I turn unstructured customer text into evidence teams act on.
Quantitative user researcher, eight years across European tech — Babbel, Qonto, DeepL, Deutsche Glasfaser. I work on the questions that set direction: who your customers actually are, what the silent majority of them are telling you, and what follows from that. Based in Germany, open to remote roles across the EU.
What I do
Research that scales past the sample size
Most companies have far more customer evidence than anyone can read, and make decisions on the fraction someone had time to skim. I close that gap — and I work on the questions that set direction, not the ones that optimise a button.
Foundational research
Segmentation, customer understanding, and the market questions a roadmap rests on. The work that decides what to build, rather than measuring something already shipped. Survey design, segmentation, driver analysis, brand and experience tracking.
Text modeling & classification
Pipelines that label, cluster and route open-ended feedback at volume — taxonomy design through to a model your team can rerun weekly. Python for building, R for statistics, SQL for the user data underneath.
Evidence that travels
Findings framed as decisions rather than decks, and built so they hold up when someone senior pushes back. Research that changes a plan, not research that gets filed.
Selected work
Recent projects
- An insight programme that runs itself 10,000 survey responses a month — anonymised, classified, scored and fed back into Qualtrics without manual handling
- Launching DeepL Voice Mixed-method B2B research behind the launch of DeepL Voice
- Brand and experience tracking across 16 markets Brand, CSat and NPS programmes for a consumer subscription business operating in six core languages
About
Short version
I'm a psychologist by training — two master's degrees in intercultural and economic psychology — who spent the last eight years doing research inside European tech companies with large, multilingual, mostly silent customer bases.
That constraint is what pushed me toward text. When you can't interview your way to an answer, you have to build something that reads at scale.
Independent work
Need a pipeline built?
I take on a small number of independent projects — text classification, taxonomy design, and feedback analysis for teams that have the data but not the time.
If you want to see it work before committing to anything: send a sample of your text data and I'll return a working classification, at no cost.
Contact
Get in touch
Hiring for a research or insights role, or sitting on a pile of customer text you haven't been able to use — either way, I'd like to hear about it.