The data model is where you define the shape of your customer information in Prisma. Think of it as a template: before you load any data, you decide what fields each customer record will have — things like name, email, date of birth, or any business-specific attribute like account balance or membership tier.

Fields

Each piece of customer information lives in a field. When creating a field, you choose:

• ID — an identifier used internally (e.g. account_balance, last_login). Use lowercase and underscores, no spaces.

• Type — the kind of data the field holds. Common types are:

  Type	Use for
  Text	Names, codes, any free-form string
  Number	Integers (age, number of products)
  Decimal	Amounts, balances, rates
  Date	Dates in YYYY-MM-DD format
  Email	Email addresses

  For the complete list of available types, see the Field Types reference.

• Cardinality — whether the field holds a single value (One) or multiple values (Many). Use Many for things like product lists, where a customer can have more than one.

Avoid duplicate fields. Before creating a new field, check if one already exists for the same concept. For example, if email already exists, don’t create email_address for the same purpose. Duplicate fields cause confusion during data loading and segmentation, and split your customer data across two places.

Categories

Fields are organized into categories that help you find them when building segments:

• Demographic — general information about the person (age, city, gender)
• Customer Profile — business-specific data (products, account type, balance)

You assign a category when creating the field.

Field validation

Optionally, you can add a validation rule to a field to prevent incorrect data from being loaded. For example, you might restrict a birth_date field to only accept dates after January 1, 1900.

When a record fails validation during a data load, Prisma shows an error message indicating which field rejected the value: