Skip to article

Planning a study

Build a study plan around a goal, scenarios, questions, and hypotheses.

Planning a study

From a business problem to a research question

For research to be useful, you need a plan — a research design. It shows in advance which answers you will get and how they will help you make a decision.

This document connects all parts of the study and makes it meaningful: you spend time not for the sake of “process,” but for clear numbers, insights, and business value.

Open the research brief template

Business problem

The first step in any study is to understand the business problem. Why is the study needed? What problem does it solve? Which product priorities is it connected to?

Focusing on the business problem helps you:

  • set priorities correctly
  • choose the right research method
  • focus only on what truly moves the team forward

Talk to the stakeholder, or ask yourself, and try to formulate the business problem as best you can. The main thing is to keep the business interests in focus, even if the wording is not perfect yet.

Example questions for clarifying the business problem
  • Why is this study needed now? What changed?
  • What business problem can the study help solve?
  • How does the problem affect the business and its metrics today?
  • How will we know the problem is solved, and what will change in the metrics?
  • What will happen if we leave the problem unresolved?
  • How does solving it relate to OKRs and quarterly or annual goals?

Dasha Lupakhina on research and business decisions

Main research goal

When the business problem is clear, move to the next step: define the main research goal.

Example questions for clarifying the main research goal
  • What exactly do we want to understand about users or their experience?
  • What aspect of user behaviour or perception is unclear?
  • What uncertainty do we want to resolve?
  • At what point in the user journey or scenario does the question arise?
  • What mechanisms, motives, barriers, or behaviours do we want to understand?
  • Is there something we observe without understanding why it happens?
  • What exactly should be clearer after the study?
  • Which focus matters most: motivation, perception, context, decision-making, or behaviour?

These answers can suggest dozens of hypotheses and research questions. Choose the main one.

As an example, let’s take business problems and add the main research goal to each of them.

Scenarios

If you are testing a prototype or a real website, start with product-use scenarios — larger user actions that people perform to solve their task.

This will help you:

  • identify the key tasks that are most important to test first
  • not forget secondary scenarios that also affect the experience
  • formulate hypotheses more precisely — they will be based on real user actions

You can document a scenario in different ways:

  • as plain text,
  • as a flowchart,
  • or as a chain of Figma screens you want to test.

The main thing is to understand and document the scenarios in advance. Make sure they are connected to the main research goal and the business problem, and that testing them will actually help solve it.

Hypotheses and research questions

A hypothesis is a reasoned assumption that needs to be tested. It is formulated as a statement about people, a product, or an interface.

A research question is needed when there is no hypothesis and you want to find a new solution or better understand behavior.

How to understand whether a hypothesis or question is formulated correctly

Check your version against these five rules:

  • Behavior is more precise than emotions
  • New knowledge
  • Formulated specifically
  • One hypothesis — one statement
  • The results will be used

If at least one rule is not met, it is better to reformulate the hypothesis or question.

Behavior is more precise than emotions. Ratings such as “liked it” do not provide reliable data — look at what people do and how they do it.

New knowledge. If the answer is already known from analytics or experience, testing it is pointless. A good hypothesis clarifies something the team does not yet understand.

Formulated specifically. You need to clarify exactly what you are testing.

One hypothesis — one statement. There should not be two ideas inside one hypothesis at the same time.

The results will be used. If nothing will change even after you learn the answer, the study is not worth launching.

If the hypothesis or research question passes all five checks, it is ready to use.

Scroll to top