Related Services
Azure Databricks, Terraform, Azure DevOps, Astronomer, Python, dbt, Git, CI/CD, Unity Catalog, Databricks AI/BI Genie
Industry
eCommerce
Company
Xebia Data
TBAuctions (TBA) is Europe’s leading digital auction platform for B2B used industrial equipment. Its team of 1200+ employees in eight brands in the Benelux, the Nordics, the UK, and Continental Europe serves buyers in 170 countries. The Group extends the life of business goods, maximizing their value and sustainability; it facilitates the circular economy. With its proprietary intelligent auctioning platform, TBA makes buying and selling more effective and efficient by using technology, AI, automation, and economies of scale. TBA’s digital auctions make this happen by creating a cross-border platform that perfectly matches the demand and supply of used business goods.
Together with its brands, Klaravik, Troostwijk Auctions, Surplex, Auksjonen, PS Auctions, British Medical Auctions, Vavato, HT Auctions & Valuations, and Auktionshuset dab, TBA brings sustainable trade to a higher level. Sourcing locally and selling globally, 2.2M lots are auctioned annually as TBA’s websites receive 176M visits and 700K unique bidders per year. Its motto: “Everything Has Value.”
The Price of Success
Since 2021, TBAuctions has undergone significant expansion through numerous acquisitions, resulting in a portfolio that now includes nine auction brands. This remarkable growth has enabled the company to reach an increasingly broad customer base worldwide, hire new talent, and scale up its daily deal volume.
Integrating the newly acquired entities and their teams presented substantial operational challenges. As the organization grew, so did the complexity of its data platform, leading to a gradual erosion of the clear boundaries between development and production environments (Figure 1). Although leadership had made well-considered decisions about tooling, the pace and scale of growth began to exceed the capabilities of the existing infrastructure. In some cases, code reached production without undergoing automated validation, introducing risks to platform stability and, consequently, the execution of critical deals.
Even with the expertise and vigilance of the data professionals, the mounting complexity inevitably slowed down delivery timelines.
Figure 1: Dependencies between Azure Databricks notebooks in the old platform.

A New Beginning
Three data engineers from Xebia joined TBAuctions to completely rebuild the data platform from the ground up. However, it soon became clear that there were too many dependencies among the hundreds of Databricks notebooks and that creating a new data platform from scratch would be more advantageous.
Following best practices, Xebia's consultants built a new data platform with Terraform, following data architecture principles the team agreed upon and supported. The consultants clearly separated development from production, set up infrastructure as code, wrote plenty of documentation, and created requests for comments for every significant design choice in the platform. This helped TBA’s team adopt an engineering mindset.
There was also an issue with the orchestration of the data platform. In particular, job scheduling was implemented in three places: ingestion, data transformation, and dashboard creation. This created many synchronization issues between the three scheduling systems, so Xebia brought in Astronomer to orchestrate all processes and implement transparency throughout the company.
“Xebia didn’t just create our new data platform; they upskilled our team so we can own and evolve it with confidence.”
Bringing in the Teams
After the first four months of work, Xebia’s consultants started migrating existing code to the new platform. While the data platform is always a living thing, it was in a good enough state to start bringing TBA’s employees in, both old and new.
Onboarding the teams was probably the biggest challenge of the project. Learning new tools for the future while having high load on the team to perform now so Xebia’s consultants directed many pair programming sessions with dbt, Git, and other tools. TBA’s employees also joined some of Xebia’s trainings, such as dbt-learn and Python for Data Analysis. Most importantly, they had weekly knowledge-sharing sessions to ensure everyone was aligned with the engineering mindset and the way to do things in the new platform.
With most of the team onboarded onto the new platform, TBA is now in a safer place to experiment with and deploy new products. Additionally, thanks to the addition of Unity Catalog to the data platform, TBA professionals can leverage Databrikcs AI/BI Genie to uncover data insights. These products and services bring business value faster and more reliably to TBAuctions while making the company future-ready.