Designing With Mental Model Diagrams — An Introduction
Better products and services start with understanding people
Understanding people and what they’re trying to achieve is a key part of a solid design process. Mental model diagrams are a powerful way to achieve this, as they help you map people’s experiences in-depth and in a scalable way.
In the first of a series of articles, we’ll give an introduction to mental model diagrams, talk about some of the theory behind them, and discuss what we believe are some of their key benefits.
Check out the follow up article on how to build mental model diagrams.
At SEEK we have a mature Design practice that engages in a lot of research. Recently, we’ve introduced Continuous Discovery which led us to run research more frequently and for more teams to be involved. Our research team has done great work to address these changes. They’ve helped educate the business about the different types of research (e.g., exploratory, generative and evaluative), refined processes, and championed standards to lift the quality across the board.
However, we still face some challenges:
- While having a lot of research insights is great, it also adds to the complexity. Even though we save insights in a centralised repository, it’s not always easy to see how they come together to tell a clear story. This is because it’s hard for teams to add their learnings on top of existing ones in a structured and straightforward way.
- We’re now looking after products shared across Australia, Asia and in the future other regions. This means that we may need to adapt solutions based on region, channel or both. This also translates in getting yet more insights, but this time from different markets. It’s crucial that we can bring together these insights and identify potential similarities or differences.
- In the past, teams have mapped experiences through customer journey maps. While useful, these methods of mapping tended to lack a high level of depth around needs and the problems people face. They also proved difficult to scale by adding new information on top of them.
In our attempt to address these challenges, they were a lot of approaches that we considered. However, we found mental model diagrams to be particularly useful to address the challenges at hand. So what exactly are mental model diagrams, why are they valuable, and how do they work?
What is a mental model
Many of the models and diagrams we show in this article are based on Indi Young’s work¹¹. However, the term “mental model” has a long history.
In 1943 the British psychologist Kenneth Craik put forward an idea: we make sense of the world and how things work by means of mental representations⁴. People constantly access and update these internal representations—i.e., mental models—in an ongoing effort to reach conclusions and predict outcomes.
Mental models then help us understand how people reason, feel, and the kind of attitudes and beliefs they bring to the table⁸. This in turn allows us to better understand what they’re trying to achieve — i.e., the intents behind their actions³.
How people construct mental models is based on several factors, such as past experiences, level of knowledge, and cultural references⁸. For example, your understanding of how a smartphone works might be different to someone who repairs smartphones for a living.
Why mental models are useful
In Human-Computer Interaction and Interaction Design, we often talk about the concept of mental models⁹. However, the discussions tend to be on how we can design a product or a service. For example, if we want to build a new ATM interface, we may consider how people think about ATMs, as this can impact their expectations and behaviours. It’s then all about the solution, as we look to optimise the end-product.
However, we can frame mental models outside any solution and use them to help us understand the problem space¹². The problem space is where we capture how people reason and express their beliefs and feelings; not in terms of how they understand a product or a service, but rather in terms of their intents and motivations. Using the example above, we would want to understand what people want to achieve when looking to access and use an ATM (e.g., access my bank account on the go), and what goes through their mind as they do it.
Mental models promote a deeper understanding of people. As we decouple mental models from solutions it further help us to break down complexity and avoid a too common sin; focus on products and services and fail to acknowledge that people use solutions as a means to an end, not as an end in itself.
The structure of a mental model diagram
Mental model diagrams are artefacts that are built around a framework that follows certain rules and processes. Commonly, we break down the structure of a mental model diagram into 2 parts:
- A problem space — Which is represented in the top section of a mental model diagram. Here, we capture people’s mental models around how they act, think and feel as they move towards their goals. We capture this information at different levels of granularity.
- A solution space — In the bottom part of the diagram, we identify how solutions and competitors align with the things people want to get done in the problem space. We can be quite flexible on how we present this information and use different ways of styling the information. For example, we use colour coding to rate how well we believe a solution is working to support a specific need.
Mental model diagrams offer us a unique way of structuring experiences. For example, customer journey maps revolve around solutions. Their focus is to “understand the myriad possibilities and paths a customer may take to complete his or her ‘job’”⁷.
On the other hand, mental model diagrams offer a broader and more in-depth understanding of people independent of an organisation. Focus and depth come mostly from having a detailed account of what happens in the problem space.
How we structure the problem space
In a mental model diagram, we build the problem space using qualitative research, using either semi-structured or unstructured interviews. We capture data as close as possible to the language people use, and we analyse it and synthesise it using a bottom-up approach. The result is a series of grounded insights that we organise using a 3-level hierarchy:
- The task/box level — As we hear what people have to say, we capture verbatims in order to identify concrete goals and emotions. These tasks or boxes are the basic building blocks of the problem space and are placed at the very bottom of the hierarchy.
- The tower level — As we identify how a series of tasks comes together through affinity mapping, we cluster them into towers. A tower is a series of tasks that explain what drives more concrete behaviours or emotional responses, but at a higher level of abstraction.
- The mental space level — At the highest level of the hierarchy, we have mental spaces. Again, by grouping together towers, we identify higher-level goals that are closer to the “why” that’s motivating people. Mental spaces are placed at the very top of the hierarchy and are its more abstract layer.
In Figure 2, we have an excerpt of a diagram that looks into how people go about improving their working lives. At the top level, we have a mental space that defines the broader intent — “Understand if this is the right job for me”. Within that mental space, we have more specific goals or towers, such as “Make sure the location is right”. Finally, within those towers we have tasks/boxes that represent more granular goals and reactions, such as “Understand how long it takes for me to get to work”.
But why are diagrams built like this and is it even a good idea? We believe there are 3 major reasons why using hierarchies is useful in structuring the problem space:
- Established theories support the notion that goals are hierarchical² and can be defined along a continuum, ranging from specific to aspirational and abstract⁴. One way to look at this is through the lens of why, what and who that defines motivated action⁶. Using techniques such as laddering we can see this in practice, as we ask people to explain why something concrete is useful to them (e.g., why is saving a job important to you?).
- Hierarchies make it easy to add new insights. They offer us a strict structure and higher-level goals tend to be more stable over time¹. The more we know about a domain, the easier it is to build on it. We move from a bottom-up approach to a top-down one, where adding new information is akin to seeing where a piece of evidence fits in the bigger puzzle. Additionally, humans easily connect with the concept of a hierarchy, as they’ve been part of how we’ve experienced the world for thousands of years⁸.
- With hierarchies, we can add new layers of analysis without changing the underlying structure. For example, we can analyse what the data is telling us at the task/box layer, identifying groups of people that share commonalities at this level (e.g., customer segments/personas). With a hierarchy, we can do this without changing anything above or below this level.
However, using hierarchies is not all roses. A major challenge is that there’s no clear pathway on how to build a hierarchy, nor a single right answer on how granular or abstract it should be. As a general rule you want to avoid being too broad or too narrow. Getting the balance right is key.
Building hierarchies does require some practice, but once you build a solid one, it will become a basis for future research. This in turn enables you to build ever more sophisticated models, without ever having to start over or ditch existing insights.
But how flexible are these diagrams in practice? For example at SEEK we run research sessions all the time. We’re accruing new knowledge on pretty much a weekly basis. Not all of this research is key to the problem space (e.g., usability testing), but eventually we’re capturing a lot of “goodies” that we can use.
Building more complex models
Mental models diagrams aren’t the only approach we can use to capture people’s experiences. Like we’ve mentioned, customer journey maps are a great way to show how people engage with an organisation⁷. However, customer journey maps tend to be static, as raw data is heavily synthesised to create the artefact. This makes it hard to include new insights without changing their structure.
From our experience, mental model diagrams do well where you want flexibility and scalability. We believe that they’re a powerful way for both building on existing insights, as well as helping you tell different stories.
In Figure 4 we have some examples on how we can tell these stories by layering different pieces of information on top of the model. For example, in the solution space we can identify whether a solution maps to something in the problem space, and how it works or not (e.g., colour coding strong/weak solutions).
In the problem space, we can identify different types of segments where people differ or converge. For example, we can learn that some people are more keen on seeing a higher number of jobs, while others have a more laser-focused attitude towards job finding and prefer to see fewer, more targeted, jobs. For the latter, providing an appropriate level of control could be crucial.
We can also add extra credibility to the insights in a mental model diagram by connecting them with other, more quantitative-oriented frameworks.
For example, Jobs-to-be-Done is an increasingly popular way to quantify needs. The premise behind Jobs-to-be-Done is that people “hire” a product or a service to help them get a “job” (i.e., a goal) done. Outcome-Driven Innovation, a specific approach to Jobs-to-be-Done, is well-suited for mapping with mental model diagrams. This is because “jobs” are implicitly organised as a hierarchy, which makes working across the 2 frameworks straightforward.
For example, at SEEK we’re in the process of layering insights from a Jobs-to-be-Done survey on top of a mental model diagram. After we do that, we’ll be able to quantify a whole range of “jobs”, how relevant they are, and what segments are emerging. Mental model diagrams enable us to visually represent these results using the structure already in place. This in turn helps us communicate effectively and powerfully within the organisation.
The opportunities that mental model diagrams offer us are exciting, and there’s much more we can do. At the moment we’re looking into how we can connect mental model diagrams with the Kano Method. The idea is to leverage the method as a way of qualifying solutions more objectively.
Hopefully, we’ve managed to shed some light on the value we believe mental model diagrams bring to organisations.
As we build mental model diagrams, we’re mapping out how people understand, reason, and feel throughout a domain of interest. Contrary to how mental models tend to be used, mental model diagrams focus mostly on the problem space. This means that we’re interested not in how people understand a product or a service, but rather what people are trying to get done and what their experiences are.
Mental models are built from the ground up using a hierarchical structure. This formality helps us to synthesise insights using a clear-cut methodology. It also helps us build on the knowledge we already have, without changing its underlying structure or meaning.
A big plus of mental model diagrams is how they make a great basis for building ever more complex models over time. By layering information, we can tell different stories about the insights and we can convey these visually and powerfully. Finally, we’ve hinted that mental model diagrams can further be aligned with and complemented by other frameworks, such as those based on Jobs-to-be-Done.
In our next article, we dig a bit deeper into how we’ve built our mental model diagram at SEEK. We provide some examples and share what worked well and what didn’t. This includes how we brought together people from different teams and backgrounds to collaborate in analysing data and building the model.
Thanks for reading. If you liked the article, check out our follow up article on how to build mental model diagrams within your organisation.
- Boekaerts, M., De Koning, E., and Vedder, P. (2006). Goal-directed behavior and contextual factors in the classroom: An innovative approach to the study of multiple goals. Educational Psychologist, 41(1):33–51.
- Carver, C. S. and Scheier, M. F. (1998). On the Self-Regulation of Behavior. Cambridge University Press, Cambridge, UK.
- Christensen, G. L. and Olson, J. C. (2002). Mapping consumers’ mental models with ZMET. Psychology & Marketing, 19(6):477–501.
- Craik, K. (1943). The nature of explanation.
- Gutman, J. (1997). Means–end chains as goal hierarchies. Psychology & marketing, 14(6):545–560.
- Hassenzahl, M. (2010). Experience Design: Technology for All the Right Reasons. Synthesis Lectures on Human-Centered Informatics. Morgan & Claypool Publishers.
- Lemon, K. N. and Verhoef, P. C. (2016). Understanding customer experience throughout the customer journey. Journal of marketing, 80(6):69–96.
- Payne, S. J. (2007). Mental models in human-computer interaction. In The Human-Computer Interaction Handbook, pages 89–102. CRC Press.
- Valdez, A. C., Ziefle, M., Alagoz, F., and Holzinger, A. (2010). Mental models of menu structures in diabetes assistants. In International Conference on Computers for Handicapped Persons, pages 584–591. Springer
- Wright, A. (2008). Glut: Mastering information through the ages. Cornell University Press.
- Young, I. (2008). Mental models: aligning design strategy with human behavior. Rosenfeld Media.
- Young, I. and Mangalam, K. (2017). Launching problem space research in the frenzy of software production. interactions, 25(1):66–69.