Managing recommendations in Incentivio allows you to fine-tune how items are suggested and ensure they align with your menu strategy. You can review recommended item pairings, adjust item categories, and control which products appear as upsell opportunities. This guide walks through how to review, adjust, and optimize your recommendation settings for the best performance.
For information on Recommendation Performance see our Recommendations Report
Our Recommendations Management is located under Menu Management > Recommendations and consists of four tabs:
Item Analysis
The first tab helps you view menu item's current recommendations, so you can identify gaps, validate pairings, and ensure your most important products are supported by relevant upsell suggestions.
To view item-specific recommendations:
- Select any item or search for a specific item.
- If recommendations have been generated for this item, it will display the top items that are shown as a complementary recommendation. However, if this item does not have a sufficient purchase history (such as for new or infrequently purchased items), there may be no recommendations available to show yet.
Recommendations are based on how often something is purchased with the Main item. The bar indicates how often the items below get ordered alongside this item.
Custom Rules
Custom rules let you override automated recommendations to spotlight strategic items, drive campaigns, or ensure key menu items appear in upsell moments.
To add custom recommendation overrides.
- Select "Add New Rule"
- Decide which items, groups, or menus trigger the recommendation and if they need to add any one, or all of the products
- Add the products that will be recommended if a user adds one of the trigger items.
- Click Save
You can view and delete existing rules on this tab as well.
Item Classifier
Our system automatically classifies your items into categories like "Sides", "Desserts", "Beverages", etc. This is done because our system recommends supplemental products to the products that are in the customer's cart already. For example, if the customer has added 4 coffees, we would sooner recommend a muffin than another beverage. With all of this said, sometimes our artificial intelligence gets it wrong! This tab allows you to classify your items into categories manually.
To edit an item's Category:
- Select the pencil icon next to the category
- Select your new category
The new Category will automatically save
Settings
These settings allow you to control where recommendations are shown and how many to display. You can select the specific locations where recommendations will appear and set limits for different recommendation types like mains, sides, beverages, and more. Use these settings to fine-tune the recommendation experience for your customers.
Settings
What is the Maximum number of Recommendations to display?
The maximum number of recommendations a customer can see at checkout
Why use this: Adjust this to control how prominent recommendations feel — showing fewer can keep checkout simple and focused, while increasing the number can encourage more discovery and boost add-on revenue.
What is the Maximum number of Recommendation Overrides to display?
The maximum number of custom rule recommendations shown in checkout
Why use this: Useful when running promotions or highlighting priority items, ensuring your strategic overrides appear without overwhelming the guest experience.
Category Maximums
The maximum number of recommendations shown in checkout for each category
Why use this: Helps maintain balance across your menu by preventing one category (like drinks or desserts) from dominating recommendations, keeping suggestions relevant and varied.
Please note, by default our new model is trained to vary in its recommendations, you should not see one category overwhelm another. Use overrides for specific use cases and reach out to our team if you have any questions about your recommendations behavior.
When to Expect Upsell Recommendation Data to Populate
The Upsell Recommendation feature has two key components that operate on separate systems and schedules:
1. Upsell Recommendation Engine (Checkout Screen)
Managed by: Machine Learning (ML) Team
Purpose: Powers the real-time upsell suggestions shown to guests at checkout.
Activation Requirement: At least 1,000 historical transactions within the current Incentivio system are needed to train the model. (Note: there is no 90-day waiting period.)
Timing: Once activated, the engine begins generating and displaying recommendations on the checkout screen immediately.
Training Frequency: The ML training job currently runs Monday, Wednesday, Friday automatically.
2. Upsell Recommendation Summary Reports (Admin Portal)
Managed by: Data Team
Purpose: Displays aggregated metrics comparing purchased vs. non-purchased upsell recommendations. These metrics appear as tiles in the Upsell Summary Reports section of the dashboard.
-
Data Availability:
The tiles display data from the previous calendar month only.
When a brand’s Upsell Recommendation Engine has been live for less than one month, it’s normal for these tiles to appear empty or incomplete: the system simply doesn’t have enough historical data yet.
Update Schedule: TBD
Example
If the Upsell Recommendation Engine went live on October 1st, the Upsell Summary Reports will begin showing populated data starting in November, once the system has a full month of tracked recommendation data.