Data layer best practices

This article reviews our best practices for implementing and managing your data layer.

Code placement

In the page code, the data layer object (utag_data) must be declared before the reference to the Universal Tag (utag.js). This ensures that the Universal Tag has all the data layer variables needed to evaluate load rules, extensions, and tags. An example of the recommended code placement can be accessed from the Code Center.

Learn More: Code Center

Naming conventions

The goal of the data layer is to provide a set of variables that are vendor-neutral and easy to understand. The following best practices apply:

  • Use lowercase, underscore-separated variable names.
    Examples: site_section, product_unit_price, login_status.
  • Avoid vendor specific naming.
    Avoid eVar1, pv_a3, or oid.
  • Use meaningful variable descriptions.
    Sometimes even a properly named variable still needs additional clarification. A good variable description includes details about when and where a variable is set, which vendor is using the variable, and possible values to expect in the variable. Examples:
    • Login status anonymous or authenticated, used for eVar1 in Adobe Analytics.
    • Order subtotal does not include tax or shipping, exclude commas and currency symbols.
  • Avoid pluralized variable names.
    Variables that contain multiple values, such as the product array variables, should use the singular form of the name. Examples:
    • product_category instead of product_categories
    • product_id instead of product_ids
  • Prefix boolean variable names with is_. or has_.
    This allows you to quickly identify variables that contain a value of 1 or 0. Examples: is_registered, is_first_time_visitor, is_logged_in

Pros

  • Creates a consistent naming convention.
  • Easier for new users to understand what is available in your data layer and how each variable is used.
  • Tealium Support is familiar with this convention.

Cons

  • Additional effort to transition vendor-specific variables to vendor-neutral naming convention.

String variables

Use string values for all non-product variables. Boolean and numeric variables should be passed as strings.

For boolean values, use 1 and 0 instead of true and false. This approach is more stable and reduces confusion about the expected values in these variables.

Examples:

Boolean Integer
Correct is_registered : "1" order_total : "1234.56"
Incorrect is_registered : true order_total : 1234.56

Pros

  • Tag templates expect strings and arrays.
  • Minimizes testing effort during and after implementation.

Cons

  • None.

Array variables

Set product variables (price, quantity, ID, etc.) as arrays. The Universal Tag (utag.js) is designed to use arrays for all product variables. While it is possible to set product values in a comma-separated string, this approach is more prone to errors.

Array (recommended):

product_id : ["prodID1","prodID2","prodID3"]

String:

product_id : "prodID1,prodID2,prodID3"

Pros

  • Same format expected by vendor tag templates.
  • Improves readability of the data layer.

Cons

  • None.

Page types

All pages of your site should include a variable called page_type. This is used to determine the type of page the user is viewing. The suggested values include, but are not limited to:

Page Value
Home Page home
Category / Product List category
Product Detail product
Search Results search
Cart / Basket cart
Checkout Flow (User Info, Billing Address, Shipping Address) checkout
Order Confirmation / Thank You order

Pros

  • Gives a clear understanding of what kind of page the user is viewing.
  • Many vendor tags utilize a page type to function properly.

Cons

  • Initial development effort to add page_type variable to the Universal Data Object.

Third-party data layer objects

You might already have a data layer object implemented on your site, such as the W3C Data Object or your own custom object. We recommend that these objects be converted to the UDO utag_data format using one of the available data layer converters.

Pros

  • Better compatibility with built-in functionality, such as load rules, data mappings, and Web Companion.
  • Reduces support costs to investigate issues with your custom object.

Cons

  • Additional effort to implement.

Was this page helpful?

This page was last updated: January 7, 2023