Overview
At the start of 2022, I joined coeo incasso in Rotterdam with a simple question that quickly became larger: how do you acknowledge vulnerable groups in a system not built for them, and remove barriers from the debt trajectory? This thesis lived inside the organization, close to operations, legal, and customer support—every idea had to work in practice.
Instead of adding more rules or more text, the aim was the opposite: lift barriers, reduce cognitive load, and make the next step unmistakable. The result is an inclusive set of service and interaction patterns—language, flows, and small interventions—that help people understand their situation sooner, act earlier, and avoid escalation where possible.
Problem
Most private debt collection assumes high financial literacy and stable circumstances. Real life is different. Letters arrive at the wrong moment. Emails blur together. Legal jargon competes with stress and shame. Vulnerable groups—people with intellectual disabilities, low literacy, multiple languages, or financial stress—are overrepresented and face higher barriers.
Many people wanted to resolve their situation but got stuck at the first message: too many instructions, too many choices, too little empathy. The result is predictable—avoidance, late engagement, and fees that pile up despite willingness to act.
Research approach
I worked across the service: speaking to people who had been through the process, sitting with legal and operations, and auditing every message and screen we could find. We prototyped small, testable changes and ran lightweight usability sessions to see what actually helped.
The method was intentionally practical: interview, rewrite, test; map a flow, remove a step, test again. If someone couldn’t tell me in 30 seconds what they were supposed to do next, it wasn’t done.
Key insights
Several ideas kept coming back:
- People don’t disengage because they don’t care; they disengage because they don’t understand.
- Tone matters. Empathetic, agency-preserving language creates better action than threat-heavy messaging.
- One clear action per step beats three competing ones.
- Clear hierarchy and chunked information reduce abandonment.
- Tiny commitments early on—like choosing a call window—make later plan adherence more likely.
Design principles
These principles guided the work:
- Make the next step unmistakable.
- Preserve agency with bounded choices.
- Explain consequences plainly—not punitively.
- Put help up front (payment plan, pause, talk to a person).
- Write for low literacy and test for comprehension.
- Keep every message self-contained and scannable.
Prototyped interventions
We translated principles into small, shippable changes:
- Plain-language letters and emails with a single next step.
- A guided portal flow that identifies intent: pay now, arrange a plan, or request help.
- Contextual support at the decision moment—not buried elsewhere.
- A simple plan explainer with sliders and immediate feedback on feasibility.
- A status timeline that shows what happened, what’s next, and by when.
Outcomes (pilot metrics)
In pilot tests, we saw earlier engagement and clearer comprehension. People set up plans faster when eligible, and support spent less time clarifying basic steps. Most importantly, participants could explain the process back to us—in their own words.
Deliverables
The work resulted in a communication system (letters, emails, SMS patterns), an interaction model for the client portal, language guidelines, and an inclusion checklist that teams could use to evaluate new ideas before they shipped.
References: TU Delft repository record