1. Problem specification: Clarify and define the stakeholder’s main problem through interviews
2. Goal definition: Write clear, concise, foolproof goals directly tied to the problem
3. Hypothesis formulation: Create testable YES/NO statements based on the problem, preliminary research, and industry practice
4. Concept and metric definition: Align on terms and measurable criteria to avoid misinterpretation
5. Research questions: List internal guiding questions to confirm or refute hypotheses
6. Participant questions: Translate research questions into user-facing interview or survey questions

The scientific method is an established, verified, worldwide practiced system of steps by which something is achieved or known. It is a repeatable way of solving a problem, while it also includes the process of planning and carrying out research. Using the scientific method means uncovering the context of the surrounding world most systematically.

To include as many eventualities as possible, we are starting from scratch. We have nothing but a vague brief from a stakeholder. Therefore in real life, you will choose only what you need from the steps listed - depending on the scope and content of the research project.

Our model situation (The stakeholder has a problem) :

• We found that when refinancing a mortgage after 15 years, the largest number of clients leave for another bank. We want to know why and if we can somehow solve it through the app.

‍

‍Step 1: Problem specification

Or formulation of the main research problem. At this point, you make sure you’ve captured the problem and the stakeholder’s expectations. This step is based on communication between the stakeholder and the researcher (or designer). This is done through stakeholder interview. This process ensures that the possibility of forgetting any essential or desirable aspect is minimized.

‍

Model situation: The results of stakeholder interviews revealed the following 6 points:

1. I want to know the reasons why clients leave when refinancing their mortgage to another bank
2. I want to know why they took out the mortgage.
3. I want to know why 15 years is such a significant point.
4. I want to know how we solve this problem through the application.
5. I want to know if they chose another bank because of their social networks.
6. I want to know if they have marketing campaign notifications turned on

Highlighted sentences represent those data that are the most related to the given problem. We can use all other data later, e.g. for the formulation of hypotheses or questions for the questionnaire.

Model situation: Our main research problem now consists of three points:

1. What are the reasons why clients leave when refinancing their mortgage to another bank?
2. Why is leaving tied to the fifteenth year of the mortgage?
3. How can we use the application to prevent the client from going to another bank in this situation?

‍

Step 2: Goal definition

A well-formulated goal is clear, concise and foolproof. The best-formulated goal is one that even a disinterested person on your team can understand. Goal definition contains keywords that give a clear description of the framework and nature of the research. The goal does not have a journalistic form and does not contain redundant words.

Model situation:
Problem 1: What are the reasons why clients leave when refinancing their mortgage to another bank?

• Goal: To map the specific reasons for the departure of clients when refinancing a mortgage to another bank. - CX

Problem 3: How can we use the application so that the client does not go to another bank in this situation? - UX

• Goal: Mapping solutions to the problem on the market. (Competitor analysis)
• Goal: Mapping the motivations for the client to continue with the mortgage in the original bank. (Analysis of secondary data + interviews)
• Goal: Obtaining the client’s opinions on the proposed solutions (Testing with users)

I included methods just for demonstration of next steps. But as you can seen , we already the broader issues of the problem becomes clear to us. As the granularity increases, so does the complexity of the project.

‍

Step 3: Formulation of hypotheses

All steps are connected. The creation of hypotheses often takes place simultaneously either with the definition of concepts or we discover some already during the stakeholder interview.

A hypothesis is an assumption -It is a statement about the relationship between two or more researched themes. With its formulation, hypothesis predicts a certain state or relationship between themes, which can be detected and investigated - empirically verified. If we cannot formulate such a hypothesis, we cannot even test it.

‍

Model situation:

Research problem: What are the reasons for clients leaving when refinancing their mortgage to another bank?
Goal: To map the specific reasons for the departure of clients when refinancing a mortgage to another bank.

Hypotheses:

• The client was motivated to leave by a campaign on social networks.
• The terms of refinancing were more favorable with the competition.
• The client does not feel loyalty towards the bank.
• The bank did not provide the client with conditions for which he would decide to stay.

Hypotheses are statements. Clear, precise and eloquent. They must have a form of YES/NO answer. Do not use “I think that … we assume that … I expect that …” . Use just the words that must be there and remove all unnecessary.

Hypotheses help to minimize the subjectivity of the researcher because their truth or falsity is confirmed independently of his wishes. We formulate hypotheses based on:

• Stakeholder interview
• Preliminary research
• Best practices / industry practices

‍

Hypotheses form a connection between the theoretical and empirical components of knowledge. They are derived from the theory and their confirmation or disproof enriches and develops the theory. Both variants — verification or falsification — are equally valuable from a scientific point of view.

Industry practice and experience tell us that stakeholders are more satisfied when hypotheses are confirmed. It’s a strange psychological trick. It doesn’t seem so distracting and doesn’t evoke the impression that we’ve failed. All those who are afraid that their hypothetical statements will not be confirmed, remember that this procedure is completely correct and normal from a scientific point of view.

Our knowledge develops not only by finding out that something is like that, but it also develops equally significantly by revealing that something is not like that.

For some types of projects, it is necessary to define individual terms to avoid misunderstandings. This is not about doubting intellectual abilities. This is a problem of interpretation. Each of us interprets the reality based on our own experiences and subjective interpretation of the world. And yes, while a mortgage is still a mortgage, the emotions and experiences associated with a mortgage are different for everyone and this also affects the further investigation of the mortgage from the perspective of the customer journey.

‍Model situation:

Hypothesis: A campaign on social networks motivated the client to leave.

• Concepts: Who is the client? What social networks? What form of contributions?

Hypothesis: The terms of refinancing were more favorable with the competition.

• Concepts: What are the conditions?

As for the metrics, we are talking about the fact that we must define everything that we have to “measure” in the hypothesis for it to be confirmed or denied.

‍Model situation:

Hypothesis: A campaign on social networks motivated the client to leave.

• Metric: degree of motivation

Hypothesis: The terms of refinancing were more favorable with the competition.

• Metric: the list of more favorable conditions
• Metric: the degree of affinity towards these conditions

The degree of anything is best quantified on a scale of 1–10, where the participant evaluates himself based on his personal inner experience and understanding.

‍

Step 5: Formulation of research questions

This step means listing all the questions we need to answer to confirm or not confirm our hypotheses. Thus, we will be able to answer the goal of the research. These questions are not for the client. They are for us. They help us not to get lost and not to forget anything.

Model situation:

Research problem: What are the reasons for clients leaving when
refinancing their mortgage to another bank?
Objective: To map the specific reasons for the departure of clients when refinancing a mortgage to another bank.
Hypothesis: The client was motivated to leave by a campaign on social networks.

Questions:

• Does the client use social networks?
• Does it monitor the websites of competing banks?
• Is he influenced by the content he sees?

‍

Step 6: Formulation of questions for the user (interview/survey/…)

The last step is the formulation of questions for the user, whether we are talking about an interview or a questionnaire. These are all the questions that we have to ask the user to be able to answer all the research questions, thereby confirming or not confirming the formulated hypotheses and thus being able to answer the research goal. And ultimately, to know the answer to the stakeholder’s problem.

‍

Model situation:
Research problem: What are the reasons for clients leaving when refinancing their mortgage to another bank?
Objective: To map the specific reasons for the departure of clients when refinancing a mortgage to another bank.
Hypothesis: The client was motivated to leave by a campaign on social networks.

Research Questions:

• Does the client use social networks?
• Does it monitor the websites of competing banks?
• Is he influenced by the content he sees?

Questions for User:

• Do you use social networks? (if so, what kind?)
• Does he follow any banks on social media? (if so, what kind?)
• Do you feel that the given content influences you in your financial decisions? If so, to what extent?

‍

All research has an inductive and a deductive dimension - induction means going into the field, observing, and gathering information about what people do. Deduction means explaining a fact based on a general hypothesis or theory. So the well known corporate double diamond is basically just method of induction and deduction.

User experience is not a natural science. It is a sociological study. We ask about a user and his specific experience. Yes, in usability testing we ask about parameters that we can classify under the large IT box. But that’s only one part. Customer journey, personas, feelings about colors, preferences towards illustrations, the overall feeling that the already mentioned mortgage refinancing evokes in us. All this is the level of investigation of the social sciences, whether we are talking about sociology, psychology, or ethnography.

Our rigorous stick up to scientific, or in our case, the science-ish method has its justification in the fact that every social experience is:

• Complex - the result is action of many factors, which makes it difficult for us to isolate it and simply describe or examine it
• Collective - results are product of the behavior of a large number of individuals
• Historical and unrepeatable - all social experiences and processes are connected with previous experiences and they can not be repeated in exact same way (as an opposite to natural science experiments)
• Open - there is no final form of a social experience

We cannot claim that we have described a problem perfectly, permanently, and definitively. At most, we can claim that we have contributed to the description of the given experience. We move along the spiral, getting closer and closer to the essence of the problem, but we will never be able to describe all the eventualities that mortgage refinancing brings.

Thanks to following the steps of the scientific (and science-ish also) method, we can claim that in the given context, based on the selected approaches and defined metrics, we found these results from this specific group of users at the given time.

… and it’s hard to argue with those statements.

‍

You can read the whole article at Medium.