Old-software testing
Usability testing the existing software with current users.
Refeyn / Macro Mass Photometry / KaritroMP
A new instrument A new software experience.
KaritroMP was a new benchtop instrument for analysing viral vectors and viral-like particles. It was also the chance to design the software properly from the ground up.
- Role
- Senior Product Designer
- Scope
- UX research, UI, field testing
- Outcome
- New software shipped with the new Karitro instrument
The starting point
When I joined Refeyn, the existing software used lightly customised default QT components. It worked, but looked dated and behaved inconsistently.
KaritroMP created a clean moment to rethink the workflows and lay foundations for the design system the company had been missing.
My role
Wireframes & prototype
Wireframes for the new software and a clickable prototype.
Prototype testing
Usability testing the prototype with target users.
High-fidelity design
Production UI design across screens and states.
Design system foundations
Established components, styles, and documentation in parallel.
Beta testing
Field testing on real instruments with real researchers.
Testing the old software
Before designing anything new, I ran moderated sessions with current users on the existing software. The test analysis tracked tasks, success rates, errors, and quotes.
Findings were mapped into an affinity diagram to surface the main pain points and prioritise the workflow issues worth solving first.
Wireframes & prototype
I designed wireframes, built a clickable prototype, and tested it with target users. A second affinity map captured what came back from the prototype sessions.
Building the Design System in parallel
The new software was also the moment to lay the groundwork for Refeyn's design system - extending the brand palette, choosing a typeface, defining components, and documenting it all for engineering. Each hi-fi screen informed the system; the system kept the screens consistent.
That work is a case study of its own.
Read the design system case studyHigh-fidelity design
The hi-fi design focused on four things: making it easy to define an experiment, guiding setup, keeping the user informed and in control, and making complex data intuitive to read.
Setup with guided instructions
Easy to define an experiment
Users informed and in control
Intuitive visualisation of complex data
Beta testing
8 participants
2 institutions
7 days
real researchers
Authentic experiments, their own samples. I observed end-to-end workflows, measured task success and error frequency, and assessed user confidence through post-task surveys.
Three improvements shipped because of beta.
Alignment Tool
Problem
Sample order
Users got lost while pipetting samples into the wells.
SolutionWith R&D and Marketing, we redesigned the alignment tool to include a well-order diagram printed on it.
Problem
Sample concentration
Users struggled to assess whether their sample was at the right concentration before running the experiment.
SolutionWith Software, we added a Concentration Check feature — users select wells to test before committing to the full experiment.
Problem
Calibration
Users struggled to assess whether a calibration was acceptable.
SolutionWe added examples of acceptable and unacceptable calibrations, plus clear criteria to judge against.
Continuous improvement
After launch, with the Product Manager, I set up a structured feedback system — customer surveys, sales-team input, direct client interviews, and software telemetry — all centralised in Confluence.
The improvement plan prioritised changes by user need. Shipped so far: workspace saving, new graph types, titer calculation.
Feedback channels
- Customer surveys
- Sales team input
- Direct client interviews
- Software telemetry
Process
- Insights centralised in Confluence
- Prioritised by user need
- Ongoing improvement plan, reviewed with the PM
Shipped so far
- Workspace saving
- New graph types
- Titer calculation
Final reflection
Designing the software in step with the new instrument meant every screen could be tested in real lab conditions, with real samples, before launch. The design system grew alongside it, ready to support the touchscreen products that came next.
OTHER PROJECTS GLIMPSES