BigQuery to Snowflake Migration
Evidence-based platform evaluation that gave leadership a confident migration decision.
Stage.in is a regional language OTT platform with 5M+ monthly active users. The data platform had been built on Google BigQuery. As the platform scaled and the data team grew, questions arose about whether BigQuery remained the right long-term foundation given the company's infrastructure direction and cost trajectory.
Leadership needed a clear, evidence-based recommendation on whether to migrate to Snowflake on AWS — including a concrete business case and a credible execution plan that would not disrupt ongoing analytics work.
BigQuery was not a deliberate architectural choice — it was a legacy decision that Stage.in had outgrown. With the rest of the infrastructure running on AWS, BigQuery was generating significant network transfer costs every time data crossed environments. The evaluation narrowed to Snowflake and Databricks. Databricks was ruled out on operational grounds: the capabilities were compelling, but Stage.in did not have the data engineering headcount to administer it properly. A well-configured Snowflake instance, by contrast, could be managed by a lean team and would give the company a stable, scalable platform for at least two to three years — sufficient for the current trajectory without over-engineering. The recommendation was made with that operational reality explicitly front and centre.
- 01Audited current BigQuery usage: query patterns, cost structure including AWS network transfer overhead, team workflows, and known performance constraints.
- 02Defined the evaluation criteria with stakeholders: cost at scale, performance, operational overhead, ecosystem fit with the AWS stack, and the team's capacity to administer what was built.
- 03Conducted a structured evaluation of Snowflake and Databricks against those criteria, modelled against Stage.in's actual workload patterns rather than vendor benchmarks.
- 04Assessed operational fit in depth: Databricks' administration overhead vs the team's capacity, Snowflake's configuration ceiling vs the company's two-to-three year growth horizon.
- 05Produced a recommendation document with full rationale, ruled-out alternatives, total cost of ownership modelling, and a risk-adjusted migration plan.
- 06Defined the target schema and phased rollout plan; presented the architecture decision directly to leadership and documented it for engineering reference.
- Migration approved and scoped based on structured, evidence-based evaluation
- Target schema and phased rollout plan defined and documented
- Architecture decision — including ruled-out alternatives and total cost of ownership modelling — socialised across engineering and leadership
- Network transfer costs eliminated by aligning the data warehouse with the existing AWS infrastructure stack
Start a conversation.
Every engagement begins with a focused discussion of your current data environment and priorities. To schedule an initial consultation, reach out directly.
Get in touch