Snowflake

Snowflake provides cloud-based data platform to drive BI, data drive apps and share data.

Product

Data Ecosystem

Data Cloud

  • single source of truth
  • storage layer: ingest massive amounts to create unified data record
  • compute later: dedicated resources to enable users to access common data sets for many uses cases w/o latency
  • cloud service layer: optimize each use case performance requirements
  • supported on AWS, Azure and GCP

Data Marketplace: buy/sell datasets

Business Model

  • 93% of revenue is consumption based on compute, storage, data transfer
    • Customer may roll over of unused capacity permitted
    • Customer may exceed contracted capacity or extend beyond original contract term
    • Consumption often exceed initial capacity commitment amounts
  • Annual deployment fee of ~1% revenue
  • After adoption, focus sales on migration of customer workloads to snowflake platform
  • Billed annually in advance
  • Contracts of 1-4 years (weighted average: ~2 years)
  • Revenue: $1.065B expected (91% YoY)
    • Geography:
      • US: 82%
      • Other: 21%
    • 95% of revenue comes from existing existing customers under capacity arrangements
    • Gross product revenue: 68%
  • Cost of revenue
    • Cost of product revenue:
      • 3rd party cloud infra
      • Personnel costs with customer support, service availability, security
      • Snowflake receives credits from cloud providers

Core Metrics

  • Revenue: $1.065B expected (91% YoY)
    • Q3 Remaining Performance Obligations: $1.5B (+120% YoY)
    • Q3 Revenue: $254M (+103% YoY)
      • Product Revenue: $254M  (94%)
      • Professional services revenue: $17M
    • Q3 Cost of Revenue: 106M (+112% YoY)
      • Gross Profit: $166M (+100% YoY)
      • Gross Margin: 61% (62% last year)
        • Product gross margin (w/o services): 66%
    • F22: $1.065
      • Gross Profit: 63M
  • Customers: 4990 (+60% YoY)
    • 212 of Fortune 500
    • Customers >$1M: 116 (+107% YoY),
      • 51% of all revenue
  • Net revenue retention: 169%

Competitors

AWS Redshift

  • Can run on AWS S3
  • Pay for dedicated workers: $3.5k per month
  • $5/TB
    • can be reduced by compression and clever queries
  • Compute layer is tied to storage layer (slower queries)
  • AWS lock in: some enterprises may be on Azure
  • No native JSON support
  • Hard to share materialized views

Google BigQuery

  • No payment for storage
  • GCP lock in: most enterprises are either in AWS or Azure so they do cannot use BigQuery
  • Compute scales in lots of 100
  • $5/TB

Azure Synapse

  • Requires dedicated SQL pool to create long living SQL database
  • Azure lock in: some enterprise may only be on Azure
  • requires much more resource management

Competitive Moat

The paradigm of big data at scale already existed, but Snowflake has perfected the craft. Snowflake has created a well engineered system that provide its customers with a extremely scalable data platform for almost all data analysis use cases. I believe Snowflake has the best solution and the product they have built will be extremely difficult to reproduce.

Analysis

Being the single source of truth for all data within a customer makes Snowflake a mission critical system and makes it extremely hard for customers to switch to a competitor. With 4990 existing customers trusting Snowflake as the core data infrastructure, Snowflake should also be able to expand and upsell their product suite upstream (data ingestion) and downstream (data analysis, BI, AI).

Snowflake’s numbers look very different from other SaaS companies as their pricing model is based on usage and revenue is recognized after usage. I believe their high NDR comes from their sales team selling the product to customers but newer customers taking long periods to migrate to the new data platform. The credits from the previous year probably roll over into the next year where they ramp up usage and actually recognize the revenue. This is pretty clear from the fact that 95% of their revenue comes from existing customers and the remaining coming from new customers.

Snowflake plans on increasing it’s product margin from 68% to 73% which is already quite impressive margins for an infrastructure product. As the core data service, Snowflake has the opportunity to tax every TB of data that goes through it’s customers systems just like how AWS can tax every hour of compute and every TB of data.

At current valuation of 90x revenue, Snowflake is quite expensive but their growth rate and potential is still incredible. If we compare Snowflakes trajectory to be similar to AWS, in 3 years, we could see Snowflake hit $6B in revenue ($180-$300B company,  2-3x upside)

tech investing