Snowflake Cloud Data Warehouse - The Good, the Bad and the Ugly
May 14, 2019

Snowflake Cloud Data Warehouse - The Good, the Bad and the Ugly

Wicks Justin | TrustRadius Reviewer
Score 9 out of 10
Vetted Review
Verified User

Overall Satisfaction with Snowflake

Snowflake is being used across our whole organization. It addresses the need to combine disparate data sources into one single data source and discover insights into our business that would not be easily possible with lots of disconnected data sources. It also delivers world-class query performance and the ability for different business divisions to query their own data quickly and create reports in a fraction of the time as opposed to a traditional data warehouse solution. We are also using it to drive machine learning, and the data scientists are gaining insights into our business data that would not be possible otherwise. It is addressing multiple business problems, from distribution of our products all the way through to billing.
  • Reporting queries run in a fraction of the time that they would in our production systems. For example, we can take the original MS SQL reporting queries , that used to take hours to run in our OLTP databases, and convert it into snow SQL, and run almost the exact same query in Snowflake in minutes, if not seconds.
  • Having all our different data sources in one data warehouse database means we can start looking for links between data sources and different business units to tie all our data together. We can see from when a customer was dialled, through to when/if they bought a product, to when they were actually billed, and identify where in the process we are most efficient, and where we can improve our services and product offerings.
  • Having a truly automated database system, without the need to create indices or maintain them, means we can spend more time in gaining insights into our data and getting actual results/data out, rather than spending time managing and maintaining the solution.
  • There is no easy way to schedule any type of task.
  • I have no idea on ROI.
We use SQL Azure as well. The internal trial of snowflake went so well, that we did not consider any other solution.
The web front end is OK, but they could improve the product with a front end GUI. We have connected DBEAVER to snowflake with an ODBC driver, but the queries are a lot slower from DBEAVER, compared to the web front end. SSIS also has an ODBC connection, but it is also a slower to pull data from snowflake. Overall the usability is good.
If you need a totally managed cloud data warehouse solution that is as fast as lightning, then snowflake is for you.
I would not use it for any OLTP production type of solution.