Summary
This article explains the decision framework behind the Research Method Explorer: how the generative/evaluative split, four research intents, components of experience, and measurement layers connect your research question to the right method, instruments, and study configuration.
The first question: do you have a solution?
Every research project starts with a fork in the road. If you don't yet have a product, prototype, or concrete solution, you're in generative research territory: understand the problem space, discover unmet needs, generate ideas for what to build. If you do have something (a wireframe, a working product, a competitor's interface), you shift into evaluative research, where the goal is to assess how well that solution works.
This single distinction determines the direction of your study. Generative research asks "What should we build?" Evaluative research asks "Does this work?" Mixing them up (running a usability test when you should be doing discovery, or conducting interviews when you should be measuring) wastes time and budget while producing answers to the wrong question.
Four research intents
Within the generative and evaluative branches, the tool distinguishes four intents that map to different method families.
Discover means exploring a problem space before you have a solution. The methods are qualitative and open-ended: contextual inquiries, ethnographic observation, focus groups, and exploratory interviews. The output is themes, mental models, unmet needs, and opportunity spaces. Discovery research is how you avoid building the wrong thing.
Evaluate means assessing how people experience something you've built. Usability testing, heuristic evaluations, card sorting, and tree testing all fall here. The critical sub-question is which aspect you're evaluating: findability, task efficiency, emotional response, or overall satisfaction. Different components of experience require different techniques and instruments.
Decide uses structured methods to reduce uncertainty around a specific choice. Feature prioritization through MaxDiff or conjoint analysis, pricing research, concept testing, product-market fit assessment: the output is decision-ready data, often quantitative, that answers "which option?" and "is there demand?"
Measure means quantifying experience with standardized instruments. The System Usability Scale, Net Promoter Score, Single Ease Question, task success rates, and UX benchmarks produce numbers you can track over time, compare across products, and report to stakeholders.
Components of Experience
When you're evaluating a product, not everything is "usability." The Components of Experience model provides a hierarchy. At the foundation: accessibility and basic functionality, prerequisites that must work before anything else matters. The pragmatic level covers classical usability: can people find things, complete tasks efficiently, and recover from errors? The experiential level addresses aesthetics, emotional response, trust, and overall interaction quality.
This hierarchy matters for method selection. Findability problems show up in tree tests. Task efficiency surfaces in moderated usability testing. Emotional response requires different instruments entirely: reaction cards, AttrakDiff, or the UEQ. The interactive tool above walks you through selecting the relevant components.
Layers of experience and measurement
The Measure path maps to a layered model. At the task level, you measure whether people can complete specific actions: task success rates, time on task, error rates, and the Single Ease Question. At the product level, instruments like the System Usability Scale or User Experience Questionnaire assess the overall experience. At the customer experience level, relationship metrics (Net Promoter Score, Customer Satisfaction Score, Customer Effort Score) capture how people feel about the brand beyond any single interaction. At the market fit level, product-market fit surveys and willingness-to-pay analysis assess whether the product meets a real need.
Each layer requires different instruments and different sample sizes. Task-level measurement works with small samples and produces actionable micro-insights. Market-fit measurement needs larger samples and produces strategic evidence. Using a task-level instrument to make a market-level decision produces unreliable conclusions because the evidence doesn't match the scope of the question.
The research profile
Every recommended study configuration produces a research profile: a five-dimensional characterization of where the study sits on the spectrum of research approaches.
The qualitative to quantitative dimension captures whether you're working with words or numbers. The generative to evaluative dimension reflects whether you're exploring a problem space or assessing a solution. The active to passive dimension distinguishes direct engagement (interviews, tests, surveys) from observation without intervention (analytics, passive data collection). The depth to scale dimension captures the trade-off between understanding a few people deeply and measuring many broadly. The attitudinal to behavioral dimension reflects whether you're capturing what people say or what they do.
No profile is "better" than another. A deep, qualitative, generative study is exactly right for early-stage discovery. A broad, quantitative, evaluative study is exactly right for benchmarking a mature product. The profile makes these trade-offs visible so you can confirm they match your intent, or adjust course.
What Comes Next
This tool gives you a starting point: a method, instruments, and a study configuration matched to your research intent. The next step is a research plan that operationalizes these recommendations: research questions, sampling strategy, data collection protocol, and analysis approach. For a detailed calculation of how many participants you'll need, use our Sample Size Calculator.
For the method building blocks behind the recommendations, see Building Blocks and Core Methods. To assess the business value of your selected method, try the Research Value Calculator.