The method follows the question.
Use the decision and uncertainty to choose the method.
Read the reasoning: Start with the question. Then choose the method. →Approach
Start with the decision. Then choose the research.
Working backward means first agreeing what you may do differently because of the result. We then identify what you need to learn and what evidence would be enough to act. That is how we choose the method, rather than starting with a test and hoping its result will be useful.
- Define the decision
- Identify what is unknown
- Design the research
The decision-first sequence
What each step is doing.
The sequence makes the reasoning visible. It can guide a focused product comparison just as well as a larger research programme.
01Decision
Name the choice the research must inform. Are you selecting a formulation, changing instructions, supporting a claim, or deciding whether to investigate further?
02Uncertainty
Identify what you need to learn before making that choice. Distinguish a missing fact from competing explanations of something you already observed.
03What to measure
Define the outcome that represents the question. A product property, perceived intensity, liking, choice, and continued use answer different questions.
04Comparison
Choose what gives the result meaning: the current product, an alternative formulation, a control condition, a baseline, or another relevant reference.
05People, context & time
Specify whose response matters, under which conditions, and over what period, from first contact to repeated use.
06Evidence needed
Agree what size of effect, level of precision, and degree of uncertainty would justify the decision or intended claim.
07Method & analysis
Choose the study design, sample, measures, controls, and analysis that can answer the question at the depth required.
08Conclusion & limits
Explain what the results support, what they do not support, and how far the findings can reasonably apply beyond the study.
09Action
Translate the answer into a next step: proceed, revise, stop, support a specific claim, or resolve a remaining uncertainty.
The experience model
A product measure is part of the answer.
Products shape the conditions for experience. Understanding the response means studying the product together with the person, the setting, and what changes over time.
Product × person × context × time
The product provides signals.
The person has the experience.
A cup of coffee makes the relationship visible. The same questions apply to other products, services, and encounters.
Product / stimulusProvides physical and chemical signals
SignalsHeat · aroma and taste compounds · surface texture · sound⟶
PersonPerceives, Feels, Interprets
Context shapes the encounter Place · social setting · occasion · information
- BeforeAnticipationExpectations, goals, past experience
- DuringThe encounterPerception, feeling, meaning, response
- AfterMemory & actionWhat is remembered, chosen, or done next
What changes with the person?
Sensory responsiveness, physiology, current state, attention, expectations, and previous experience can change what is perceived and how it is interpreted.
Why context and time?
The setting and occasion can alter the encounter. Within one use episode or across repeated use, adaptation, learning, familiarity, and memory can change the experience again. Earlier experiences also shape future expectations.
Research principles
Checks that keep the conclusion honest.
These are working checkpoints used to frame questions, design studies, review evidence, and interpret results. Each links to a referenced article explaining the reasoning in more depth.
In the project
Apply the principle to a specific design choice, interpretation, or decision.
In the articles
Explore the underlying science, examples, and references.
Browse the Perspectives →Perception is the interface.
Study what people actually perceive and interpret, not only what the product was designed to deliver.
Read the reasoning: Products are designed. Experiences are emergent. →The measure is part of the experiment.
Check whether wording, scales, timing, order, and response format could change the answer.
Read the reasoning: The measure is part of the experiment. →Controlled research isolates. Contextual research tests whether the finding travels.
Use controlled and real-world studies for their different purposes, then connect the evidence deliberately.
Read the reasoning: Controlled research isolates. Contextual research tests whether the finding travels. →Context is not automatically noise.
Treat context as a possible cause when it can change perception, behaviour, or the decision.
Read the reasoning: Temperature Is Not Background →Variation can be error or explanation.
Determine whether differences reflect noise, stable individual differences, changing states, or product–person interactions.
Read the reasoning: The Average Person Is a Statistical Convenience →Liking is evidence. It is not adoption.
Match the outcome to the decision: liking, choice, trial, repeated use, switching, and adoption are not interchangeable.
Read the reasoning: Liking is evidence. It is not adoption. →Statistical significance is not the action threshold.
Define how large, precise, and practically important a result must be before it changes the decision.
Read the reasoning: Start with the question. Then choose the method. →Claims must stay inside the evidence.
Make sure the study design, population, comparison, outcome, context, and time support the exact wording used.
Read the reasoning: What can this evidence actually support? →Uncertainty belongs in the answer.
Report what is known, what remains uncertain, and which limitation matters for the next decision.
Read the reasoning: What can this evidence actually support? →A useful next conversation
What does your next decision depend on?
We can work out what is already known, what still needs to be learned, and how to get a useful answer.