How did Snowflake reach 158 percent net revenue retention, and what is actually copyable?
Consumption pricing did it. Customers paid for what they ran, so as they moved more workloads onto the platform their spend grew with no upsell conversation, no renegotiation and no salesperson in the room. At S-1, roughly half of revenue growth came from existing customers expanding usage, and it had been above 200 percent the year before. What is copyable is the principle: tie your price to a unit that grows when the customer succeeds. What is not copyable is the context, since a cloud data warehouse rides a once-in-a-generation migration. Do not benchmark your seat-based B2B tool against it.
Go deeper
5 resources, 1 India-specific, 5 link-checked.
📄 Article
✓ Link checkedFreeAdvanced
The primary-source teardown behind the famous 158 percent net dollar retention, including the detail most people miss: it was above 200 percent a year earlier, and half of revenue growth came from existing customers expanding usage.
The net retention numbers by pricing model (122 percent top quartile for usage-based versus 109 percent without) plus the warning that switching is an org change, not a pricing page change.
The good, better, best NRR bands (100, 110, 120 plus) that boards actually benchmark against, sitting alongside the efficiency metrics they get judged with.
Accel's argument that expansion revenue is the specific thing that carries a company past $50M ARR, with the Twilio and Shopify revenue-mix numbers to show what it looks like when it works.
Snowflake is one data point, and this session shows the pattern behind it: seven of the nine recent IPOs with the best net dollar retention priced on usage. That is the copyable part, separated from the once in a generation cloud migration.