Independer, the largest independent comparison platform in the Netherlands, helps 21 million annual visitors compare insurance, energy, mortgages and telecom. Its challenge was that customer data was fragmented across legacy on-premises systems and organizational silos, blocking the single customer view the business needed to deliver proactive, personalized experiences. With Lakebase, Independer eliminated its last-mile export bottleneck, cut development cycles in half and unified its data, application and AI teams on one governed foundation.
Scattered customer data blocked a single view of 21 million visitors
For 25 years, Independer has helped people across the Netherlands make smarter decisions on insurance, energy, mortgages and telecom. With roughly 21 million unique visitors a year, every digital experience depends on understanding the full context of a person's journey: what they compared, how they engaged with service teams, what messages they responded to and what permissions they granted. Building that single customer view, what the team calls Customer 360, meant bringing all of that context together in one place.
But that data was scattered across website events, customer service records, account profiles and ERP systems. "You had to go shopping across the entire organization just to assemble a complete picture of a single customer," said Gijs Gennissen, Tech Lead and Solution Architect at Independer.
Independer's on-premises architecture had served the business well for two decades, but as the company migrated to Azure it needed a more scalable, event-driven approach. Yet even after the migration, releases still stalled at the last mile. After engineers built gold-layer tables, they had to write export scripts, load a separate SQL Server instance and only then begin work on a new API.
"Every new feature hit the same wall. You couldn't touch the API until the export scripts were written, tested and loaded into a separate database," said Bas Verburg, Data Engineer on the Customer 360 project.
At the same time, Independer's users increasingly wanted proactive alerts about savings opportunities on their contracts and more personalized, conversational interactions rather than static comparisons. Delivering those experiences required faster access to richer, consent-based customer data, and the export bottleneck was standing directly in the way.
Lakebase brings Customer 360 from data lake to production API
Independer centralized its data inside Atlas, its internal data platform built on Databricks, pulling together data from across the business and processing it through layered jobs into a unified set of gold tables. Every customer profile is built on the permissions and consent that person has granted, tracked and enforced through Unity Catalog.
With the data centralized and enriched, the remaining bottleneck was getting it into the hands of the teams that needed it. Lakebase eliminated that last-mile step entirely by synchronizing gold tables directly with the services that consume them. "Once the data reaches the gold layer, we are building. No scripts, no staging, no waiting for a handoff that used to take longer than the feature itself," Bas said.
Because Lakebase runs on PostgreSQL, it also opens doors that the previous SQL Server layer could not. Teams now have access to capabilities like Vector Search and geodata support, expanding what the serving layer can do well beyond basic queries. And with the full stack unified in Databricks and Python, the gap between data engineers and application developers has closed significantly. "Now everyone is building inside the same system. Data science, GenAI, platform engineering, they all see the same data and want to use it," Gijs said.
Faster delivery, stronger governance and a foundation for AI
With the export layer gone, Independer cut development work for new Customer 360 features by 50%. The team now iterates faster on the experiences powering the website, customer service tools and internal analytics. "Lakebase turned the longest part of every release into the simplest. Enriched data moves from Databricks into production APIs in half the time it used to take," Gijs said.
The impact is already extending beyond Customer 360. The GenAI team is actively building proof-of-concept agents that store state in Lakebase, with plans to move to production in the coming months. Data science is incorporating Lakebase into its MLOps architecture, replacing API-mediated access with direct queries against governed, enriched data. And because Lakebase can synchronize data in real time, Independer now has a path to real-time customer data flows without building a custom change data capture solution, a step that would have been a significant challenge under the old architecture.
Centralizing everything on one platform has also strengthened governance across Independer. All customer data flows through permissions and consent checks within Databricks, and eliminating replication across multiple systems reduces data exposure. That simplifies compliance and enables teams to move faster with confidence. "Databricks has become the unifier for our data departments. One platform, one way of working, one place where every team can build on the same foundation," Gijs said.