Skip to main content

Creating and managing attributes

This article explains what attributes are, how they relate to taxonomy, channels, and attribute groups, and how to configure them for enrichment, validation, and publishing.

S
Written by SKULaunch Support
Updated over 4 months ago

What is an attribute?

An attribute defines a structured piece of product information that is used for enrichment, validation, and output. Attributes sit at the heart of your product data model and represent the clean, normalised data you want to manage and publish.

Examples include:

  • Brand

  • Ingredients

  • Protein percentage

  • Life stage

  • Product type

  • Energy (kcal/100g)

Attributes are typically populated using a combination of:

  • Input attributes

  • AI enrichment

  • Rules and validation

  • Manual review

Attributes vs input attributes

Attributes and input attributes serve different purposes.

  • Input attributes capture raw incoming data

  • Attributes represent the refined, structured version of that data

This separation allows SKULaunch to ingest inconsistent source data while maintaining a clean and controlled target schema.

Viewing attributes

To view your attributes:

  1. Navigate to Product settings → Attributes

  2. Use search, filters, or sorting to find specific attributes

  3. Each row shows:

    • Attribute ID

    • Name

    • Type

    • Creation date

This view gives you a complete overview of your product schema.

Creating a new attribute

To create an attribute:

  1. Click New attribute

  2. Choose Single attribute or Bulk upload

  3. Configure the attribute parameters

  4. Save the attribute

Once created, the attribute becomes available for enrichment and channel mapping.

Defining general parameters

When creating an attribute, you will define:

Taxonomy nodes

Select the taxonomy nodes where this attribute applies. This ensures attributes only appear where they are relevant.

Channels

Select one or more channels to define where the attribute is used, such as eCommerce or Source Data.

Attribute group

Assign the attribute to an attribute group to control how it is organised in the UI.

Type

The type defines how values are stored and validated. Common types include:

  • Text

  • Textarea

  • Simpleselect

  • Multiselect

  • Number

  • Measurement based types

Choosing the correct type is critical for data quality and downstream compatibility.

Description and AI guidelines

Each attribute supports:

  • Description
    Explains what the attribute represents and how it should be populated

  • AI guidelines
    Instructions for AI extraction and enrichment, including formatting rules, exclusions, or edge cases

Clear descriptions and guidelines significantly improve enrichment accuracy.

Content generation and AI enrichment

Attributes can be configured for:

  • Content generation
    Allows the attribute to be used when generating descriptions or other content

  • AI enrichment
    Allows AI agents to populate or improve the attribute value

These settings control how attributes participate in automation workflows.

Managing attributes at scale

Attributes support bulk operations:

  • Bulk upload for faster schema creation

  • Bulk actions for updates and management

  • Sorting to control display order

This makes it easier to manage large, complex schemas without manual repetition.

Editing and maintaining attributes

Existing attributes can be updated over time:

  • Descriptions and AI guidelines can evolve

  • Channel assignments can change

  • Taxonomy coverage can be refined

Attributes should be treated as living schema components that evolve with your business and integrations.

Best practices

  • Design attributes around business meaning, not supplier formats

  • Keep attribute names clear and unambiguous

  • Use taxonomy scoping to avoid irrelevant fields

  • Leverage AI guidelines to reduce manual correction

  • Review attributes regularly as channels and requirements change

A well designed attribute model is essential for scalable enrichment, reliable automation, and clean downstream publishing.

Did this answer your question?