Data-driven decisions: The evolution of the BI team at Nitrado

Content

How did you build the team at Nitrado?

Peter Waleczek: When we joined Nitrado, there were already some metrics, but they were quite basic. After our acquisition, we decided to invest heavily in the Business Intelligence (BI) area. You built the BI team from scratch. What did you find, and how did you proceed? What were your initial goals?

Dominik Braun: Initially, we encountered a grown structure, where many things were poorly documented. The first task was to get an overview: What data is available, where does it come from, and why does it look the way it does? Then we built a team – initially, we were four people: two Data Engineers, a Data Analyst, and me. Our focus was on building a BI infrastructure, as there was hardly anything in place before. Much was geared towards operational work, with little thought given to long-term analysis or historical data. So, we laid the foundation to make use of well-modeled data.

How has the BI team and your work evolved over time?

Peter Waleczek: How has the BI team and your work evolved over time, particularly in the last year?

Dominik Braun:
Working with data is an ongoing process, and data modeling never really stops. There are always new data points to integrate into the model. The amount of data in a company is vast, and the deeper you dig, the more potential you discover.

In the first six months, our main focus was building the team and establishing a basic BI infrastructure. We had to ensure that the necessary systems were in place to collect, process, and analyze data.

From month seven to about month 24, we focused on deeply understanding the available data. This involved data cleaning, quality checks, and close collaboration with stakeholders to understand how the data was generated and what it meant. At the same time, we conducted ad-hoc analyses to provide value from the beginning and learn how best to link and interpret the data to meet the company’s needs.

The result is a solid data foundation, allowing for detailed financial reporting and granular analyses. After two to two and a half years, we reached a milestone by creating a time series that enables deep analysis going back to the founding of Nitrado and Apex.

To illustrate the scope of these 1.5 years: Our largest table, which contains the central data model, holds all currently defined relevant information in 140 columns. This table is based on 150 source tables that we load and process. The table is over 500 GB, or 0.5 TB, and currently contains about 120 billion data points.

Now, the focus is on effectively using these data and models within the company, accompanied by change management that alters how people work with the data. We’ve identified a group of 35 Data Leads with whom we work closely to support their decision-making through our analyses.

Can you provide a concrete example of where your work made a difference?

Peter Waleczek: Can you provide a concrete example of where your work made a difference?

Dominik Braun: A good example is the KPIs “Churn Rate” and “Customer Lifetime Value.” These metrics didn’t exist before. Now, we can analyze in detail how customers behave across different games, whether changes are market-driven or influenced by our product decisions. These in-depth analyses weren’t possible before.

How do you collaborate with stakeholders? Does the company actively use BI now?

Peter Waleczek: How do you collaborate with stakeholders? Does the company actively use BI now?

Dominik Braun: We work closely with Finance, Product, and Marketing. These areas regularly request information and analyses. Other stakeholders are still getting used to the new data and possibilities. We introduce them gradually by regularly presenting new dashboards and explaining how they can be used. A central part of our current work is breaking old habits and thought patterns so that stakeholders understand the advantage of accessing a shared, clean, and highly detailed data foundation with a historical perspective, rather than continuing to rely on operational snapshot data.

You’ve expanded your team from four to six people. Where did you invest, and why?

Peter Waleczek: You’ve expanded your team from four to six people. Where did you invest, and why?

Dominik Braun: The shift from foundational work to more visibility and use of the data was crucial. In addition to another Data Analyst, we hired someone solely responsible for creating and further developing dashboards. This role helps us better integrate the results of our work into the company and meet the specific needs of different departments.

Is it really necessary to build a large team for BI, or can you start smaller?

Peter Waleczek: For a mid-sized company, this might sound like a significant investment. Is it necessary to build such a large team, or can you start smaller?

Dominik Braun: You can definitely start building a BI function with a small team, even with just one or two people. However, the question is how quickly you want to see results. Companies have a lot of complex data, and processing it takes time. The more resources you have, the faster it goes. If you invest too little, projects can drag on for years without being fully completed.

That’s why it’s important to define clear goals from the start. For us, the financial process was the starting point. We wanted to establish clean financial reporting that also allows for in-depth analysis. This focus helped us guide the development of our BI function. It’s crucial to have a clear goal in mind and build on that step by step to make meaningful progress and advance the company.

How can a company that doesn’t yet have a BI function get started?

Peter Waleczek: How can a company that doesn’t yet have a BI function get started?

Dominik Braun: If there’s someone in the company who’s data-savvy and has worked with data in previous jobs, that can be a good starting point. Areas like performance marketing or finance often inherently involve working with data. Involving such employees in Business Intelligence can be the first step.

It’s important to understand that BI is a company-wide issue and can only realize its full potential when data from all areas is connected. Data and information silos in individual departments make it difficult to achieve full value. Therefore, general access to data, information, and company-wide collaboration is essential, even if many companies initially struggle with it.

The first and most important step is to create awareness of the importance of data and BI. Afterward, the decision can be made whether to tackle the issue internally or seek external help. Internal employees are already familiar with the infrastructure and processes, which can speed up the establishment of a BI department if the necessary skills are present. It’s crucial to have someone leading the topic strategically and operationally who knows what they’re doing. Otherwise, it can quickly lead to chaos that takes more time to fix.

At Nitrado, we drive the BI topic as a team, but we constantly bring in experts from various departments to move projects forward and discuss specific issues. For certain projects, we expand the circle, and my goal is for the entire company to eventually see itself as part of the BI topic. Everyone can contribute – whether through careful data capture or by using the resulting insights.

You mentioned infrastructure. What’s needed as a basic setup, and what are the investment costs?

Peter Waleczek: You mentioned infrastructure. What’s needed as a basic setup, and what are the investment costs?

Dominik Braun: Thanks to cloud solutions, you can now quickly and cost-effectively build a BI infrastructure. Initial costs are mainly for the team, not the infrastructure. I recommend starting with a cloud solution because it’s flexible and scalable. We use Azure as our cloud solution and Databricks with its modern Data Lakehouse approach as the core technology for all our BI processes. This combination offers many advantages, including better performance and cost control.

What tools do you use for data visualization?

Peter Waleczek: What tools do you use for data visualization?

Dominik Braun: For dashboards, we use Tableau, which is directly connected to our Azure infrastructure. We mainly conduct ad-hoc analyses in Databricks but sometimes deliver and visualize the results in Google Sheets, as it’s often faster than creating a Tableau dashboard directly. These ad-hoc analyses often evolve into new dashboards and models.

Many companies are already linking BI with AI. How do you see the future of this combination?

Peter Waleczek: Many companies are already linking BI with AI. How do you see the future of this combination?

Dominik Braun: I don’t think we’ll ever reach a point where data analysts become obsolete or where no one is needed to make informed decisions based on data. The idea that AI could fully take over these tasks seems unrealistic to me. AI can support and simplify processes, but it won’t replace human expertise.

A good example is anomaly detection. In a company, there are hundreds of data dimensions that can be viewed in countless combinations, leading to thousands of potential perspectives. No one can oversee all these perspectives simultaneously. Here, AI can help by identifying patterns and deviations that might otherwise go unnoticed.

For example, AI might point out that a revenue drop correlates with a marketing budget cut. While professionals in their respective fields might notice such anomalies, it’s often difficult to grasp the bigger picture. AI can therefore not only highlight problems but also point out opportunities that might be overlooked without its support.

Peter Waleczek: Are you doing this today, or do you plan to implement it in the future?

Dominik Braun: We are currently building dashboards that can detect anomalies. The next step will be to integrate these anomalies into an alerting system that tells us what to focus on. This will likely be a topic for us by the end of the year.

Is BI the key to optimal corporate decision-making, or are there aspects that BI can’t cover?

Peter Walezcek: Is BI the key to optimal corporate decision-making, or are there aspects that BI cannot cover?

Dominik Braun: BI is a powerful tool and can definitely help make decisions more informed and precise. It reduces the number of decisions that, in hindsight, turn out to be suboptimal, and enables companies to make better decisions based on data. However, BI is not a cure-all, nor is it a guarantee for always making the right decisions.

There are aspects that BI cannot fully cover, particularly when it comes to strategic decisions where “entrepreneurial intuition” and weighing risks and opportunities play a major role. While BI can be very effective in areas like performance marketing, where you can work heavily data-driven, other areas still require human judgment.

That’s why I advocate for companies to work “data-informed” rather than relying solely on data. BI should be seen as a tool that helps leaders make informed decisions, but it doesn’t replace the need to complement complex decisions with experience and intuition. Our BI mission at Nitrado reflects this balance: “Connecting the dots to provide actionable insights, empower data-informed decision-making, and drive the data-inspired transformation in the organization.”

Peter Waleczek: Thank you for your time and insights, Dominik.

Share this article
Erhalten Sie innerhalb von zwei Werktagen eine indikative Bewertung Ihres Unternehmens.
Wir berechnen auf dieser Basis gerne einen indikativen Wert Ihres Software- oder Tech-Unternehmens. Für eine genauere Bewertung müssen wir im zweiten Schritt weitere sensible Daten erfragen, für die wir Ihnen vorab eine Vertraulichkeitsvereinbarung zukommen lassen.

*FLEX Capital ist ein Private-Equity-Buyout-Investor mit Spezialisierung auf den Softwaresektor. Wir verfügen über maßgebliche Expertise bei der Unternehmensbewertung in diesem Segment.

Get a reliable valuation of your business within 48 hours.

We are happy to calculate an indicative value of your software or tech company on this basis. For a more detailed assessment, we need to ask for further sensitive data in the second step, for which we will send you a confidentiality agreement in advance.

*FLEX Capital is a private equity buyout investor specializing in the software sector. We have significant expertise in company valuation in this segment..