VaultSpeed supports a data-driven approach at HDI Global

The HDI story

Kolja Sadowski is the head of the Industrial Lines BI team that – organized together with their colleagues from the business departments - provide the BI systems for the industrial insurer HDI Global. Birgitta Kuehn, who is part of that team, works directly with VaultSpeed on a daily basis. The team is responsible for data integration, data modeling, and BI reporting in mainly controlling and accountancy and aims to provide added value to other lines of business. The team is in charge of 35+ systems. Some legacy systems have already seen the best part of their lives, while others are relatively new. As part of the company’s digitization strategy HDI Global consolidates many of its former reporting solutions and aims for extended capabilities around data and analytics. This is accompanied with a massive cloud shift for most of the data-centric systems. And that is where VaultSpeed comes into play.

Challenge

The company wants to build a data marketplace where every department could access the data necessary e.g. for reporting and auditing purposes, risk management, and claims handling. HDI started out its data cloud initiative with a data lake which already went live in 2021. But sure enough this was just a first step, and to achieve all the other goals like providing refined, historized, and harmonized data, a data warehouse is needed. A couple of years ago, HDI started a big data warehouse project: they wanted to have daily harmonized data from local and global sources. "That was a greenfield approach, and we believed that Data Vault was state of the art, so we started modeling a data warehouse with it. At first, we had a partner who started modeling everything manually," says Kolja. But, after a while, it became evident to the team that manual modeling wasn't an optimal solution.

So, the next step for the company was acquiring a custom build automation tool. This approach served for a few years but didn't stand the final evaluation: "The generator was a matter of complexity, and only one person was very deep into the generator logic and could help with creating new releases of the tool; we wanted to avoid that dependency to an external resource. Also, there were some limitations in creating the interfaces to the consumers generated by the tool. When now starting to redesign the data warehouse and bring it to the cloud HDI decided to go for a cloud-native solution and leave the on-premise obsolete legacy behind. As Kolja says, they wanted a standard software on the market, a tool that his team could learn to use and would be able to get help online when needed.

Why VaultSpeed

Even if their first try with Data Vault bore much room for improvement, HDI was still convinced that the methodology of this approach would suit their needs. At that time, reeeliance, their local service partner, proposed VaultSpeed to HDI as an efficient cloud-native solution. The encompassing pilot, our experienced team, and an overall "fresh and modern feel" helped us win the software contest despite being an early technology. Kolja says, "I must admit, I hadn't heard about VaultSpeed before then, but now I can say that our experience with them is great. The team loves working with VaultSpeed." Kolja and Brigitta note that developing a Data Vault model became more straightforward with VaultSpeed. With our solution, they have no problem implementing individual use cases, and the ability to generate code for various target systems minimizes a vendor lock-in.

Solution

HDI's first experience with VaultSpeed was when they tried our pilot and followed the training. "We all took part in VaultSpeed’s automation training. For me, it was one of the best trainings I've had so far. Even the developers that were on the project for months and years highly rated this training. We started with the basics of Data Vault and moved on to the tools and options. That was very valuable." Brigitta and Kolja both found it helpful that we also provided educational videos. Diving into the world of Data Vault can be challenging, so having those tutorials available to watch on-demand was definitely an asset to the team.

Before VaultSpeed, HDI had a quite complicated development process. If they wanted to deliver something to UAT or in production, there were many manual steps, which took them days or weeks to build. Adopting the new approach wasn’t just changing the platform from on-premise to cloud, it was a different way of looking at the data pipelines they were building. "You can create new releases and go ahead and feel done within an hour. This is how VaultSpeed supports technical agility. And that's the most crucial part," comments Kolja. VaultSpeed has helped HDI move to the cloud and automate processes, providing high-quality data faster and easier. As a result, HDI has already been recommending VaultSpeed.

Reliable service partner brings additional value

VaultSpeed is backed by a strong network of local service partners that help our customers to deliver the best solutions. For example, reeeliance helped HDI get the DWH team up to speed with VaultSpeed and Data Vault. They have well-trained experts in VaultSpeed; they delivered Data Vault modeling assistance and assisted in modeling data marts on top of the Data Vault layer.

They dramatically changed how HDI's developers deliver code to the production systems. All generated code is transferred directly to a GIT repository by VaultSpeed's agent. From there, all builds are automatically deployed to UAT and production systems, leaving no margin for deployment errors. In addition, integration testing scripts are automatically executed. They completely moved HDI's team from a trunk approach to a state-of-the-art CI/CD pipeline.

"With automated CI/CD there is no more writing off changes, no more tickets, everything is documented automatically. And that's so much faster, quicker, better, and more satisfying than before. We've got early adopters in our team who are thrilled and wonder, how could they have ever worked differently?" says Kolja.

reeeliance onboarded new HDI team members into the project with every product increment up until the point where HDI's data engineers were fully self-sufficient. Today, they have a supporting and advisory role when new options like event-driven loads or streaming are considered.