Begin with the decision
A research plan should start with what the team needs to decide. The method follows the uncertainty, not the other way around.
User research
User research at HAAM
Research should change what gets built.
HAAM uses research to reduce uncertainty before it becomes expensive. The goal is not a larger research archive. The goal is a clearer product decision, tested early and carried into production.
Know enough to make the next decision responsibly.
The research loop
Research is a loop between the world and the product.
The strongest process moves repeatedly between observation and making. Each pass sharpens the question, exposes another assumption, and gives the team something more concrete to evaluate.
- 01
Frame
Name the decision, the risk, and the assumption that matters most.
- 02
Observe
Study what people do, say, avoid, misunderstand, and work around.
- 03
Interpret
Turn patterns and contradictions into a model the team can challenge.
- 04
Make
Give the interpretation form through flows, language, and prototypes.
- 05
Test
Put the idea back in front of people before confidence becomes expensive.
- 06
Measure
Keep learning from real use after the product leaves the research session.
Eight working principles
Methods matter less than the quality of the next decision.
Research assumptions, not just opinions
People can describe preferences, but products fail because of hidden expectations, missing knowledge, social pressure, and practical constraints. Research should expose those conditions.
Listen for contradictions
What people say matters. Hesitation, workarounds, skipped steps, and choices under pressure often reveal even more.
Treat context as part of the interface
Language, culture, money, place, trust, access, devices, and institutions shape how a product is understood and used.
Combine scale with depth
Surveys reveal patterns. Interviews and observation explain why they exist. Prototypes show whether the interpretation survives contact with use.
Segment by behaviour and values
Age and location are rarely enough. Decisions become clearer when people are grouped by motivations, trade-offs, habits, and confidence.
Prototype before certainty
A prototype turns an abstract opinion into a concrete reaction. People can respond to a flow, a word, a delay, and a consequence.
Carry evidence into production
An insight is only useful when it changes a priority, interaction rule, content model, component, measurement plan, or shipped decision.
Mixed methods, one question
No method sees the whole product.
Each method creates a partial view. Confidence comes from combining views, understanding their limits, and following important contradictions rather than averaging them away.
Ask
Surveys and interviews
Useful for language, motivations, self-reported behaviour, value conflicts, and patterns that deserve deeper investigation.
Watch
Observation and usability testing
Useful for noticing hesitation, false starts, environmental constraints, workarounds, and differences between intention and action.
Make
Concepts and prototypes
Useful for turning assumptions into something specific enough to misunderstand, reject, compare, and improve.
Measure
Analytics and product signals
Useful for seeing what happens at scale, where journeys break, and whether a research-led change survives real use.
Research learned by doing
Green Filter became the deepest research programme behind this practice.
The work explored how young adults in Taiwan make decisions when price, sustainability, trust, and limited attention compete. It moved between survey design, data cleaning, behavioural segmentation, interviews, prototypes, in-person testing, remote self-testing, and repeated redesign.
986
Completed survey responses
876
Filtered Gen Z respondents
3
Research-derived personas
32
Face-to-face prototype tests
100+
Remote self-tests
7
Universities represented in prototype testing
Cross-cultural practice
Context is not a localization layer added at the end.
Living and working across Estonia, Portugal, São Tomé and Príncipe, and Taiwan changed how I understand research. People do not meet a product as abstract users. They meet it through language, institutions, devices, habits, histories, and expectations about who can be trusted.
Cross-cultural research therefore requires more than translating a questionnaire. It requires noticing which assumptions belong to the product team, which belong to the research setting, and which only become visible when the product moves between places.
Localization starts during research. Concepts, examples, scales, categories, and even the meaning of a successful outcome may need to change before the interface does.
What repeated testing taught me
Small moments often carry the biggest product lesson.
A single word can hide a whole feature
Language is not surface polish. An unfamiliar label can make a useful capability invisible, even when the interface is otherwise clear.
Dense evidence can still be unusable
More data does not automatically create more understanding. People need hierarchy, comparison, explanation, and a visible next action.
Surprise exposes a hidden assumption
When a user is shocked by a result, the reaction may reveal a gap between the product model and the person’s existing mental model.
Trust needs provenance, not polish
A confident answer is not enough. People need to understand where information came from, how recent it is, and where uncertainty remains.
Values change what better means
Price, convenience, climate, health, animal welfare, privacy, and risk are weighted differently by different people. Personalization should respect that.
Discovery is part of usability
A feature cannot help anyone who never realizes it exists. Guidance, feedback, and progressive disclosure belong inside the experience.
The first test also tests the test
A session reveals weaknesses in the prototype, but also in the task, question order, explanation, and assumptions of the researcher.
Outliers are not automatically noise
A minority experience can expose an accessibility issue, trust failure, cultural mismatch, or future need that averages make invisible.
Research quality
Evidence needs a visible chain of custody.
Research becomes dangerous when interpretation is presented as observation, confidence is hidden, or an AI summary becomes more authoritative than the people and sources behind it.
- Separate observation from interpretation
- Keep contradictory evidence visible
- Record source, date, sample, and confidence
- Avoid leading participants toward the preferred answer
- Do not confuse a polished prototype with a proven proposition
- Use AI to organize evidence, never to invent participants or erase consent
What research should produce
Not a report that waits to be interpreted.
HAAM turns research into artifacts that guide design, engineering, content, prioritization, and measurement without losing the original evidence.
- A decision map that names the uncertainty
- A behavioural journey grounded in real situations
- Research-derived segments with visible differences
- Prioritized opportunities with explicit trade-offs
- A prototype that makes assumptions testable
- Interaction and content rules for implementation
- A source trail connecting evidence to decisions
- A measurement plan for learning after launch
Evidence before theatre
This is the research practice behind the Why HAAM promise.
Research, analytics, prototypes, and technical constraints should improve the next decision. The same person who hears the evidence stays close enough to carry it into the product.