Every shopper, every channel, one customer view.
How Walmart Labs Canada built a digital marketing platform on Google Cloud and Power BI that unifies customer data and turns it into targeted, measurable campaigns.
About the client
Walmart Labs Canada builds data and advertising technology for one of the country’s largest retail operations, where online and in-store activity must come together into a single, actionable customer view.
The Challenge
A marketing platform built like a product
Alphavima built the platform on Google Cloud, unifying customer data across online and in-store touchpoints into a single profile store. A rule-based audience generation engine lets marketers compose segments from behaviour, purchase history and channel, with no data team required per campaign. A custom React application is the marketer’s cockpit, with Power BI Embedded delivering real-time behavioural and campaign insight inside the same experience. Integrated tracking closes the loop: abandoned carts and campaign responses feed straight back into retargeting.
From blind campaigns to closed-loop marketing
Customer data unification
Online and in-store sources joined into one profile model on GCP.
Audience engine
Rule-based segmentation engine for campaign targeting.
Marketer cockpit
React application with Power BI Embedded analytics.
Tracking & retargeting
Abandoned cart capture and campaign retargeting live.
What changed
- Multiple systems with disjointed customer data
- No audience targeting for campaigns
- No behavioural visibility or cart tracking
- Unified customer data across online & in-store
- Rule-based audience generation engine
- Real-time Power BI insight and retargeting
Questions about projects like this
Why Power BI inside a custom app?
Power BI Embedded gives enterprise-grade analytics inside the product experience, so marketers never leave their workflow to understand performance.
Does this pattern require Google Cloud?
No. The same architecture maps to Azure, and we recommend the cloud that fits your existing estate and licensing.
How is privacy handled in a single customer view?
Profiles are built on consented, governed data with role-based access, and the model supports regional privacy requirements by design.
Can this work for a smaller retailer?
Yes. The pattern scales down: one profile store, one audience engine, embedded analytics. Scope shrinks, the architecture holds.