Building mental models (and how to leverage them for sound design)
Mental models are one of the most important concepts in product design but we rarely use them to their full potential.
The term “mental model” has become everyday product design parlance. Designers want to design for the user’s mental model. Researchers want to understand the user’s mental model. You hear it used by product managers hoping to build a product that “serves the mental model the majority of our customers have.” However, despite how much we speak about mental models, many UX practitioners fail to get as much value from them as they could because they focus too much on what users’ mental models are right now as opposed to how they change over time.
A very brief primer on mental models
If you take a look at the appearance of “mental models” in academic literature, there is a rapid uptick starting in the early ’80s that is still widely seen in the present day. But why all the mention of mental models? Simply put, because they are one of the primary drivers of human behavior.
A mental model is a mental representation of something in the world that humans use to predict what will happen when we interact with it. Psychologists and UXers have known this for decades. Kenneth Craik is usually credited with first conceiving of the idea in his 1943 book¹. Fast forward 40 years and Don Norman wrote an article that kickstarted widespread use of the term outside academia². He wrote about it again 5 years later in his seminal book The Design of Everyday Things³. Right about here is where mental models started to pick up steam in applied fields like design.
Norman famously pointed out three kinds of models that exist for a given design: the model as it exists in the mind of the designer (conceptual model), as it exists in the mind of the user (user’s mental model), and as it exists in the thing itself (system image). The typical design goal should be to have all three of these in alignment but, at a minimum, the user’s mental model and the system image should be consistent. If there is misalignment, UX issues will abound and the experience will suffer as a result. The takeaway: it is a long-established design best practice that the user’s existing mental model needs to be taken into account when designing something.
Figure 1: Norman’s visualization of three types of models
Users aren’t born with mental models — they build them
The definition I just sketched is the common one, and one that hasn’t changed all that much over time. In part, this constancy can be attributed to its being simple and useful — two great qualities in a theory. In its simplicity, it focuses on one type of mental model: the one that exists, statically, in the minds of users. But mental models are not eternal and they are not static. They are constantly being adjusted and created anew as users gain new experiences with novel products or changes in products they already use.
When users approach a product, they can choose one of three actions: 1) use an existing mental model as it stands or 2) make changes to an existing mental model or 3) create a brand-new mental model. How users decide which path to take depends on a number of factors that we don’t have time to explore here, but an opportunity arises when we realize that design practitioners have been focused almost exclusively on the first.
It is of course valuable for designers to take existing mental models into account. However, they can find even greater value in understanding how users are building their mental models and how those models change over time. Armed with this kind of knowledge, they can design products that not only account for existing models but empower users to easily adjust to changes, engage in new interactions, and confidently approach innovative designs that would otherwise be overwhelming.
3 stages to build a mental model
So, how do users build mental models of the products they use? In 2018 I published the findings of a study that took a deep dive into how mental models change with experience⁴. I won’t rehash all the details of that study here, but I do want to take a look at what we can learn from its conclusions. In particular, let’s look at the “Three-Stage Model” of how users build mental models of new products as they use them.
Let’s assume users are just starting to engage with a new product or a new feature. As they interact with it, their mental models will begin to form and change over time. This change over time can be thought of as flowing through three stages.
Figure 2: Three stages of building a mental model
Stage 1 — Getting familiar: At this starting point, users decide if they can use an existing mental model (maybe one used for another similar product) or if they need to start from scratch. The existing mental model may be void of useful knowledge or worse, contain the wrong knowledge. The core components missing at this stage include an understanding of the language the product uses and the actionable steps needed to complete core tasks.
Stage 2 — Becoming proficient: Users begin to understand the language of the product, especially if this language uses technical or unfamiliar vernacular. Understanding terminology and use of product language is one of the first building blocks that form the foundation of the model. This is also when users start to understand the step-by-step procedures required to perform basic operations to get some desired outcome. Here, they are usually learning the bread-and-butter operations of the product so they can receive enough value to motivate them to keep using it. Later, they will be able to build on this knowledge.
Stage 3 — Mastering: Users build on the core step-by-step tasks they already know, expanding their knowledge to more complex or niche capabilities. If they make it this far they most likely have an accurate understanding of most or all of the language the product uses.
Something else interesting happens here. Users already understand the language and core interactions the product offers them. However, assuming the product experience has some depth to it, users will start to understand more abstract ideas, strategies, and functions within the product. These are meta-level tasks or ideas that aren’t displayed outright in the UI, but regardless users discover them as they try to accomplish more of their goals.
For example, a medical professional may have learned to use the “Other” health problems field in a health record app as a “reminders” box for uncompleted tasks rather than, as the designer intended, a place to record information about unlisted health issues. She does this because she finds it simple and effective and understands the benefits of this approach over the designed to-do list feature. This kind of out-of-the-box thinking occurs as users construct a much more sophisticated mental model on top of the foundation they built from the earlier stages.
Lastly, at this stage users develop small “pockets of expertise” that serve specific situations or contexts. Instead of merely having broad knowledge that applies to the whole product, they develop specialized knowledge that applies to particular scenarios or contexts. Going back to the healthcare example, a physician has been using a single strategy for recording patient information during office visits. Now, however, she begins to understand how the product handles nuances between types of patients such as new vs. existing, complex vs. simple health conditions, and so forth. She already knew all the features available to her for filling out a patient chart during and after a visit. But now she has also developed unique understandings of sub-cases of different types of patient visits with respect to the charting software that allow her to tackle different kinds of context-driven tasks more effectively.
You can see from this three-stage model that mental models develop by first establishing a foundational understanding of a product and then using that initial model as a launching point to construct a more sophisticated one. The more evolved a model becomes, the more it can do what all mental models are designed to do: help users accomplish goals and satisfy those motivations that pushed them to engage with the product in the first place.
Applying what we’ve learned to product design
Now that we understand the basic process of how users build their mental models, let’s finish with a few ways of applying this model to improve product design.
1. Use the 3 stages as a guide for onboarding or training design
Onboarding new users to a product screams mental models. While users are being onboarded or trained they are creating a mental model that will be used to guide their future choices and determine success. It can be helpful to look at how well the onboarding experience or training program supports the three stages in order. If you find the product is encouraging or even requiring users to understand more complex tasks before learning essential product language, for example, things are probably out of order. Users may struggle to build an accurate mental model quickly or may avoid interacting with the product or certain features altogether. Remember, it is much more difficult for users to “unlearn” a poor mental model than it is for them to build an accurate one the first time around.
2. Use each stage as a set of heuristic criteria to evaluate a product
The different stages of development can also be used as a set of heuristics for evaluating a UI, interaction, or task flow. Using them is pretty straightforward and works just like other popular heuristic evaluation criteria that rely on a list of established design principles⁵⁶.
3. When a new product or feature is launched, measure mental models regularly
When releasing a new product or feature it will take time for users to progress through the stages of building a mental model. One way to gain insight into the state of user mental models is to run regular research studies, perhaps even with the same participants. This will give you a snapshot of what their mental models look like at various points in time. Taking a regular snapshot of user mental models, especially for new products or features, can help you recognize and diagnose issues quickly and identify ways to help users develop accurate mental models. If you are wondering how to measure mental models, there are many ways — too many to talk about here. However, if used carefully and appropriately, common UX research methods such as interviews and surveys are capable of giving an accurate glimpse into the mental models of users.
Mental models are a core component of our psychological decision-making process. This simple idea has long been incorporated into design best practices and changed the way we approach understanding our users. If we take the next step beyond merely looking at mental models that currently exist and consider how mental models of our products are actively evolving, we will uncover a deeper level of value that can drive products further towards exceptional experiences.
References
[1] Craik, K. J. W. (1952). The nature of explanation. CUP Archive.
[2] Norman, D.A. (1983). Some observations on mental models. Human-Computer Interaction, 241–244
[3] Norman, D.A. (1988). The design of everyday things. Basic Books.
[4] Furlough, C. S., & Gillan, D. J. (2018). Mental models: structural differences and the role of experience. Journal of Cognitive Engineering and Decision Making, 12(4), 269–287.
[5] Nielsen, J. (1995). How to conduct a heuristic evaluation. Nielsen Norman Group, 1, 1–8.6Gerhardt‐Powals, J. (1996). Cognitive engineering principles for enhancing human‐computer performance. International Journal of Human‐Computer Interaction, 8(2), 189–211.
[6] Gerhardt‐Powals, J. (1996). Cognitive engineering principles for enhancing human‐computer performance. International Journal of Human‐Computer Interaction, 8(2), 189–211.
