The options on the Global Recommendation Settings tab affect every recommendation strategy in the account. Follow these steps to access the settings.
- Click COMPONENTS in the top navigation bar, and then select Recommendations.
- Click the Global Recommendation Settings tab.
On this tab you can adjust how much historical data a recommendation strategy considers as well as create product filters that apply to all recommendation strategies in the account.
Recommendation Lookback Period
The default amount of historical data considered in a recommendation strategy is 30 days if that strategy uses a recommendation algorithm with lookback period options. This setting only determines what lookback period option appears preselected for a recommendation algorithm with lookback period options when you create a new recommendation strategy. This setting does not prevent you from selecting a different lookback period for a recommendation algorithm when configuring or modifying a recommendation strategy that uses a recommendation algorithm that has lookback period options.
You can change this default setting on the Global Recommendation Settings tab from 30 days to 2 days or 7 days. Click SAVE after you select a new setting.
Be aware that a 30-day default lookback period in some situations might cause new products to take longer to appear in recommendations. For example, if you frequently update a product catalog in the account, then recommendations drawn from it based on 30 days' worth of historical data might include products that are no longer in that catalog. If you update product catalogs frequently, consider a shorter lookback period.
Offline Purchase Data
If your account has the Offline Purchase Data add-on and if you have uploaded at least one Offline Purchases dataset, then the Global Recommendation Settings page includes an optional setting so that you can make one Offline Purchases dataset the default used for every recommendation strategy configured to include offline purchase data.
Even if you don't select an Offline Purchases dataset on the Global Recommendation Settings page, the options to include offline purchase data still appear on the recommendation strategy configuration page. If you configure a recommendation strategy that in some way includes offline purchases data but haven't selected an Offline Purchases dataset on the Global Recommendation Settings page, you don't see an error message, and the strategy will not work as expected.
General Filters
Product filters allow you to further refine the items included in recommendations. The general filters you create on the Global Recommendation Settings tab are applied to all recommendation strategies in the account, including existing strategies that are in use in experiences.
Furthermore, you cannot modify general filters once they are applied to a recommendation strategy. As demonstrated in this screenshot, the filters you create on the Global Recommendation Settings tab appear in a recommendation strategy, but the ability to edit their values is disabled.
Creating a General Filter
Follow these steps to create a general filter to apply to all existing and future recommendation strategies.
- Click COMPONENTS in the top navigation bar, and then select Recommendations.
- Click the Global Recommendation Settings tab, and then click ADD FILTER in the General Filters section.
- Choose an option from SELECT ATTRIBUTE, and then complete the filter equation.
- Optionally, repeat steps 2 and 3 to add additional filters, if necessary. Click the logic connector toggle from AND to OR as necessary.
- Click SAVE, and then click OK to acknowledge the warning modal reminding users that general filters apply to all recommendations.
See Filters in Recommendation Strategies for guidance on creating filters and understanding the logic behind them.