Resolving 1M rows: $500. Querying them: 1¢.
Data Cloud consumption is not evenly priced, and the spread between the cheapest and most expensive operations is enormous. Querying a million rows costs approximately one cent. Running identity resolution across the same million rows costs approximately five hundred dollars.
Most teams do not discover this ratio from the documentation. They discover it from an invoice. By then the pattern that caused it has been running on a schedule for two months.
The billing model is the architecture constraint
The critical thing to understand about Data Cloud is that consumption is driven by operations, not by storage. That inverts the instinct most people bring from a data warehouse, where the expensive thing is keeping data and the cheap thing is asking questions of it.
Here it is the reverse. This kit leads with the full seventeen-operation rate card, with dollar values against the March 2026 model, so the cost of every architectural choice is visible before you commit to it.
Eleven ways the meter runs without anyone noticing
Every entry in the burn list shares a shape: something reasonable-looking is scheduled or configured once, and quietly bills forever. Editing an Identity Resolution ruleset, which re-bills the entire ruleset rather than the delta. Resolving identity across junk profiles that should have been filtered upstream. Streaming ingestion where a nightly batch would have been indistinguishable to the business, at a fraction of the cost.
Hourly Calculated Insights and over-frequent segment publishes belong to the same family: the configuration is defensible in isolation, and expensive because it repeats.
Filter before you resolve
The single highest-leverage change in the kit is ordering. Identity resolution should never be the first thing that touches your data — it should be the last, after junk, test records and unusable profiles are gone.
The kit includes the SQL to do that filtering before resolution runs, plus a burn worksheet you can populate with your own volumes to see what the current architecture actually costs per year.
What’s inside
- Full 17-operation credit rate card with dollar values (March 2026 model)
- The eleven operations silently draining your credits
- Why editing an Identity Resolution ruleset re-bills the whole set
- Resolving identity on junk profiles — and how to filter first
- Streaming versus batch: when streaming costs 40× for zero benefit
- Hourly Calculated Insights and over-frequent segment publishes
- SQL to cut identity-resolution cost before it runs
- A burn worksheet with real dollar maths for your own volumes
