I don't agree with the assement that compute can be commoditized and be cheaper on SF and DBX side. This is the fundamental difference between OLTP (eg. salesforce) and OLAP (Warehouse). One is a simple operation and the other takes tons of processing.
If you want to go back to cheap compute, we can go back to commodity hardware in Hadoop. There is a reason that spark and MPP processing works well in a cloud world. You can get workloads done at massive scale with pay by the second.
Great writeup here. The problem I have is comparing to dbt too much to Salesforce/SaaS business logic. Despite building tooling for dbt for over 3 years, I could never wrap my head around the business. You're either selling compute or your building software products but somehow dbt is in this uncanny valley of doing neither.
It's remarkable they've been able to get as far as they have especially when a business has to buy a bunch of tools and expensive people to make it all work well.
the final form of dbt is basically portable business logic tho
if you pay attention to companies like the big health insurers or companies w super fragmented data like banks, their “system of record” is basically thousand of layers of stored procedures sitting on a sql server database
great writeup ethan. ultimately it comes down to what's the commodity - for internal enterprise workloads, its compute. for vertical apps, its business logic.
if anything, having this "codified" will get both camps that they need for the right ICP.
I don't agree with the assement that compute can be commoditized and be cheaper on SF and DBX side. This is the fundamental difference between OLTP (eg. salesforce) and OLAP (Warehouse). One is a simple operation and the other takes tons of processing.
If you want to go back to cheap compute, we can go back to commodity hardware in Hadoop. There is a reason that spark and MPP processing works well in a cloud world. You can get workloads done at massive scale with pay by the second.
I wish the dbt and fivetran side luck
Love this!
Loving it. A very lucid analysis - from a former, long-time Snowflake employee.
I think the acquisition timeline comes from Simon Späti, unless he got it from somewhere else https://x.com/sspaeti/status/1977796652552376594
This was great, thanks.
Great writeup here. The problem I have is comparing to dbt too much to Salesforce/SaaS business logic. Despite building tooling for dbt for over 3 years, I could never wrap my head around the business. You're either selling compute or your building software products but somehow dbt is in this uncanny valley of doing neither.
It's remarkable they've been able to get as far as they have especially when a business has to buy a bunch of tools and expensive people to make it all work well.
yea i don’t disagree
the final form of dbt is basically portable business logic tho
if you pay attention to companies like the big health insurers or companies w super fragmented data like banks, their “system of record” is basically thousand of layers of stored procedures sitting on a sql server database
aka the predecessor of dbt
Awesome! thx for useful resource👍
Very cool! But are the charts on increasing compute share on 1$ based on real data?
great writeup ethan. ultimately it comes down to what's the commodity - for internal enterprise workloads, its compute. for vertical apps, its business logic.
if anything, having this "codified" will get both camps that they need for the right ICP.