To get value out of it, a developer has to write sql and integrate some other reporting tool to actually view results.
How are they positioning the product to customers? As a tool for developers? For business analysts (will business analysts want to even write sql)?
How are they picking up million dollar contracts when the output of their product is essentially a stream of data that needs another tool to visualize it and the input needs to be written by an engineer?
To answer your question, what's great is that anybody who can SQL can use this, developers and business analysts alike. Developers who want data to back up product decisions will use this to answer, eg, how many customers use feature A vs feature B, and then use that information to choose to build either feature C or feature D. Business analysts want answers about the business, and if they have to write SQL to do it, then they'll partner with someone who can, or more likely, just learn SQL. (This is why "data scientist" is a hot new career path.)
Corporations that are in a place to be giving Snowflake a million dollar contract most likely already have a similar system already in place, either homegrown or via a competitor. (Eg Amazon Redshift.) Thus, their positioning depends on the customer, but for customers that previously already had a million dollar contract just with a different vendor, their focus is on why their product is super-amazing-awesome compared to their competition. (Which, and the market agrees, it is.)
patio11 has a thread on just how unseen some of the niches of software is - https://twitter.com/patio11/status/1235919318329643008
Software is like an iceberg - there's a wild world of software that none of us have heard of. It seems strange that it's possible for Snowball to succeed given that it's "essentially a stream of data that needs experts in order to use it", but the same thing is true about the Internet in the 1980s.
To date, I haven't had a specific need for Snowflake, but that's down to my client's situations, not any issue with the company or its product. If that situation changed, it would indeed result in contracts.
It's not winning on performance or price/performance. I suspect it's other features such as easy onboarding and an ETL (or ELT) toolchain. They claim it's self-managing, as so many products have before.
B2B over the web/cloud circumvents traditional enterprise sales (and billing!) channels, so I guess that's part of it.