Also published on Medium : https://medium.com/@paulamatdesign/psycho-socio-methodology-canvas-for-ux-research-d622c8e0cd2d
One of the observations I have made in my few business experiences is the difficulty of setting up a UX Research operation upstream or during a project.
This difficulty is due to two things for me: the lack of knowledge of the ins and outs of UX Research on the part of companies, stakeholders, but also the opacity of the methods of structuring UX Research operations for some UX Designers. I am one of those UX Designers who do not have a background in social science but who wish to evolve in knowledge and operational capacity in the field of UX Research. I’m not just talking about writing an interview protocol or a questionnaire in good practices, I’m talking about applying methods in psychology and sociology in the most literal and complete way possible. However, this development can be complicated in the absence of a structured methodology on which to rely constantly, as psycho/socio methods can be difficult to theorize, and above all, as the application of UX Research can be difficult to “scientize” and systematize (assuming a UX Research application model can be systematic).
Democratizing access to psycho/socio methods
My project is therefore to propose, at my small level, a structuring of the psycho/socio method applied to user-centric design with a “ready-to-use” canvas, supporting user research from the first hypotheses to the formalization of a Persona supported by data. Indeed, the canvas presented here is not intended to be an end in itself.
Goals
From a project point of view, this canvas should allow three essential things: to make the right user assumptions and facilitate their verifications, to formulate the right design problems, and finally to install KPIs product/service consistent with the first two points.
From a user search point of view, this canvas should allow us to choose the right variables to study the target user, to make them interact effectively, and above all to structure these interactions between variables. This canvas offers a very literal, even down-to-earth, approach from applying methods in psychology and sociology to user-centric design.
In short, this canvas should be used to formulate a user-product equation that will prove to be fair if the assumptions it contains are not rejected by the user data.
Expected deliverables
The potential deliverables following the realization of this canvas are multiple:
- User hypotheses retained or rejected;
- Sampling criteria for a research survey;
- Complete persona with sociological profile, behaviors and opinions, motivations and objectives…
- Product/service design issues;
- KPIs produit/service ;
- Synthetic methodology presented to the sponsor/stakeholder to communicate about the process.
Advancement
Before getting to the heart of the matter, it should be pointed out that this canvas is a methodological framing theorization of UX Research. It was on the stroke of a failure, a workshop for framing target users having turned to disaster because too direct, that I improvised a first version of this canvas. I am in discussions with professionals to enrich it and I hope to soon be able to test it in a concrete case, within a company, but I have not yet had the opportunity to do so in the whole methodology.
The Canvas
In a nutshell
The use of this canvas is done in several stages:
- Describe the end user by listing observed or assumed discriminating characteristics (in the form of independent variables invoked: IVi);
- Describe the product/service or product/service concept by planned or actual characteristics (in the form of independent variables caused: IVp);
- Predict observed or expected occurrences of measurable user behaviours and/or opinions under real or experimental conditions (in the form of dependent variables: DV);
- Explain each behaviour or opinion by underlying motivations, objectives, social mechanisms… hypothetical or reported by the users themselves (in the form of causal inferences: CI);
- To extract from each behavior or opinion a product/service design problem (P);
- Link these different variables to formalize assumptions about the user;
- Prepare a quantitative survey from IVi/IVp and DV;
- Reveal a trend by crossing IVi/IVp and DV with a statistical analysis of the data collected;
- Prepare a qualitative survey from IVi/IVp and DV;
- Reject or retain hypothetical explanations given to these correlations (CI) if they exist with the justifications given by the users themselves;
- From the retention or rejection of hypothetical correlations between IVi/IVp and DV, and from the retention or rejection of hypothetical explanations given to these correlations, reject or retain the assumptions about the user.
Example of use
The illustrations presented throughout this article are based on the example of my graduation project: the support of a startup offering its users the possibility to repatriate and manage its personal data from different networks. During this project we have as a team formalized a Persona from a quantitative and qualitative study, but we would have had great interest to rely on a canvas to frame our methodology.
Describe
Who, Where, When, With what ?
Describe the end user by listing observed or assumed discriminating characteristics (in the form of independent variables invoked: IVi)
Describe the end user means first and foremost listing its distinctive characteristics. These are sociological first: we can rely on the heel regularly updated by the INSEE to define an age range or income level if these variables are deemed relevant. As you will have understood, I am one of those UX who think that in the realization of a Persona, the sociological profile has its importance. These variables may not allow us to justify behaviour or opinion by social determinants, but at a minimum they can help to target a panel with concrete qualifiers. It should be noted that while these variables will sometimes be measurable, and therefore potentially perceived as quantitative variables, we will prefer to consider them as qualitative variables in this exercise, simply because we do not need to measure their terms and conditions. Thus, rather than noting the age in years of the user (real or hypothetical), we will prefer to categorize the user in a given age range. The goal is not to give a unit of measurement to a variable.
Beyond the sociological profile, this part of the canvas can collect any information inherent to the subject, to the person being studied. Any categorization has its place as a variable modality: for example, we can imagine a “hesitant” user profile and another “determined” as two modalities of the same “buyer profile” variable.
| Type | Format |
| Independent variable invoked | A priori qualitative |
Describe the product/service or product/service concept by planned or actual characteristics (in the form of independent variables caused: IVp)
The same exercise can be carried out by taking the product as a subject or rather, because we are here in advance of the design, the concept of product or service. Using independent variables here allows us to begin to establish the possible framework for future experimental research. The variables will be the first parameters on which to play to study the subject or subjects.
| Type | Format |
| Independent variable caused | A priori qualitative |
Predict
Do/Think what ?
Predict observed or expected occurrences of measurable user behaviours and/or opinions in real or experimental conditions (in the form of dependent variables: DV)
Here we will note the variables used to measure the occurrence or non-occurrence of the user’s behaviours or opinions about or via the product/service. It is necessary to insist on the measurable nature of this occurrence or non-occurrence: in fact, the objective is not only to check whether a behavior or an opinion appears, but to decide the keys that will measure its appearance. This will be an opportunity to specify variables to use during a field survey and, later, during a user test. The measurable nature of these variables also reveals the first IPPs that are essential for the project, which will measure the effectiveness of the project once it is put into production. Unlike the variables used previously, they must include a unit of measurement: time spent in seconds on a page, number of clicks to reach information, number of errors made before reaching a goal…
To collect the data necessary for our analysis, it will be necessary to implement a study under experimental conditions, in “laboratory”.
| Type | Format |
| Dependent variable | A priori quantitative |
Explain
Why ? how ?
Explain each behaviour or opinion by underlying motivations, objectives, social mechanisms… or reported by the users themselves (in the form of causal inferences: CI)
So far we have only pointed out or hypothesized a causal relationship between a descriptive trait of the user and/or the product (independent variable) and the occurrence of a behavior or opinion (dependent variable). In this section, this causal relationship should be explained, i.e. its functioning. The meaning of the relationship is already known: the modalities of the independent variables are at the origin of the occurrence studied. To explain this relationship, we observe or hypothesize the motivations, personality traits, goals, environmental influences or underlying social mechanisms to which users are subjected. The study of social mechanisms is the reason why, in our definition of Persona, we do not exclude the sociological characteristics of the subject.
Extract ingupeer of a product/service design problem (P) from each behavior or opinion
Characterizing the product or product concept allows, among other things, to extract one or more design problems. These take the form of a question to be solved:
“How can the subject [IVi] be allowed to do [DV] with [IVp] to achieve its [CI] goal?”
Assumptions and problems
Link these different variables to formalize user assumptions
To extract a hypothesis from the canvas, once it is filled and the variables linked by hypothetical relationships, we can simply read the latter from the first to the third column as a single sentence, following the thread of these relationships:
“The subject [IVi] with the product [IVp] will do [DV] because [CI].”
Before continuing
We hypothesized that a characteristic of the user and the product (IVi and IVp) are the cause of product behavior (DV) because the user is motivated by something (CI). Now is the time to test this hypothesis with a field investigation. The hypothesis will be retained if these two conditions are true:
- There is indeed a link between the user characteristic and the occurrence of a behavior. In this study model, this link takes the form of a statistical correlation.
- The explanation given by the users themselves to this link corresponds to that hypothesis.
The first condition will be verified by a quantitative survey. The second condition, and this is the beauty of UX Research, will be verified by a qualitative survey. If the second condition is not verified, the initial hypothesis will have to be rejected in favour of a redefinition including user feedback.
Quantitative study
Prepare a quantitative survey from IVi/IVp and DV
Once the hypotheses have been made, we can begin drafting study protocols to validate or invalidate them. The advantage given by the canvas is that it has already fixed the variables to be studied with their modalities. As a reminder, we have independent variables relating to the user profile; these can be used to filter a sample of people. In a quanti survey, independent variables are used in the same way as the sociological heel, so on the face of it, in a questionnaire: a variable, a question (unless variable modalities represent a score or index).
The other typology of variables are the dependents. These should each be the subject of several questions. Questions of fact, corresponding to assumptions about practices, must be doubled or counterbalanced in order for responses to be accepted as actionable data. Questions of opinion, which correspond to assumptions about opinions, must multiply to explore the broadest possible spectrum of a notion, thought or imagination.
Reveal a trend by crossing IVi/IVp and DV with a statistical analysis of the data collected
Once a block of data has been collected by a questionnaire, it is time to manipulate the data to draw conclusions. The objective is to cross-reference data between independent variables (which materialize by the answers given to sociological questions) and dependent variables (which materialize by the answers given to questions of fact or opinion). If the variables, dependent and/or independent, are quantitative (so measurable), it will also be possible to compare the lists of answers between them to reveal possible correlations.
Qualitative survey
Prepare a qualitative survey from IVi/IVp and DV
In a quali survey, independent variables, in addition to assisting segmentation, may sometimes help define the scale of study: the more discriminating they are, the smaller the scale, sometimes from macro to micro (the choice of scale must be however, be done according to the starting brief and cannot be systematic).
Dependent variables will be able to provide a basis for designing a maintenance guide. The aim will be to explore the question of the hypothetical practices and opinions of users in a less direct setting than the questionnaire to get to the heart of the motivations, technical or social constraints, the determinants that are at the origin of these practices and opinions.
Reject or retain hypothetical explanations given to these correlations (CI) with the justifications given by the users themselves
With quantitative data with cross-and-correlational analysis, as well as qualitative data from interviews, we are now ready to objectively retain or reject the assumptions we have extracted from our canvas. For this, we note the veracity of our two starting conditions:
- There is indeed a link between the user characteristic and the occurrence of a behavior.
- The explanation given by the users themselves to this link corresponds to that hypothesis.
Closing the study
From the retention or rejection of hypothetical correlations between IVi/IVp and DV, and from the retention or rejection of hypothetical explanations given to these correlations, reject or retain the assumptions about the user
