Boosting rules offer you the ability to influence the scoring of your items in recommendations per a placement. Boosting rules are a part of the Constraints area of the Placement because the rule you upload will alter the results of recommendations.

When do you need boosting rules:

• When you configure your placement and want to make sure items with a certain characteristic will be favored. You can use any tag that you have to create boosting rules.

The difference between Boosting and Tag Rules:

• Tag rules work as filtering leaving you with altered results ensuring that the placement will display or will not display items with specific. Boosting works to help you give an advantage to certain items.

Uploading Boosting Rules 

Before you can use boosting rules, you need to upload them via a .csv file in the Data section of the platform. Follow these steps:

• Navigate to the Data section in the menu, click Datasets, and click ‘Add new’.

• Select "Gaming Boosts" in the Templates dropdown. Give it a name and Description. The name is how you will see this option in a placement whenever you choose to add the boosting rule.

• After saving navigate to Editor and add boosts within the interface. The names of the tag_type and tags should be exactly like the ones you are sending us.
  E.g. if your game has the tag ‘Ancient Worlds’ then you should type exactly ‘Ancient Worlds’ and not ‘ancient world’.

• Or you can upload .csv file according to the schema tag_type, tag, multiplier in the Upload tab.

• Once uploaded or edited and saved, the boosting rules will be available for use in your placement.

Understanding Templates and Multipliers 

The .csv file should adhere to a specific format for successful upload:

• tag_type: Indicate the type of the tag used for classification.

• tag: This represents the tag associated with the boosting rule.

• multiplier: A numeric value that multiplies the model score. Higher multipliers lead to more significant influence.

You can choose different tag types within one template.

You don’t have to list all tags that belong to the tag_type in a template.

The multiplier in a boosting rule is essential for determining the impact of that specific rule on the model's output score.

• The multiplier value ranges from 0 to ♾️ (infinity).

• A higher multiplier leads to a stronger impact on the recommendation output.

Utilizing Boosting in Constraints

The "Boosting" field within the Constraints section allows you to integrate boosting rules into your placement configuration.

• In the Constraints section, locate the "Boosting" field.

• Select the relevant boosting rules from the uploaded dataset.

  If you have just one configuration for your placement - you can select one boosting rule. However, if you configured different target groups and/or variants for your placements - each of them can have different boosting rules applied if you wish.

• Apply the selected boosting rules to a specific target group.

By applying the chosen boosting rules to the appropriate target group, you'll influence the model's score calculation and recommendation output for that group.

If you have any questions or need further assistance, please don't hesitate to reach out to your CSM.