Worldpanel by Numerator helps the world’s largest consumer brands understand how people shop through demographically representative panels across more than 30 countries. Its teams collect billions of transaction records, but the scale of that data makes valuable insights difficult to access. By building Item Picker on Lakebase and the Databricks Platform, Worldpanel by Numerator created a fast, governed experience that teams can use directly. The result is broader self-service access, faster global reporting and a foundation for AI-powered insights.
Rich shopper data was trapped behind legacy systems
Worldpanel by Numerator turns the everyday choices of real shoppers into the insights brands use to guide product, pricing, and promotion decisions. But answering any single client question can mean navigating an extraordinary amount of proprietary data. In one UK panel alone, teams work with roughly 18 billion transaction records and about 8,000 attributes. The insight was never in doubt, only whether the people who needed it could actually reach it.
Years of growth through acquisitions had left each market operating with its own tools, structures and processes. Bringing that panel data together was essential to building truly global products. Yet much of the work still relied on country-specific environments, peer-to-peer CSV transfers and a small group of specialists who knew how to navigate the data.
As demand for faster, self-service access grew, those limits became impossible to ignore. Some global reports could still land 10 to 20 days after the period ends, behind the pace clients increasingly expect. When the team set out to build Item Picker, a self-service tool for navigating and selecting from that data at scale, the question was whether the existing foundation could support it.
"With the legacy stack, it simply wasn't possible," said André Garrido, Head of Technical Engineering at Worldpanel by Numerator. "Every country had grown up with its own tools and ways of working, and at this magnitude of data, a single global approach wasn't something we could even start to think about."
Lakebase gives Worldpanel by Numerator a fast, governed foundation for Item Picker
The first step was getting the foundation right, and a global ETL pipeline and gold layer in Databricks gave the team a way to do it. Data from the company's markets and acquired businesses was brought together and standardized into a single, consistent format, making it easier to govern, integrate, and make available across the organization for the first time. For Worldpanel by Numerator, that common structure is what makes a true global product possible.
On top of that foundation, the team built Item Picker as its first major application on Lakebase. Item Picker lets users define the exact dataset they need from the company's full data lake, while Lakebase provides the speed, permissions, and direct connection the experience requires, all without copying data into separate systems or stitching together custom sync processes behind the scenes.
"We needed it to be fast, we needed it to perform at scale, and we needed it to be something we could build on again and again. That's exactly what Lakebase gave us," André said.
For the people answering client questions every day, the change is tangible. Teams can now move through billions of records and thousands of attributes, build the right selections, and work in parallel rather than wait on specialists. "Item Picker is the window into the data for a user," said Mike Nyhan, Product Director for Data as a Service at Worldpanel by Numerator.
Faster insight delivery and a foundation for AI-ready growth
Lakebase has accelerated Worldpanel by Numerator's ability to deliver insights while laying the groundwork for future AI-driven experiences.
Global reports delivered 11 days faster
The most immediate impact has been faster reporting. One global report that previously arrived 12 days after a reporting period ended is now delivered in one day. Worldpanel by Numerator aims to move more reporting toward near-zero-day delivery so clients can act while the data is still fresh.
Lakebase also makes Worldpanel by Numerator’s massive datasets easier to use, not just store. Through Item Picker, more teams can access and navigate the data without relying on a small group of technical specialists. They can answer questions faster and package data more flexibly based on each client’s needs.
“Clients can now get the data in the format they need, at the scale they need it, and do something with it, whether that’s with our help or entirely on their own,” said Mike.
A reusable foundation for future AI experiences
Worldpanel by Numerator is already looking beyond Item Picker. The company sees Lakebase as a reusable architecture for future tools that connect directly to its core data without custom sync pipelines or manual intervention.
The same foundation can support AI agents and richer self-service experiences. Worldpanel by Numerator’s goal is to eventually surface relevant insights before clients even ask for them.
“Where we’re heading is a place where the interesting insights are in front of clients before they’ve even asked the question. That’s the future Databricks is helping us build toward,” said Mike.