The business
Daikin is one of the world’s largest makers of heating and cooling equipment. Its North American business, which sells under the Daikin, Goodman and Amana names, is run from a campus outside Houston.
Most of that equipment never goes straight to a homeowner or a building manager. Independent dealers and contractors sell and install it. The better a dealer understands the local market, the more of it they sell, so sales data that dealers can read is worth real money to the company.
The problem
Marketing data lived in its own Snowflake instance, separate from the main one. Two warehouses mean two bills, two sets of access rules to keep straight, and questions that cannot be answered without copying data from one to the other.
The summer’s work was to bring the marketing instance into the main one, tidy how data was laid out once it got there, and turn commercial and residential sales data into something a dealer could use.
What I built
- The migration. Moved the marketing instance into the main Snowflake instance and rewrote the ETL that loads it, so the same data arrives by a cleaner route.
- The remapping. Reworked how data in the main instance is mapped, which made storage cheaper, easier to reach for the people who need it and better secured against those who should not have it.
- The dealer views. Built Tableau visualizations of commercial and residential sales so dealers could see how their market was moving.
How the data moves now
- Marketing instance
- Cleaner ETL
- Main Snowflake instance
- Tableau sales views
- Dealers and marketing
What changed
- 5%
less time to run queries after the migration and the new ETL.
- $15,000
saved every year by remapping data in the main instance.
- 2 to 1
Snowflake instances holding the company’s marketing data.
The sales views gave dealers market insight they could act on and fed a more targeted marketing strategy.