let's work together
514-589-4563
4388, rue Saint-Denis, Suite 300,
Montréal, QC, H2J 2L1
Case Study

Reduce customer churn through predictive modeling and AI

Impact

In a high-growth context where retention is key to profitability, we helped WeCook turn its data into a competitive advantage.

By building a unified data warehouse and developing an AI-powered churn prediction model, we enabled the team to identify at-risk customers and act proactively, with insights directly usable by marketing teams.

Results: 85% accuracy in detecting at-risk customers, a full implementation completed in just one month despite complex data sources, and a stronger ability to generate advanced insights to optimize marketing operations.

Proof that a strong data foundation, combined with AI, can turn complex data into actionable business decisions.

Client Objectives

Your AI Integrations Your profitability Your growth

Services and Technologies

Artificial Intelligence Our Artificial Intelligence Solutions BigQuery Data Collection and Engineering Data Strategy Data Warehouses and Data Lakes Data and Technology Data Platforms Customer Data Platforms (CDPs) AI Pilot Projects Predictive Intelligence Customer Engagement & Loyalty AI Strategy Customer Intelligence

Industry

Retail Consumer Goods

The mandate

WeCook aimed to reduce churn and maximize customer lifetime value by leveraging its existing data.

The challenge: transforming a rich but fragmented data ecosystem into a concrete performance driver by developing a predictive model capable of identifying at-risk customers and informing marketing actions.

All within a short timeframe, with a scalable and cost-efficient solution.

 

85%

for churn detection

The Strategy

Build a high-performing data foundation

The project began with the implementation of a robust Google Cloud architecture capable of centralizing and efficiently processing large volumes of data. 

Connect a complex data ecosystem

Data from multiple platforms (GA4, CRM, media, transactions, customer service, payments) was integrated and harmonized to create a unified customer view. 

 

Structure data for activation

A medallion architecture was deployed to organize data and enable its use for both analytics and artificial intelligence.

Develop an actionable predictive model

The churn model was designed to be directly usable by marketing teams, incorporating a wide range of behavioral and transactional signals. 

Accelerate deployment

Tools such as Fivetran, dbt, and Cloud Run helped optimize data ingestion, transformation, and model execution while meeting budget constraints. 

A well-structured data foundation enables the rapid activation of high-impact AI use cases.

The Steps

Define the target architecture

Assessment of the existing ecosystem and identification of an architecture aligned with data volume, activation needs, and budget constraints. 

Integrate data sources

Connection and ingestion of multiple sources: GA4, Klaviyo, media platforms, transactional data, customer service (Zendesk), and payments (Stripe), using native connectors, Fivetran, and custom APIs. 

 

Centralize in BigQuery

Migration of data into a centralized warehouse capable of supporting multiple terabytes and enabling advanced analytics. 

Structure with a medallion architecture

Organization of data into layers (raw, transformed, enriched) to facilitate usage and ensure quality. 

Develop the predictive model

Training of a churn model using Vertex AI, based on a combination of behavioral, transactional, and relational signals. 

Deploy to production

Scaling execution with Cloud Run and integrating model outputs into marketing processes.

Our Approach

 

Start with the business problem

Align the solution with a clear objective: reducing churn and increasing customer value. 

Simplify the architecture

Prioritize high-performing technology choices that are also pragmatic and cost-effective. 

Unify the data

Create a 360° customer view from multiple, heterogeneous data sources. 

Activate quickly

Deploy a model that can be rapidly used by marketing teams. 

Collaborate closely

Bring IT and marketing together to maximize impact. 

Go further with our experts

4 MarTech trends that are redefining the future of marketing
Artificial Intelligence

4 MarTech trends that are redefining the future of marketing

May 19, 2025 4 min read
Accelerate the adoption of artificial intelligence to support business growth
Artificial Intelligence

Accelerate the adoption of artificial intelligence to support business growth

October 30, 2020 4 min read
adviso_cta

You know where you want to go.
We know how to get you there.


Tell us what you need and we’ll help you find the best solution.
Contact us or 514-589-4563

Find and leverage what really matters.

Get our insights and recommendations to stay ahead in the digital landscape.