Unsecured debt purchase: How data becomes a portfolio valuation

At first glance, the data package appears unremarkable: a long Excel list containing millions of data points. Each line represents an outstanding receivable from a delinquent payer. For Malte Janzen, this is where the analysis begins – an analysis that can ultimately result in an investment decision worth millions. 

He has been responsible for the Unsecured Investments Division at EOS since 2025. Together with a team of credit and data analysts, he evaluates receivables portfolios offered for sale by banks across Europe. Malte’s expertise lies in investments. He wrote his doctoral thesis on investment decision-making before going on to work at an M&A consultancy and later at a Hamburg-based corporate group, where he helped build the Infrastructure Projects Division.

“We analyze the receivables package and develop a business plan. On this basis, we make a recommendation on the purchase price to the EOS Board, which then makes the final investment decision,” Malte explains. Each year, EOS evaluates between 500 and 600 receivables portfolios across more than 20 European countries.

The process typically begins when a bank puts a portfolio of unsecured receivables up for sale. The experts at EOS’s national subsidiaries contact the seller and receive an anonymized data package for the due diligence review. On this basis, the EOS company in the respective country prepares a detailed analysis and valuation of the portfolio. 

At the same time, Malte’s team reviews the portfolio from a Group-level perspective: How long has the receivable been overdue? When would it become time-barred? What has been done so far to collect the receivable? The portfolio’s context also plays an important role: Does the portfolio originate from a period of crisis? What does the seller’s customer structure look like? Every portfolio has its own story. Gradually, a complete picture emerges from the many individual data points. “Ideally, we already have experience with the seller. That allows us to assess the customer structure and the quality of the receivables even more accurately.” 

Given the sheer volume of data, manually analyzing each individual receivable would be impossible. “Instead, we begin by identifying key characteristics and grouping receivables accordingly,” says Malte. Mathematical and statistical methods, as well as machine learning models, are used to do this. Drawing on extensive historical data, these models simulate how different groups of receivables are likely to perform over time. 

Additional factors are also incorporated into the analysis, such as changes to the regulatory environment or macroeconomic trends. “With our models, we can map a wide range of different scenarios. This helps us identify which factors have the greatest impact on the portfolio and what to pay attention to later during processing.” 

However, even the most sophisticated models cannot replace expert knowledge. Like Malte, his team members bring many years of experience in credit and data analysis to the job. They also work closely with colleagues from the national subsidiaries. “They have in-depth knowledge of the local markets and are able to put many trends into context more effectively.”

If the investment decision is positive, the analysis does not end there. The portfolio’s actual performance is continuously monitored and compared with the original forecast. If performance deviates from expectations, the underlying causes are investigated. “We feed these insights directly back into our analytical tools, which allows us to continuously refine and improve our valuation process.”