5 research methods every PM should know
A practical guide to the five research methods that give PMs the clearest signal at each stage of the product lifecycle.
5 research methods every PM should know
The five research methods product managers use most are: user interviews, concept testing, usability testing, surveys, and Jobs-to-Be-Done research. Each one answers a different kind of question, and knowing when to reach for which method is one of the most practical skills a PM can build.
This guide covers each method, when to use it, what good output looks like, and how to get signal quickly without a dedicated research team.
Why research methods matter for PMs
Most product failures are not engineering problems. They are research problems: the team built something users did not actually want, could not figure out how to use, or would not pay for. Good research, done at the right stage, eliminates most of that waste.
The challenge is that PMs are usually working with limited time and no dedicated researcher. That means choosing the right method matters even more. A usability test run at the discovery stage gives you the wrong information. A concept test run after you have already shipped is too late. Method selection is half the work.
For a broader grounding in how research fits into product management, the user research in product management overview is a useful starting point.
Method 1: User interviews
User interviews are one-on-one conversations designed to understand how users think, what they care about, and what problems they experience. They are the foundation of qualitative product research.
When to use them:
- Discovery: understanding the problem space before you have a solution
- Validation: testing assumptions embedded in your roadmap
- Post-launch: understanding why adoption is lower or higher than expected
What you get: You get mental models, language users use to describe their problems, and unexpected insights that no survey could surface. The signal is rich but not statistically representative.
How to run them as a PM:
- Write a discussion guide with 5 to 8 open-ended questions focused on behavior, not opinions.
- Recruit 5 to 8 participants per user segment using a panel platform.
- Keep sessions to 30 to 45 minutes. Record with consent.
- Debrief immediately after each session while observations are fresh.
AI-moderated interview platforms now let you run asynchronous versions at scale. A participant answers your questions via voice or text at their own pace, and the AI probes for depth on interesting answers. This is useful when scheduling is a bottleneck or when you need to cover more than a handful of participants quickly.
A step-by-step guide to structuring interview sessions is available at how to conduct user interviews.
Common mistakes:
- Asking leading questions (“Don’t you find it frustrating when…?”)
- Spending more than 10% of the session talking yourself
- Treating five interviews as statistically valid data for executive presentations
Method 2: Concept testing
Concept testing lets you validate an idea, feature, or positioning message before you build it. You show participants a description, a mockup, or a simple prototype and measure whether they understand it, value it, and would use it.
When to use it:
- Before starting a build sprint, to confirm the concept resonates
- When choosing between two or three different product directions
- When testing messaging before a launch campaign
What you get: Quantitative scores for comprehension, appeal, and purchase intent, plus qualitative explanations of why something does or does not land.
How to run it: Keep concepts simple: a one-paragraph description or a single screen mockup is enough for most concept tests. Ask participants to describe what they think the product does before asking what they think of it. That comprehension check reveals messaging gaps you would otherwise miss.
A full walkthrough of concept testing approaches is at concept testing: how to validate ideas before development.
What good output looks like: A concept test report should include a comprehension rate (percentage of participants who accurately described the product’s function), an appeal score, the top three reasons for and against adoption, and direct participant quotes to support each finding.
Method 3: Usability testing
Usability testing measures whether users can accomplish specific tasks using your product. It uncovers friction, confusion, and drop-off that analytics alone cannot explain.
When to use it:
- Before shipping a new flow or feature
- After a redesign, to confirm the new version performs better
- When analytics show a drop-off at a specific step but you do not know why
Two main formats:
| Format | Best for | Time required |
|---|---|---|
| Moderated (live, with a facilitator) | Complex flows, exploratory diagnosis | 45-60 min per session |
| Unmoderated (self-guided, recorded) | Known flows, fast benchmarking | 15-20 min per session |
Moderated sessions give you the richest signal because you can ask follow-up questions in real time. Unmoderated sessions scale faster and are cheaper per participant.
Sample size guidance: For qualitative diagnosis, five participants per segment reveals the majority of usability issues. For quantitative benchmarking (task completion rates, time on task), plan for 20 to 30 participants per segment.
For a detailed breakdown of usability testing formats, see usability testing methods: how to choose the right framework and the how to do usability testing guide.
Common mistakes:
- Facilitating in a way that helps users rather than observing them struggle
- Testing a polished prototype instead of the actual product
- Running tests after development is complete, when it is expensive to change anything
Method 4: Surveys
Surveys let you collect structured data from a large number of people quickly. They are the right tool for quantifying something you already understand directionally.
When to use them:
- Measuring customer satisfaction (CSAT, NPS, CES)
- Validating whether a qualitative finding applies broadly
- Prioritizing features across a large user base
- Segmenting users before recruiting for deeper research
When NOT to use them: Do not use surveys to understand a problem you have not already explored qualitatively. You can only ask about things participants can articulate, and users often cannot articulate their real reasons for behavior. Surveys confirm hypotheses. They rarely generate them.
Survey design basics:
- Keep surveys to 5 to 10 questions for response rates above 40%.
- Use closed questions for quantitative analysis, but add at least one open-ended question to capture unexpected signals.
- Pilot your survey with three to five colleagues before sending it to real users. Confusing questions become obvious fast.
Platforms to consider: Typeform, Google Forms, and SurveyMonkey handle most PM survey needs. For NPS specifically, tools like Delighted or Sprig provide in-product delivery.
Method 5: Jobs-to-Be-Done research
Jobs-to-Be-Done (JTBD) research, popularized by Clayton Christensen, frames user behavior around the progress users are trying to make in their lives. Instead of asking “what features do you want?”, JTBD research asks “what were you trying to accomplish when you hired this product?”
When to use it:
- When defining a new product strategy or roadmap
- When you want to understand switching triggers (why users left a competitor)
- When your feature requests feel scattered and you need a unifying framework
What you get: A structured understanding of the circumstances, motivations, and desired outcomes that drive product adoption. This becomes the foundation for product briefs that actually map to user needs.
How to run a JTBD study: The core interview format asks participants to walk through a recent purchase or switching moment in detail: what triggered the search, what alternatives they considered, what pushed them to decide, and what they hoped would change after using the product. This is called the “switch interview” and takes about 45 to 60 minutes.
A guide to applying JTBD in user interviews is at Jobs-to-Be-Done research: how to apply JTBD in user interviews.
Output: A set of job statements in the format: “When [situation], I want to [motivation], so I can [desired outcome].” These statements replace feature-level roadmap items with needs-based ones.
How to choose the right method
The simplest way to choose is to match the method to the question type:
| Question type | Best method |
|---|---|
| What problems do users have? | User interviews |
| Will users value this idea? | Concept testing |
| Can users use this feature? | Usability testing |
| How widespread is this pattern? | Survey |
| Why do users hire or fire this product? | JTBD research |
Most product cycles benefit from a mix of methods at different stages. Discovery calls for user interviews and JTBD. Ideation and prioritization benefit from concept testing. Pre-launch and post-launch both benefit from usability testing and surveys.
For a more detailed decision framework, how to choose the right user research methods covers the key variables including timeline, budget, and research maturity.
Running research without a dedicated researcher
Most PMs do not have a researcher on the team. That is not a blocker, but it does require a disciplined approach:
Panel access matters. Recruiting the right participants is usually the biggest time sink. Using a platform with a pre-verified panel removes weeks of back-and-forth. Platforms like CleverX provide access to 8M+ verified B2B and B2C participants across 150+ countries, with results available in days rather than weeks. For B2B research especially, being able to specify exact job titles, industries, and company sizes saves significant screening time.
Keep studies small and frequent. Five-participant usability tests run every two weeks give you more signal than a 50-participant study run once a quarter. Continuous small-batch research is more actionable than periodic large studies.
Timebox analysis. Set a two-hour limit for analyzing any single study. Write findings in plain language, not research jargon. Include direct quotes. Make one primary recommendation per study.
The Nielsen Norman Group’s guidance on five-user tests is a useful reference for justifying small-sample qualitative studies to skeptical stakeholders.
Frequently asked questions
Which research method is best for validating a product idea before building?
Concept testing is the most direct method. Show potential users a description, mockup, or prototype of the idea and measure their intent to use it, perceived value, and willingness to pay. Running concept tests before a single line of code is written saves significant rework later.
How do PMs run user interviews without a research team?
PMs can run interviews themselves by preparing a short discussion guide, recruiting through a panel platform, and keeping sessions to 30-45 minutes. AI-moderated interview tools now let you run asynchronous interviews at scale, removing scheduling bottlenecks entirely.
What is Jobs-to-Be-Done research and why do PMs use it?
Jobs-to-Be-Done (JTBD) research focuses on the underlying goal a user is trying to accomplish, not on features or demographics. PMs use it to identify unmet needs, understand switching triggers, and write product briefs that map to real user motivations rather than assumed ones.
When should a PM use a survey versus a user interview?
Surveys work when you need to quantify something you already understand qualitatively, such as satisfaction scores, feature prioritization votes, or demographic breakdowns. User interviews work when you need to understand why users behave a certain way or uncover problems you have not anticipated yet.
How many participants do you need for usability testing?
For qualitative usability testing, five participants per distinct user segment is a widely cited baseline that surfaces most critical usability issues. For quantitative benchmarking, 20 to 30 participants per segment gives you statistically stable task completion and time-on-task data.
Can PMs combine multiple research methods in a single study?
Yes, and it often makes sense to. A common pattern is to run a survey to segment your audience, then follow up with user interviews for the most interesting segments, and finish with a usability test to validate specific design decisions. The key is to match each method to a distinct question.