How we segmented 20,000 consumers across 20 European markets and discovered that psychology, not income, predicts financial health
Intrum, Europe's leading credit management company, partnered with FT Longitude to understand the psychological drivers behind consumer payment behaviour across 20 European markets. The challenge was formidable: 20,000 respondents, hundreds of survey variables, and the need to create actionable intelligence that would work consistently across different economic systems, languages, and cultures. Traditional demographic segmentation had hit a wall. Income brackets, age, and employment status didn't explain why consumers in apparently similar circumstances made fundamentally different financial decisions.
We applied the Data Alchemy Framework to compress complexity into clarity, turning raw survey data into four distinct Money Manager personas that now guide Intrum's engagement strategy across every market.
Intrum's analysis teams were drowning in data. They had detailed financial behaviour measurements from 20,000 respondents, but the traditional approach of segmenting by demographics was failing them. A 45-year-old earning 35,000 euros might be financially resilient, while another earning 65,000 might be desperately stressed. Two people with identical salaries made completely different decisions about borrowing, saving, and managing debt.
The real question wasn't "how much do they earn?" It was "what drives their relationship with money?" But that requires moving beyond income brackets into the psychological territory that traditional market research struggled to access. Intrum needed a framework that could identify and act on the psychological patterns underneath the economic data.
Demographic segmentation assumes homogeneity within groups. If you're 40, earn 50,000 euros, and have two kids, you're grouped with everyone else who meets those criteria. But it turns out that's a category without much behavioural coherence. One 40-year-old panics about money and avoids opening statements; another is methodical and comfortably ahead of payments; a third is drowning in debt despite the steady income. Behavioural differences that matter far more than shared salary brackets.
Even the more sophisticated approach of clustering on raw survey responses, which is what most research agencies would offer, has a fundamental flaw: people are poor judges of what drives their own decisions. Rating scales are plagued by bias, compression, and social desirability. When you cluster on those raw inputs, you are grouping people by how they filled in a questionnaire, not by how they actually behave. The segments look clean in a presentation but fall apart the moment you try to act on them.
Intrum needed to answer the underlying question: "Which consumers are genuinely financially healthy, and which are vulnerable, regardless of their income?" The answer required outcomes-based segmentation: modelling the behavioural outcome first, generating individual driver profiles, and then clustering on motivation patterns rather than raw survey responses.
We deployed the five-step Data Alchemy process, each step building on the previous to move from raw data to actionable intelligence. At the heart of this framework is Knowsis's proprietary outcomes-based segmentation methodology: model the behavioural outcome first, infer individual driver profiles for every respondent, then cluster by the shape of those motivation patterns rather than raw survey responses. Steps 2 through 4 below follow this sequence precisely.
The survey wasn't our design, but we worked with Intrum and FT Longitude to ensure it captured the full spectrum of financial health variables. Beyond typical income and credit questions, we mapped spending habits, savings discipline, attitudes toward debt, stress around money, financial planning behaviour, social media influences, childhood money memories, and emotional spending triggers. Over 30 variables, each measuring a different facet of how consumers relate to money in practice.
This is where our outcomes-based approach begins. The 30+ variables couldn't be used as they were. They correlated with each other, some contradicted others, and raw variable counts are impossible to act on operationally. Rather than treating every variable equally, which is what traditional segmentation does, we first modelled the behavioural outcome: financial health itself. We built the Money Management Index, a composite 0-100 score that used advanced machine learning to identify the latent structure in how consumers related to money, compressing dozens of variables into one meaningful dimension. This step is critical because it forces the analysis to answer "what actually matters?" before deciding how to group people. A consumer with a Money Management Index score of 73 was consistently more financially resilient than one scoring 41, across all 20 markets. The index became the analytical anchor for everything that followed.
This is the step that fundamentally separates our approach from traditional segmentation. With the outcome modelled, we generated a unique behavioural fingerprint for every single respondent: an individual driver profile showing the relative influence of each variable on that person's financial health. Not averages across groups. Not correlations across the whole sample. A unique influence profile for each of the 20,000 consumers. Catalyst methodology distinguishes between stated importance (what consumers say matters) and statistical impact (what actually predicts behaviour). Consumers consistently rated "earning enough money" as the top factor in financial health. But at the individual level, the data told a different story. For some consumers, childhood money stress was the dominant driver. For others, emotional spending triggers outweighed salary entirely. Social media pressure predicted debt struggle better than age or employment type in certain profiles. The Catalyst phase didn't just flip conventional wisdom at the aggregate level. It revealed that each consumer's relationship with money is shaped by a unique combination of psychological forces, and that combination is what we needed to segment on.
Here is where the proprietary approach pays off. Traditional segmentation would cluster people by how they answered the survey: their raw scores on spending, saving, and anxiety questions. We clustered on something fundamentally different: the shape of each person's behavioural fingerprint. Two consumers might both score moderately on financial anxiety, but if one is driven primarily by childhood money stress and the other by social media pressure, their motivational profiles are completely different, and they belong in different segments. By clustering on the pattern of influences rather than the raw responses, four distinct Money Manager profiles emerged, each characterised by genuinely different psychological relationships with money. These profiles cut across geography, age, and income. A self-disciplined optimiser in Warsaw behaves like one in Barcelona, even though their economic contexts differ dramatically. That cross-market consistency is a direct result of segmenting on underlying motivations rather than surface-level survey responses.
Raw clusters are analytically pure but operationally abstract. Staff at Intrum's call centres, compliance teams, and policy divisions needed to understand these segments as people, not data points. We developed AI-generated personas for each Money Manager type, bringing them to life with realistic behavioural profiles, communication preferences, typical stressors, and effective engagement strategies. These personas are now the language Intrum uses across all 20 markets to discuss and serve consumers.
This was the most counterintuitive finding. The Money Management Index revealed that 21% of consumers earning average salaries fell into the most financially vulnerable group, while only 26% of high earners qualified as truly resilient. Income matters, but it doesn't determine behaviour. A person earning 30,000 euros with strong financial discipline and low stress is more resilient than someone earning 60,000 who lives in constant anxiety about money. For Intrum, this meant they couldn't segment customers by income alone. They needed to measure actual behaviour and psychology to understand who was struggling.
The Catalyst analysis showed that psychological factors were far stronger predictors of financial vulnerability than economic circumstances. Consumers who grew up in households where money caused stress were three times more likely to struggle financially as adults, regardless of current income. Emotional spending patterns, social media influence, and anxiety around debt matters more than what someone earned. This shifted the conversation from "how much do they earn?" to "what's their relationship with money?" It also shifted Intrum's engagement strategy fundamentally, moving from purely economic solutions toward psychological support, clearer communication, and empathetic guidance.
The most powerful operational finding was that these behavioural profiles worked consistently across all 20 markets. Instead of maintaining separate segmentation frameworks for Spain, Poland, and Sweden, Intrum could use one persona-based system. A self-disciplined optimiser in any market prefers minimal contact, digital tools, and straightforward communication. A financially fragile consumer everywhere needs human contact, jargon-free guidance, and empathetic support. This consistency across cultures meant one framework, one set of engagement strategies, and one way of thinking about consumer relationships across all European operations.
Financially resilient, proactive, minimal anxiety about money. Prefer digital tools, minimal contact, and straightforward information. These consumers rarely need support, and when they do, they want efficiency and data, not hand-holding. Engagement strategy: keep it simple, digital-first, respect their time.
Careful with money, anxious about the future, build safety nets methodically. They want reassurance and clear planning frameworks. These consumers respond well to structured guidance, savings products, and explicit timelines. Engagement strategy: provide certainty, transparent processes, and evidence that their caution is justified.
Struggling with income instability, stretched between necessary expenses and limited cash. High anxiety but often working hard to manage. They need flexible solutions, accessible communication, and recognition of their effort. Engagement strategy: acknowledge their struggle, offer practical help, and build trust through reliability.
Most vulnerable segment. Often grew up with financial stress, carry childhood money anxiety into adulthood. High emotional spending, avoidance of financial information, deep anxiety. They need human connection, non-judgmental support, and clear jargon-free guidance. Engagement strategy: build psychological safety first, then address practical solutions.
The Money Manager personas moved from analysis into operation. Intrum teams across 20 markets now use these profiles to:
Perhaps most importantly, the Money Management Index became an operational tool. It's now integrated into Intrum's global consumer tracking programme, deployed across dozens of markets to assess strategic importance and prioritise which segments and geographies require most attention. An analytical output became an infrastructure for decision-making at global scale.
The current framework measures behavioural financial health and groups consumers by their psychological relationship with money. The next layer would be motivational profiling. Our Impulse Engine framework could overlay emotional drivers onto these behavioural profiles, revealing not just "which persona are you?" but "what would actually move you to change behaviour?" A cautious saver might be motivated by security; a self-disciplined optimiser by efficiency; a fragile consumer by immediate relief. That level of personalisation exists in our analytical toolkit but was beyond the scope of this project. It would create even richer engagement intelligence, moving from "this is your type of money manager" to "here's what would actually matter to you."
The full European Consumer Payment Report 2025 is publicly available on Intrum's website.
