer