How Snowflake has helped customers in the Product Industry
Nike ~ Sportswear
Nike is a sportswear company that probably needs no introduction.
They had numerous silo-ed data marts, which were expensive and hard to maintain. They had high concurrency requirements, and had test suite with 99 concurrent queries that they used to validate upgrades of the on-prem appliances. Any vendor interested in challenging the on-prem appliance had to prove their worth on this test suite.
The impact of this materialized in a high cost of maintaining multiple data mart appliances, keeping data synchronized, etc. High concurrency caused query failures and missed SLAs.
With Snowflake, each virtual data mart runs in its own cluster. Snowflake passed the 99 concurrent test suite by choosing the right T-shirt size and then letting MCWH auto-scale out to 10 of those clusters. MCWH load balanced the queries across these 10 clusters and a few hundred dollars of Snowflake runtime beat a 7-8 figure appliance. Nike’s JSON and Parquet data loads directly into Snowflake VARIANT columns, and can optionally be distributed into existing schemas using bulk MERGE or Multi-Table Insert operators.