How to scope a generative research study when stakeholders only want validation research

You have been asked to run a research study. The brief from your stakeholder is clear: test this design, validate this concept, confirm this direction. They have a solution in mind and they want evidence that it works.

But something feels off. Maybe the problem has not been well defined. Maybe the team is operating on assumptions about user needs that have never been examined. Maybe the proposed solution is answering a question nobody has actually asked users yet.

This is one of the most common tensions in applied research. Stakeholders ask for validation, but the situation calls for generative work. The challenge is not just methodological—it is political, relational, and practical. Scoping a generative study under these conditions requires a specific set of skills that go beyond writing a good research plan.

Why stakeholders default to validation research

Before trying to change the request, it helps to understand why stakeholders gravitate toward validation research in the first place. Their preference is rarely arbitrary.

Validation feels faster and more certain

Validation research has a clear input (a design, a prototype, a concept) and a clear output (it works or it does not). The deliverable is easy to imagine and the timeline feels contained. Generative research, by contrast, can feel open-ended. Stakeholders who are under delivery pressure often see discovery work as a detour rather than a shortcut.

The solution already has organizational momentum

By the time a stakeholder asks for validation, the team has often already invested time designing, debating, and aligning on a direction. Suggesting that the team pause to explore the problem space can feel like a challenge to decisions that have already been made. The sunk cost is real, and the political cost of revisiting those decisions is something stakeholders are understandably reluctant to pay.

Generative research is harder to scope and sell

Validation research is easy to describe in a stakeholder update: "We tested the new checkout flow with eight users and here is what we found." Generative research sounds less concrete: "We explored how small business owners think about inventory management." Stakeholders who need to justify research investment to their own leadership find validation studies easier to explain.

Past experience with poorly scoped discovery work

Some stakeholders have been burned by previous generative research that was too broad, took too long, or produced insights the team could not act on. If their mental model of generative research is "three months of interviews that resulted in a journey map nobody looked at," their reluctance is reasonable.

How to tell when the situation actually calls for generative research

Not every validation request is misguided. Sometimes the problem is well understood and the team genuinely needs to test a specific solution. The skill is in recognizing when the request for validation is masking a deeper gap in understanding.

Here are some signals that generative work is needed:

The team cannot articulate the user problem clearly. If you ask three stakeholders what problem the solution is solving and you get three different answers, the problem has not been defined well enough to validate anything against it.

Assumptions about user needs are untested. The proposed solution rests on beliefs about what users want, how they behave, or what frustrates them—but those beliefs have never been checked against real user data.

The solution scope keeps shifting. When teams are constantly changing what they want to validate, it often means the underlying problem is not stable enough to anchor a solution.

Previous validation research produced confusing results. If a usability test showed that users could complete the task but did not seem to care about the feature, the issue may be with the value proposition rather than the interface.

The team is entering an unfamiliar domain. New markets, new user segments, or new problem areas all warrant generative exploration before solution design begins.

Reframing the conversation with stakeholders

The goal is not to win an argument about research methodology. It is to help stakeholders see that generative research reduces risk and increases the chances that their solution will succeed. Here are approaches that work in practice.

Start with their goal, not your method

Stakeholders care about outcomes: shipping something users want, hitting a metric, reducing churn, entering a new market. Start from the outcome they care about and work backward to the research approach that best supports it.

Instead of saying "we need to do generative research first," try something like: "You want to make sure this feature drives adoption. Before we test the design, I want to spend a week understanding how users currently solve this problem so we can make sure we are testing the right thing."

This positions generative work as a precaution, not a delay.

Quantify the risk of skipping discovery

Stakeholders understand risk in terms of time and money. Frame the cost of skipping generative research in those terms:

• Building a feature that solves the wrong problem wastes the entire development investment, not just the research timeline.
• Validation research on a flawed concept produces misleading confidence—the team believes users want something they do not.
• Post-launch pivots are far more expensive than pre-build discovery.

If you have examples from your own organization where this has happened, use them. Concrete precedent is more persuasive than abstract principle.

Propose a tightly bounded study

One of the most effective ways to make generative research palatable to skeptical stakeholders is to scope it so tightly that it feels low-risk.

• Limit the timeline. A one-week sprint of five to eight interviews is enough to surface meaningful patterns without feeling like a long detour.
• Limit the questions. Instead of proposing to "explore the problem space," commit to answering two or three specific questions that the team needs answered before they can design with confidence.
• Commit to a clear deliverable. Tell stakeholders exactly what they will receive and when. A one-page summary of findings with design implications is more useful than a 40-slide deck.

Offer a combined approach

If the stakeholder is unwilling to delay validation entirely, propose a study that includes both generative and validation components. Structure each session so the first half explores the user's context, workflow, and unmet needs through open-ended conversation, and the second half gathers feedback on the proposed design.

This approach has a real methodological trade-off—showing a prototype can anchor participants' thinking and limit the generative insights—but it is often the pragmatic compromise that gets discovery work into a project that would otherwise have none. Always run the generative portion before showing any solution artifacts.

How to scope the study itself

Once you have buy-in for generative work, the next challenge is scoping it so it produces actionable insights within the constraints you have negotiated.

Define what you need to learn, not what you need to prove

Validation research starts with a hypothesis to test. Generative research starts with questions to answer. Write your research questions as genuine open-ended inquiries:

• How do users currently approach this task, and what are their biggest pain points?
• What workarounds have users developed, and what do those workarounds reveal about unmet needs?
• What criteria do users use when evaluating solutions in this space?

Avoid research questions that smuggle in assumptions, like "How do users feel about our proposed approach to X?" That is a validation question wearing a generative disguise.

Choose the right number of participants

For a tightly scoped generative study, five to eight participants from a single user segment is usually sufficient to identify the dominant patterns. If you are exploring multiple distinct segments, plan for five to eight per segment and be transparent with stakeholders about why.

Plan for analysis from the start

One reason generative research sometimes fails to deliver value is that analysis is treated as an afterthought. Before you begin recruiting, know how you plan to synthesize your data.

Decide on your analysis approach—thematic analysis, affinity mapping, or another framework—and build that time into your project plan. Tools like Dovetail can help significantly here, giving you a central place to tag, organize, and synthesize qualitative data across interviews so that patterns emerge faster and your findings are traceable back to source material.

Write a one-page research brief

Document the following in a single page and share it with your stakeholders before fieldwork begins:

• Background: What prompted this research and what the team currently believes
• Research questions: The two to three specific questions you will answer
• Method: How you will conduct the research (interviews, contextual inquiry, diary study, etc.)
• Participants: Who you will talk to and how many
• Timeline: When fieldwork, analysis, and readout will happen
• Deliverable: What the stakeholder will receive and in what format

This brief serves as a contract between you and your stakeholders. It makes the scope visible and gives them confidence that the work will not expand indefinitely.

Presenting findings so stakeholders act on them

Generative research is only as valuable as the decisions it informs. How you communicate findings matters as much as the quality of the research itself.

Lead with implications, not observations

Stakeholders do not need a comprehensive account of everything you heard. They need to know what the findings mean for the decisions they are making. Structure your readout around implications:

• "Users are not aware this problem exists, which means the feature as currently scoped would need significant onboarding investment."
• "Users have developed effective workarounds, so the value proposition needs to be substantially better than their current approach to drive adoption."

Connect findings to the original validation request

Remember that your stakeholder originally asked for validation research. Close the loop by explaining how the generative findings should inform what gets validated next. You might recommend changes to the design before testing, suggest validating a different concept entirely, or confirm that the original direction is sound and ready for usability testing.

This positions generative research as a precursor to validation rather than a replacement for it, which reinforces the idea that both types of research serve the same goal.

Make the data accessible

Stakeholders are more likely to trust and act on findings when they can see the evidence themselves. Sharing tagged and organized research data—interview clips, highlighted quotes, and thematic clusters—builds credibility in ways that a summary alone cannot. Dovetail is designed for exactly this kind of transparency, making it straightforward for stakeholders to explore the underlying data behind your conclusions without needing to sit through every interview.

Building a long-term case for generative research

Each well-scoped generative study that produces actionable insights makes the next one easier to propose. Over time, you are building a track record that demonstrates the value of understanding problems before jumping to solutions.

A few practices that help:

• Track decisions influenced by research. Keep a simple log of product decisions that were informed by generative findings. When a stakeholder asks why they should invest in discovery, you can point to specific outcomes.
• Share participant quotes and clips regularly. Exposing the broader team to real user language and behavior—outside of formal readouts—normalizes curiosity about user needs and makes generative research feel like a natural part of the product process.
• Involve stakeholders in research sessions. Inviting a product manager or designer to observe even one interview gives them firsthand exposure to the kind of insight generative research produces. Direct observation is more persuasive than any slide deck.

The real goal: building a balanced research practice

The tension between generative and validation research is not a problem to solve once—it is an ongoing negotiation that reflects the pressures and priorities of your organization. The goal is not to eliminate validation research or to insist on discovery before every decision. It is to build a practice where the right type of research is applied at the right time, and where stakeholders trust the research team to make that call.

Every time you successfully scope a generative study within a validation-oriented culture, you expand the organization's understanding of what research can do. That is a compounding return worth investing in.