What Is Snowflake, and Is It Worth It?

Original article was published by Frederik Bussler on Artificial Intelligence on Medium


Is it worth it?

By the end of 2019, Snowflake had 3,400 customers and reported $264.7 million in revenue — an average of $78,000 per customer.

On Snowflake’s pricing page, examples of lighter-use and heavier-use customers are given, paying around $23K and $119K per year, respectively.

So, what do you get for all that cash, and is it worth it?

At a certain point in data volume, tools like Tableau alone won’t cut it, so large companies need to turn to cloud vendors, like Azure, BigQuery, and Redshift, which offer:

  • Tons of cheap storage
  • Ability to scale up or down
  • Outsourcing data warehouse management and security
  • Paying on-demand

Besides vendor-neutrality, Snowflake is a useful alternative in that it separates compute and storage requirements, so you can scale them up or down independently, giving greater flexibility.

Further, Stitch Data writes that “Snowflake is a better platform to start and grow with,” compared to Redshift, as it’s an overall easier software to approach.

However, Snowflake isn’t always the best option. If your workflow is already integrated heavily with Amazon, Google, or Microsoft, then it probably makes sense to use their native cloud offerings.

Snowflake also isn’t necessarily the best option for highly intense workflows. Stitch Data and Xplenty both agree that if you’re dealing with “massive workloads,” Redshift is the way to go.

What Snowflake Is NOT For

We’ve covered that Snowflake is a powerful cloud-based data warehouse, but it’s not an analytics, dashboarding, or AI tool.

That means you can’t make visualizations, predictions, or really any analysis in Snowflake. Other data warehousing tools, like BigQuery, have some native analytics functions, but you’re best off using a more feature-rich analytics tool like Apteo.

I’ve written about how to analyze BigQuery data in Apteo, and the same can be done with data from Redshift, Snowflake, or any other tabular data source.

Summary

If you want an easier approach to data warehousing, without vendor lock-in, Snowflake may be your best bet. If you have extremely huge workloads, and/or need analytics functionality, however, you may want to go with Amazon, Google, or Microsoft.