Anatomy of Smart Recommender Settings

Recommendation Settings enable you to configure several options for your Smart Recommender campaigns. Thanks to them, you can configure

• Recommendation Widget
• Recommendation Strategy
• Reporting
• Filters

In the Design step of your campaign, add a new variant > select your template > select an element to insert your template. Then, you see the Recommendation Settings button in the top right corner of the page.

Recommendation Widget

Using these settings, you can:

• Set a Number of Products to Recommend

You can choose the number of products to display in the recommendation widget for your users. If the API fails to recommend enough products to meet this number, fallback algorithms will generate additional recommendations to reach the specified quantity.

• Hide the recommendation widget 

You can hide the recommendation widget once there are not enough products to recommend. The number of recommended products can be low for different reasons such as having few products in the category, out-of-stock products, or way narrow filters. Although the Smart Recommender backs these problems up with fallback algorithms, sometimes to reach the desired number of products might not be possible. For such cases, Insider has the Hide the widget if the number of products is less than input area to define the minimum number of products as the displaying condition.

By default, the widget is set to display if at least one product is recommended. For instance, if you set the number of products to 10 and the number of products to show in the recommender to 7, the widget will appear when 7 or more products are available, including fallback recommendations. However, if 6 or fewer products are available, the widget will be automatically hidden.

• Hiding is valid both for Web Smart Recommender and API-based campaigns. 
• Hiding is a front-end operation that occurs after the recommendations are fetched from the recommendation API. 
• For the API-based campaigns, recommendations are served via API after applying this filter. 
• For the products that use the recommendation API directly (e.g. App Recommender or Email, Web Push Recommendations, etc.), you should manage this configuration. For example, if you want to apply such a rule on your mobile app, it is expected that you limit the output. 

• Shuffle products 

You can shuffle products to display them to your users in one recommendation widget in a shuffled way. All recommendations are served with a pre-defined sorting logic. According to the selected algorithm logic, the sorting can be ordered with the most purchased to the least or most viewed to the least. Using personalized affinity which amplifies the user’s tendencies regarding a product attribute, you can alter sorting. However, when the shuffle option is selected after the recommended products list is received, the products are randomly shuffled within this list and any rules mentioned above are not valid anymore. 

Recommendation Strategy

Under these settings, you can:

• Select your Recommendation Strategy 

You will see the algorithms available for the page type that you select to create your campaign. 

• Personalize your recommendations with attribute affinity

This feature enables you to offer more relevant content to your users. Attribute Affinity Model runs every day to process new users and new behaviors of existing users. The model calculates attribute affinity based on a user’s visit, add to cart, and purchase behavior. For example, if a user is more interested in a specific brand, the products from that brand are prioritized in the recommendations if applicable.

To make Attribute Affinity start automatically for Product Attributes, on the InOne panel, navigate to Components > Product Catalog Management> Product Attributes, under the Affinity Modeling, enable the toggle to start affinity work that specific product attribute.

• Apply dynamic product exclusions to exclude some products from recommendations based on the dynamic user data.

Reporting

You can enable Google Analytics Enhanced E-commerce Integration to send the product-based sales and impression logs to your Google Analytics Enhanced E-commerce account. Enhanced e-commerce is an advanced event integration provided by Google that enables product purchase and pageview performance tracking via Google Analytics. Smart Recommender can send those events if the campaign is viewed and/or clicked by the user, and Google can attribute the add to cart and purchased events to the clicked item in the same session. 

Google Analytics enhanced e-commerce is a good way of presenting the performance of Smart Recommender if you would like to show the contributed revenue based on the clicked items and the item-based performance.
To use Google Analytics Enhanced E-Commerce for your campaign, on Reporting setting, enable product list performance integration for Google Analytics Enhanced E-Commerce.

Filters

You can add standard or dynamic filters to filter out your product recommendations.