Churn Prediction and Classification for Online Concierge Systems

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional online concierge systems fail to accurately identify the events causing customer churn, which hinders the implementation of optimal remedial actions to mitigate churn.

Innovation Solution

The online concierge system evaluates customer interactions over time to identify churn by applying a churn prediction model to prior interactions and attributes describing order fulfillment, and then uses a churn classification model to attribute churn to specific events, allowing for targeted remedial actions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If conventional churn identification models are used, then churn detection is achieved, but the ability to identify specific cause events is lost

Engineering Contradiction:

Improvechurn detection accuracyVSAvoidcause event identification

Core Design Contradiction:

Measurement precisionVSLoss of information

Solution Approach 1:

The patent segments the churn analysis into two distinct models: a churn prediction model that detects whether churn will occur, and a churn classification model that identifies the specific cause events. This segmentation allows each model to focus on a specific aspect, improving both churn detection accuracy and cause event identification capability.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent introduces an intermediary classification layer between the prediction model and the remedial action system. The churn classification model acts as a mediator that translates predicted churn into specific cause event categories, enabling targeted remedial actions based on identified causes rather than generic responses.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Reliability

If generic remedial actions are implemented, then some churn mitigation is achieved, but optimal targeted remedial actions cannot be implemented

Engineering Contradiction:

Improvechurn mitigation effectivenessVSAvoidremedial action targeting

Core Design Contradiction:

ReliabilityVSAdaptability or versatility

Solution Approach 1:

The patent applies local quality by tailoring remedial actions to specific cause events identified by the classification model. Different cause events (e.g., slow fulfillment, item unavailability, pricing issues) receive customized remedial actions appropriate to their nature, rather than applying a uniform remedial strategy to all churn cases.

Inventive Principle:

Principle #3Local quality

Solution Approach 2:

The system performs preliminary classification of churn causes before implementing remedial actions. By identifying the specific cause event in advance through the classification model, the system can prepare and implement the most effective remedial action for that specific cause, improving both reliability and adaptability.

Inventive Principle:

Principle #10Preliminary action

3Productivity

If multiple remedial actions are available for different events, then optimal remedial actions can be selected, but the system complexity increases

Engineering Contradiction:

Improveremedial action optimizationVSAvoidsystem structure

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The patent implements universality through a multi-functional churn classification model that handles multiple cause event types within a single unified model. The model processes various input features (fulfillment time, item availability, pricing, communication) and classifies them into different cause event categories, providing a universal approach to handling diverse churn causes without requiring separate specialized models for each event type.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Data Source

PatentUS20250156930A1Predicting causes of churn and allocating fulfillment resources based on prediction

Publication Date: 2025.05.15 MAPLEBEAR INC

AI summary

An online concierge system identifies churn of a customer, which occurs when the customer does not perform a specific action within a threshold time period. The online concierge system determines an event causing churn of the customer based on characteristics of the customer and attributes describing prior fulfillment of an order for the customer. To mitigate different events causing churn, the online concierge system maps areas of expertise of pickers for different aspects of order fulfillment to corresponding events. Through a trained picker scoring model, the online concierge system determines picker scores for different pickers fulfilling an order for a customer using characteristics of pickers, including an expertise, characteristics of the customer, and an event causing churn of the customer. Based on the picker scores, the online concierge system selects a specific picker for fulfilling a subsequent order from the customer.