Classification sets overview

Classification sets provide a single interface to manage classifications and rules. This workflow combines the creation of classifications in Report suite settings with the Classification importer. The result is a single intuitive interface to create and manage classification data.

Classification sets versus legacy classifications

The main difference between classification sets and legacy classifications is classification sets combines all functiinality in one interface where legacy classifications relies on three interfaces.

Legacy classifications

Legacy classification

In legacy classifications, classifications Schema (like for traffic, conversions, marketing channels, and more) each have their own dimenson (key Key ). You define these classifications as part of your Report suite settings.

You define rules BidRule separately in rule sets as part of the Classification rule builder interface. In that interface you associate a rule set with one or more report suites.

You use the Classification importer to download a template DocumentFragment , to import UploadToCloud classifications to, or export Download classifications from a report suite - key (dataset) combination.

Classification sets

Classification sets

Classification sets combines all the legacy classification interfaces into one. Each classification set defines:

  • One or more subscriptions, which is the combination of a report suite Data and the dimension Key (key), that you want to classify. If you want to classify products based on a product SKU, you can define all the report suites with an applicable product SKU dimension. And you don’t have to replicate classifications across report suites like in the legacy classifications interface.
  • A list of classifications Schema (schema) for the the key. For example, for products classifications you can specify category, color, size, gender, and more. Once you have defined your classifications, you can download a template DocumentFragment , upload UploadToCloud classification data, download Download classification data, and more.
  • One or more rules BidRule to support the classifications.

To access Classification sets from the Components menu in the Adobe Analytics interface, you must be a product admin or belong to a product profile containing the permission item Report Suite Tools > Classifications . Note that legacy classification management interfaces are available from the Admin menu.

Classification sets consist of three functional areas:

  • Classification Sets: Create, edit, and delete classification sets.
  • Jobs: View the status of classification sets jobs.
  • Consolidations: Combine multiple classification sets into a single classification set.

Workflow

The workflow for classification sets typically involves the following steps:

  1. Consider for which report suite and dimension combinations you want to create a classification set. An example is to define a product classification set that you create for any report suite for which you want to classify products with more details. For example, details like category and color.

  2. Create a classification set with subscriptions for one or more report suite and key dimension combinations that identify products. For example:

    table 0-row-2 1-row-2 2-row-2
    Report Suite Key Dimension
    Report Suite 1 Product ID
    Report Suite 2 Product SKU
  3. Add the classifications that you have identified to the classification set schema. For example:

    table 0-row-2 1-row-2 2-row-2
    Classification Name Identity Name
    Category category
    Color color
  4. Manually create a file containing classification data. Use a template to ensure you use the supported file format and columns for the file. Then add the data to the template file.

    Alternatively you can export data directly from your product catalog in the supported file formats with columns that adhere to the template. For example, a CSV file, like:

    code language-none
    Key,Category,Color
    Adobe Nike Tech Fleece Full-Zip Hoodie - Men's,Men,Black
    Adobe Nike Tech Fleece Full-Zip Hoodie - Women's,Women,Black
    Men's North Face Adobe Jacket,Men,Black
    Nike Air Hybrid 2 Golf Bag,Equipment,Blue
    STITCH® Ultimate Garment Bag,Equipment,Brown
    Adobe Analytics Training Tee - Navy,Men,Navy
    AirPods Pro 2,Electronics,White
    Adobe Analytics Training Tee - Green,Men,Green
    Women's North Face Adobe Jacket,Women,Blue
    Adobe Analytics Training Tee - Grey,Men,Gray
    Adobe Analytics One Million Views - Grey,Equipment,Grey
    Adobe and MGM Tee - White,Women,White
    Adobe and MGM Tee - Charcoal,Women,Charcoal
    

    In the classification data file, you refer to the key dimension for each report suite (for example: Product ID and Product SKU) using Key. And you refer to each classification using the Classification Name (for example Category or Color).

  5. Upload the file that contains the classification data into the classification set schema.

  6. Set up rules to automatically classify incoming data and data from the past.

  7. Automate the process of updates to your product catalog that you want to see reflected in classification data through the use of a cloud location.

  8. Download your classification data to validate the content.

  9. Inspect the job history to see the results of your actions (upload, download, template, and more) on classifications.

  10. If you have multiple similar classification sets as a result from a migration from the legacy classification functionality, consolidate these classification sets.

Improvements

The backend architecture released with classification sets contains several improvements:

The backend architecture released with classification sets also contains several changes:

  • When using the browser or automated import, Overwrite on conflict is always enabled.
  • When using the browser or automated import, the option to export immediately after import is no longer supported. Exports must be initiated separately.
  • The Analytics 2.0 GetDimensions API endpoint now returns string identifiers for classifications instead of numeric identifiers. Numeric identifiers can still be used, but the recommendation is to use the new string identifiers where possible. Numeric identifiers can be retrieved by using the ?expansion=hidden query string parameter.
IMPORTANT
The performance of classification sets mainly depend on the number of unique key values that contain data. Be careful when you do have variables that contain large numbers of unique values. Especially when you combine such variables from multiple report suites and dimensions into a single classification set.

Limitations

  • Classification sets do not yet support rules. Rules functionality is added to the classification sets interface before the legacy rule builder functionality becomes unavailable.
  • There is no migration of legacy classification rules and configurations to classification sets. A migration utility is added to the classification sets interface before the legacy classification functionality becomes unavailable.
recommendation-more-help
46b8682c-fda6-4669-9355-1a44923e549e