To enable this early access capability, contact your technical account manager.
Bring your business practice into our hyper-personalization platform
You can onboard your unique business logic into Dynamic Yield as recommendation algorithms. These can be the result of a recommendation algorithm that your data scientists developed, or manually curated by your master stylists to "complete the look". All models must be contextual to an SKU. For example, related products, similar products, bought together, and so on. After the logic is onboarded, you can use this model exactly like any other algorithm. Meaning, you can apply filters, use any recommendation template, test, and more.
How it works
Add your product logic (list of SKUs) to your product feed as a column. Contact your technical account manager to get more information about the column name format. You can upload up to 10 algorithm columns (within the limit of 300 columns for the feed).
After uploading, your custom algorithms are available for use in the strategies you build, and you can apply all the Dynamic Yield tools to campaigns that use these strategies.
Note the following:
- These algorithm columns can't be used anywhere else in the platform (filters, targeting conditions, audiences, and so on).
- Column names can’t be changed or deleted after they are added.
SKU to multiple SKUs column structure
| Column Name | Description | Values Format | Example |
|---|---|---|---|
| For the column syntax, contact your technical account manager. |
Values: List of product SKUs (up to 100) First SKU is highest rank, and is shown first. If this column has empty values, the recommendation is displayed empty. There is no fallback. |
Strings (separated by pipes with no spaces; up to 1,000 characters; no special characters) |
Algorithm values: 243|9804|135|6320|413
|
Note: In CART context, only the first SKU is considered.