The lifting-and-dropping process from transactional stores into analytics stores is usually the first step in analytics data pipelines. This appears to fulfill that initial ingestion step specifically for Google BigQuery. A lot of times these days this ingestion step is ran through some third party service: Fivetran, Stitch, Matillion, etc. Similar to AWS's offerings for Redshift, this looks like Google's "native" ingestion tool for BigQuery.

I'd be curious to know how it handles the wrinkles Fivetran effectively solves (schema drift, logging, general performance) and if it does so at a much cheaper price.

If you use some sort of ETL to copy data from your OLTP to data warehouse you might use this to stream data from your OLTP to Big Query so you can use Big Query as a data warehouse.

Few come to mind!

- Startup that's running analytics off their production Postgres DB can use this as a gateway drug to try out BigQuery. Bonus for not having to ETL the data themselves or pay FiveTran / host Airbyte to do it

- Replace existing internally maintained sync job / ETL pipelines with this, if you're somewhat all-in on BigQuery / GCP for analytics (BQ is quite popular in analytics because of the pricing structure)