Overview: the Define stage
The Define stage is where the Data Product Studio turns business knowledge into structure. You describe a business process you understand, and the platform generates the entities, data products, and process map that the rest of the workflow builds on.
Applies to: Domain expert
What the stage produces
A run of Define produces four things, all grounded in the domain and process you picked:
- A list of entities, each tagged Core or Supporting.
- A grouping of those entities into data products.
- A process map showing the flow of work and where each entity fits.
- A set of governance defaults attached to the process, ready to flow down to the data products you refine in the Designer.
No source data is connected at this point. The output is a blueprint, not a deployment.
The inputs the stage needs
Define is an inputs game. Better inputs produce better output. The platform asks for four:
- Industry, Domain, Process, and Functional Area. Where you are in the business and what you’re working on.
- Context. A few sentences describing what you need to answer with this data.
- Required components. A short list of artefacts the process must include.
- Source system. Optional, but it shapes the language the platform uses.
Each input has its own article in this section.
Why governance starts here
Most data platforms attach governance to tables, after the data has landed. The Data Product Studio attaches governance to the business process, before any data is connected. The pattern is intentional.
Governance defined at the process level cascades down: Process > Data Product > Snowflake Object. Each level can break inheritance to set its own values. An Employee data product inherits its security and privacy defaults from the HR process. An EmployeeBasic view inside that data product can break inheritance and carry its own settings, exposing only ID, email, and name for relationships and reporting while the full Employee table stays Confidential.
Setting governance once at the process means every data product the wizard generates from that process starts with sensible defaults. The work in the Designer is then about exceptions, not about tagging every object from scratch. This is the model the data management discipline calls business-led governance, and it’s why the platform asks for process and domain before it asks for tables.
How to do it
The fastest path through Define is the Discovery Wizard, which collects all four inputs and produces the output in three steps.
The articles in this section are reference behind the wizard. If you want to understand what each input does, why entities are tagged Core or Supporting, or how to extend the platform’s defaults with custom domains and processes, this is where to read.