Outcome Prediction for Unavailable Item Replacement Decisions

Solution Overview

Problem

Online concierge systems face increased costs and potential customer dissatisfaction when pickers replace unavailable items, as it may lead to additional time and potential refunds, making it more costly than simply refunding the customer for the unavailable item.

Innovation Solution

An online concierge system employs an outcome prediction model trained on order and customer features to determine whether to allow a picker to replace an unavailable item, using predicted outcome metrics to influence picker actions and minimize negative outcomes.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of energy

If a picker obtains a replacement item for an unavailable item, then lost revenue from refunding the customer is mitigated, but additional time for order fulfillment is incurred and potential refund costs increase

Engineering Contradiction:

Improvelost revenueVSAvoidorder fulfillment time

Core Design Contradiction:

Loss of energyVSLoss of time

Solution Approach 1:

The system performs preliminary actions by proactively notifying the customer about item availability status before the picker completes fulfillment. When items are predicted to be unavailable, the system advances the refund decision-making process to occur before the picker arrives at the retailer, allowing customers to approve or disapprove replacement items in advance, thereby eliminating time spent searching for replacements during fulfillment.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system introduces an intermediary communication channel between the online concierge system and the customer. This intermediary mechanism (notification system with approval workflow) mediates the replacement decision process, allowing the system to coordinate with customers about replacement items without requiring the picker to make real-time decisions at the retailer location.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Quantity of substance

If a picker obtains a replacement item for an unavailable item, then revenue from the replacement item is gained, but the cost of obtaining the replacement item increases

Engineering Contradiction:

ImproverevenueVSAvoidoperational complexity

Core Design Contradiction:

Quantity of substanceVSDevice complexity

Solution Approach 1:

The system performs preliminary actions by proactively notifying the customer about item availability status before the picker completes fulfillment. When items are predicted to be unavailable, the system advances the refund decision-making process to occur before the picker arrives at the retailer, allowing customers to approve or disapprove replacement items in advance, thereby eliminating time spent searching for replacements during fulfillment.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system introduces an intermediary communication channel between the online concierge system and the customer. This intermediary mechanism (notification system with approval workflow) mediates the replacement decision process, allowing the system to coordinate with customers about replacement items without requiring the picker to make real-time decisions at the retailer location.

Inventive Principle:

Principle #24Intermediary (Mediator)

3Loss of energy

If the system allows pickers to replace unavailable items, then customer satisfaction may decrease due to unsatisfactory replacement items, but refunding every unavailable item results in lost revenue

Engineering Contradiction:

Improvelost revenueVSAvoidcustomer dissatisfaction

Core Design Contradiction:

Loss of energyVSObject-affected harmful factors

Solution Approach 1:

The system implements feedback loops where customer responses to replacement notifications are captured and used to refine future replacement decisions. The system learns from customer approval/disapproval patterns, feedback on suggested replacements, and appeal outcomes to improve the accuracy of availability predictions and the quality of replacement recommendations, thereby reducing customer dissatisfaction while maintaining revenue.

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The system dynamically adjusts parameters such as replacement thresholds, notification timing, and item selection criteria based on learned patterns from historical data. By changing these parameters adaptively, the system optimizes the balance between preventing revenue loss through replacements and maintaining customer satisfaction by avoiding unsatisfactory replacement choices.

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS12475503B2Replacing an unavailable item in an order using a trained outcome prediction model

Publication Date: 2025.11.18 MAPLEBEAR INC

AI summary

An online system receives orders from users and dispatches pickers to fulfill the orders by obtaining ordered items at a retailer. If an ordered item cannot be found by a picker, the picker may refund the item or attempt to find a replacement item. While obtaining a replacement item may increase revenue to the online system, it can also cause a bad outcome for user experience (e.g., an unacceptable replacement item, a refund request of the replacement item, etc.). To balance these interests, the online system trains a model to predict an outcome metric comprising a likelihood of a bad outcome from replacing an item or an expected amount of profit to the online system from a replacement item. The online system compares the outcome metric to a threshold to determine whether to promote or dissuade the picker from replacing a not-found item.