60-second story
From a complex scenario calculator to clear product requirements
Problem
A new agritech product needed validation in usability and comprehension. Terminology, model logic, manual inputs and unclear calculations made results hard to interpret.
Approach
Briefed by the lead UX researcher, I localized the materials into Polish and moderated seven combined remote interview and usability sessions.
Key insight
The decisive visualization was not more data, but a clear cause-and-effect chain.
Decision
After every few sessions I shared interim findings with the team. Where feasible, critical bugs and UI/UX blockers were fixed before the next interviews.
Outcome
Acute frictions fed into the next prototype version. Deeper patterns around trust, comparison logic and model comprehension went into the strategic synthesis.
01 CONTEXT
A new agritech decision tool needed to show that farmers could understand – and trust – its scenario calculator
A German product team was developing a tool that lets farmers compare a baseline with alternative scenarios and see the financial and modeled impact of changing a practice. Before broader validation or further development, the team needed to know whether Polish farmers could work with the scenario calculator, interpret its results and trust them enough for real planning.
01
Product
A scenario calculator that compares a baseline with alternatives and shows the financial and modeled effects of changing farm practices.
02
Challenge
Terminology, model logic, manual inputs and unclear calculations made it difficult to quickly interpret results.
03
Research setup
Seven combined remote interview and prototype usability sessions in Polish over 2.5 weeks, for a German product team.
02 RESEARCH APPROACH
I handled the local research execution – from briefing and product onboarding to Polish moderation
The lead UX researcher briefed me on objectives, product and test logic. Before the sessions, I familiarized myself with the prototype and translated the research materials into Polish. I deliberately combined the seven sessions as interviews and usability tests: farmers navigated through concrete scenarios, thought aloud and simultaneously explained their existing planning, calculation and trust logic.
01
Current workflow
How are costs, crops and decisions planned today?
02
CONCEPT
What do farmers believe the tool does for them?
03
SCENARIO
Walk through baseline and alternative.
04
RESULT
Have them explain numbers, impact and recommendations.
05
ADOPTION
When would the tool be useful – and what would prevent adoption?
RESEARCH IN PRACTICE
Tech issue is not a lost session
In interview #4, screen sharing failed on the participant’s side. Such technical issues can occur in remote research. The farmer navigated independently and described each step. I intensified think-aloud probing and was still able to capture mental models, uncertainties and interpretations.
This was not a replacement for visual observation, but a deliberate protocol adaptation to the situation.
Evidence boundary: The interviews reveal recurring qualitative patterns and decision logics. They are not a representative market validation and do not demonstrate business impact.
03 RAPID RESEARCH LOOP
Research didn’t end with the last session – findings fed back into the product before the next interview
I documented each session directly alongside the tested product: screens, observations, exact quotes and open questions remained linked. As soon as a signal recurred in subsequent sessions, I flagged it as a recurring pattern. After every few interviews, I distilled the status for the team. This meant no end-of-project report, but a living evidence system that kept research and product iteration synchronized across three update rounds.
01
2–3 SESSIONS
Interview + Usability Test in Polish
02
Interim synthesis
Distill patterns, bugs and blockers
03
Team update
Share urgency and next changes
04
Fix → next round
Integrate relevant UI/UX fixes before subsequent sessions
MICRO · NEXT SESSION
Ensuring immediate testability
Bugs, misleading UI, blocking interactions or necessary UX fixes were flagged early. The product team implemented prioritized changes – where feasible on short notice – before the next sessions.
MACRO · STRATEGIC SYNTHESIS
Don’t fix away the bigger product questions
Value proposition, trust, calculation logic, local fit or multi-season effects were not treated as quick UI fixes. I carried these patterns forward across the interview series in the strategic synthesis.
ROLE BOUNDARY
Study framing and briefing came from the lead UX researcher. I owned product onboarding, Polish localization, moderation and interim readouts. Implementation and fix prioritization rested with the Product, Design and Development team.
METHOD IN PRACTICE · LIVING EVIDENCE
Screens, observations and quotes remained directly linked
Rather than reconstructing raw notes later, I worked directly in the context of the tested pages. Recurring observations were flagged as patterns across sessions. This kept it traceable what was a single observation, an open signal or a recurring pattern.
SESSION-LEVEL EVIDENCE
Example from the live analysis board: tested screens, observations, exact quotes and patterns flagged as recurring remained visible in the same context.
DELIVERY CADENCE · TEAM LEVEL
Three readouts kept research and product in sync throughout the study
After interviews 1–3, 4–5 and 6–7, I distilled new findings, open questions, quick wins and product implications. The team could respond early, while larger themes were deliberately carried forward across the full interview series.
UPDATE 01 · AFTER INTERVIEWS 1–3
First interim readout: quick wins, open questions and initial product implications were shared directly with the team.
UPDATE 02 · AFTER INTERVIEWS 4–5
New findings were checked against existing patterns and translated into concrete product implications for the next round.
UPDATE 03 · AFTER INTERVIEWS 6–7
In the final round, the latest signals were merged with previous patterns and distilled for the overall synthesis.
EXECUTIVE SYNTHESIS · LEAD LEVEL
Finally: a high-level summary for decision-makers with limited time
The final cross-session synthesis reduced the entire interview series to the decision-relevant patterns and implications. It did not replace the collaboration during the study, but gave leads a quick overall picture afterward.
FINAL HIGH-LEVEL SUMMARY
The client-requested final readout distilled the interview series for leads – after three collaborative interim updates, not as a static report without ongoing exchange.
DELIVERY PRINCIPLE
The report was a communication system – not a final document.
The project required high communication density and operational agility: evidence was continuously translated, prioritized and adapted to the right audience – detailed for the working product team, condensed for leads.
04 WHAT WE LEARNED
Four patterns changed how the prototype needed to be conceived as a product
The interviews delivered more than UI feedback. Recurring patterns revealed what information farmers need for a reliable decision – and where the concept was still working against their mental and operational reality.
The value was comparison – not explanation
Farmers described the product primarily as a financial or comparison calculator. The strongest hook was: What changes economically if I change a practice? Usage appeared more seasonal and decision-driven than continuous.
Transparency was the prerequisite for credibility
A result only became reliable when unit, time frame, scope, formulas, data sources, default assumptions, included costs and where applicable CO₂/payment logic were comprehensible.
The comparison had to make cause and effect visible
Baseline and alternative alone were not enough. Farmers wanted to see at a glance: What was changed? What is the financial or modeled effect? Why did the result change?
Local fit was part of the product logic
Hectares, crop shares, multiple crops and fields, local terminology and explainable fit and cost logic determined whether recommendations felt realistic. Mismatched recommendations weakened trust.
EVIDENCE SIGNAL
A recurring pattern was distrust whenever values did not respond as expected after a change or when the underlying calculation could not be explained. In one walkthrough, a farmer interpreted unchanged values as a possible prototype error.
05 KEY INSIGHT
The decisive visualization was not ‘more data,’ but a clear cause-and-effect chain
A scenario comparison only becomes actionable when a change in the model is directly linked to its outcome. For farmers, the core reading logic was: What did I change, what changes as a result in numbers – and what explains this difference?
01
CHANGE
Change practice / assumption
02
DELTA
Show difference from baseline
03
IMPACT
Make financial / modeled effect visible
04
WHY
Explain calculation and rationale
MICROCOPY SIGNAL
‘Compare’ was more concrete than ‘Add scenario’
The language should name the user intent: compare and understand impact – not the product’s internal object logic.
06 TRUST & ADOPTION
For the tool to fit into real planning, three conditions had to be met simultaneously
TRANSPARENCY
Make formulas, sources, defaults and included costs visible or explainable.
Farm fit
Connect hectares, crops, local terminology and realistic recommendations to the specific farm operation.
Workflow fit
Reduce repeated manual input: import, save and reuse data for future scenarios.
INTERVIEW #7 · MODEL LOGIC
Not every effect occurs in season 1
A later interview added an important time dimension: soil and biological effects can develop over multiple seasons. For a credible model, impact should therefore be presented gradually, with realistic delay, uncertainty and where applicable break-even – rather than as immediate ROI.
07 PRODUCT RESPONSE
The interviews produced six requirements for the next product iteration
I translated recurring observations not into a long list of individual UX/UI fixes, but into product requirements that address multiple problems simultaneously.
01
Scenario comparison
Explicitly juxtapose baseline and alternative; make changed inputs, delta, effect and rationale visible together.
02
Calculation transparency
Make formulas, sources, units, time frame, default assumptions and included costs comprehensible.
03
Farm model and localization
Reflect farm reality with hectares, crops/fields and local terminology; account for full Polish localization.
04
Lower input friction
Save, reuse or import data instead of rebuilding the same farm state for every comparison.
05
Explainable recommendations
Justify fit scores, cost logic and recommendations; verify practical feasibility.
06
Time-aware model
Model multi-year effects, delay, uncertainty and break-even where agronomic logic does not permit an immediate effect.
ANONYMIZED RECONSTRUCTION · SCENARIO COMPARISON
Change
→
Delta
→
Effect
→
Rationale
Change, delta, effect and rationale in a single reading logic.
ANONYMIZED RECONSTRUCTION · RESEARCH SYNTHESIS
Observations
→
Pattern
→
Requirement
Observations were distilled into patterns and subsequently into requirements.
STATUS: These are research-based recommendations and requirements. This case does not claim full implementation status.
08 OUTCOME
Research impact emerged simultaneously in the next session and in the longer-term product logic
The synthesis distilled seven interviews into a shared logic for value proposition, scenario comparison, trust and adoption. This made it possible to discuss more concretely which product questions needed resolution before broader validation or further development.
DECISION IMPACT
The team received findings not just at the end. Interim updates made visible which problems were immediately blocking the next research round and which questions needed strategic follow-through.
PRODUCT IMPACT
Critical bugs and necessary UI/UX fixes were implemented between interview rounds by the team, where they could be prioritized on short notice. Larger patterns were translated into product requirements.
RESEARCH IMPACT
The ongoing separation of micro-blockers and macro-themes kept subsequent sessions testable, without reducing deeper questions about value proposition, trust or model logic to quick UI corrections.
Evidence boundary: Documented are the qualitative findings, the ongoing team readouts and short-term fixes between research rounds. A measured post-launch, adoption or business impact is not available.
09 MY ROLE & LEARNINGS
My role was the linguistic, methodological and operational bridge between the German product team and Polish farmers
MY ROLE
Based on the lead UX researcher’s briefing, I independently familiarized myself with the product and prototype, localized research materials into Polish and moderated all seven sessions. In parallel, I kept evidence structured and visible, prepared three interim readouts for the product team and subsequently distilled the full interview series into a high-level synthesis for leads.
Interview Facilitation & Think-Aloud Probing
Qualitative Synthesis Across the Interview Series
Separating UI, Model and Value Proposition Issues
Translation into Product Requirements and Priorities
WHAT I TOOK AWAY AND LEARNED
01
Localization is part of research quality
Research in another language is more than translation. Product terms, moderation logic and follow-up probes need to work naturally in conversation.
02
Explainability is product UX
For calculated results, transparency is not supplementary documentation – it is part of usability and trust.
03
Research delivery is part of the method
Clean analysis alone is not enough in an iterative product project. Evidence must flow at the right cadence and level of detail so teams can act quickly without waiting for the final report.
04
Time can be a UX problem of the model
When real effects are delayed, the presentation of impact must also respect this temporal logic.
PROJECT UXR
Study Framing, Briefing, Overall Research Responsibility
MY RESPONSIBILITY
Product Onboarding, Polish Localization, Moderation, 3 Interim Readouts, Executive Synthesis
PRODUCT TEAM
Prioritization, UI/UX/Bug Fixes, Ongoing Product Decisions for the Prototype
