Comparison
Cost attributionvsCost per request
Cost attribution
every query is tagged with a team, so the monthly bill can be broken down by who actually spent it.
Assigning warehouse spend to the teams, pipelines or dashboards responsible for it. Making the numbers visible per team is usually enough on its own to change behaviour, which is why showback tends to work about as well as chargeback and causes fewer arguments. It requires tagging at query time, so it has to be built into how workloads connect rather than reconstructed afterwards from logs.
Full entry →Cost per request
you finally attribute spend per feature and find that a summariser almost nobody uses is a third of the bill.
The unit-economics number: input tokens times input rate plus output tokens times output rate, aggregated per feature and per customer. Instrument it from day one; per-feature cost attribution is what tells you which product surface is quietly burning money.
Full entry →