Designing with memory in mind
Introduction
While there are many human mechanisms used to determine our interactions with the outside world, one of the most important is memory. Without memory, we could not determine, process or recognise the elements of our surroundings. This information processor is of continuous study, and a vital consideration not just within the tangible world, but within any digital user experience. However, memory isn’t as straightforward as one may perceive, in fact its arguably complex, and impacted by our surroundings. This article will discuss human memory, as well as its importance and benefits within user experience design.
Memory is deceptive because it is colored by today’s events
Albert Einstein
Human Memory
So how does it work? At its most basic level, memory is the storage of data we collect from our surroundings. However data is not simply and indefinitely stored, in fact data can be processed in different ways and influenced by many factors. Studies have shown that there are 2 primary types of human memory; long term memory and short term memory. We can also consider more specific memory types including sensory and working memory, however our utility within UX systems can be simplified by focusing our understanding on LTM and STM¹.
A recent psychological study which appropriates existing study and evidence on human memory describes short term memory as data stored in our minds temporarily, which may be accessed consciously or unconsciously, and stored without repetition². However, there have also been significant studies to suggest that the human mind can only store an approximate amount of data, and only hold this data for an approximate amount of time. Whereas long term memory can be described as data that is stored for longer periods of time, possibly even indefinitely. However the full capacity of our brains has not quite been fully identified. To describe in detail, data is stored as a pattern of neurons. The stronger the neuron’s imprint, the stronger its existence in our memory. A key method to increase this strength is repetition. We can reactivate or recall a LTM by stimulating neurons in the same pattern as when the memory was formed. In comparison, most STM’s are forgotten³.
Despite several studies, there has been little confirmed deep scientific proof of the exact mechanisms involved in memory. However, neurobiologist Eric Kandel believes that the existence and methods of storing data in both short term and long term lies at a chemical level⁴. In 2012, Kandel conducted a study that revealed the existence of a competition between two agents to throw a ‘switch’ in the brain, a switch that allows a memory to convert from STM to LTM. To define, an agent named CREB-1 attempts the conversion, but CREB-2 attempts to prevent it, forming a balance. With enough repetition or definition, CREB-1 is able to bypass CREB-2 to create a LTM. Kandel also found that emotions were involved in the competition and balance between the agents, proving that emotions can affect memory at a neurological level.
Memory has been considered as a crucial component of the modern computer⁵. We can even loosely compare computer memory to human memory; STM — RAM, LTM — Storage. Our access to this memory was previously restricted to command lines and physical keyboard input, with a gradual progression towards digital GUI systems. These interfaces ease our cognitive load, reduce learning and allow less errors or mistakes to be made. With the development of VR interfaces more recently for example, we are lesser required to think about our steps and decisions. With this comparison, we can further understand the importance of memory consideration in UX design.
When we consider the flow of data from its initial perception to either a STM or LTM, we can strengthen our own UX workflows, consequently creating a product that promotes optimum ease of use. Our priority to design for STM or LTM can depend on the context of the interface. It would be ideal if users could store as a LTM, however, it could be argued that creating a LTM requires significant cognition⁶, maybe even resulting errors. Therefore, designing for a STM could promote an easier and quicker experience, particularly if data does not need to be effectively remembered. Given our understanding of how memory works, we can anticipate that if a STM is repeated, it will eventually become a LTM. When we create our designs, perhaps we should actively consider whether our aim is for users to store data in their minds as a STM or LTM, and consider the conversion of a STM into a LTM, therefore minimising cognition as much as possible.
Principles
There are several principles within interface design that allows us to ethically manipulate memory, and our utility of these principles can determine the possibility of the memory’s term. There are two important and proven principles to consider within the psychology of UX design.
Miller’s Law
In 1956, cognitive psychologist George Miller released a study that would become of the most cited papers in psychology; The Magical Number Seven, Plus or Minus Two: Some Limits on Our Capacity for Processing Information³. The study analysed our memory capacities and limits, particular in short-term memory, and created a conclusion that the average human can hold about seven objects or chunks of data, give or take two (7 ± 2) during a task that requires cognitive effort, hence the term ‘chunking’. Miller suggests that we consider data in forms of ‘chunks’, and limit the amount of data displayed to accommodate our STM storage capacity optimally, without overloading cognition, which could consequently result in a failure of data storage.
It is important to understand that this guideline amount of data is for memory recognition and processing, and not for recall. It has been reported that many designers misinterpreted Miller’s findings by strictly displaying 7 ± 2 chunks of data; for example, seven interactive items in a navigation menu⁷. Miller later responded to this misinterpretation to clarify his theory.
Miller also made several other discoveries, including the primacy and recency effect phenomenon, also named the serial position effect, which describes the importance of the order of these chunks of data⁸. He concluded that data chunks at the beginning or end of an experience have a stronger likelihood of being stored effectively, lesser for data presented mid-experience.
More recent studies have also concluded that the context or nature of the data can also affect the suggested numerical limit. Richard Shiffrin and Robert Nosofsky discovered that while numerical digits can be held on average in a quantity of seven, the number is reduced to six for letters, and five for words; suggesting that we are able to store more digits as a STM than letters, and even less when we combine letters into words⁹.
As discussed in Miller’s law theory, we process data in our minds in ‘chunks’. We can reflect this ability in design to promote processing of these chunks, effectively creating a set of data to imprint in the user’s mind. In design we can achieve this by breaking up content into digestible chunks to allow more effective processing. In psychology, a chunk is described as an ‘organisational unit in memory’¹⁰.
This memorisation technique can be applied in our designs by dividing information into clearly distinct groups of related content. For example, chunking text content into digestible paragraphs, shortening lines of text, grouping visually related items together or even simply grouping strings of digits or letters.
A specific example in online design is separating articles with short paragraphs and separating imagery. Remembering to group content visually, as per Gestalt Law of Proximity, to help identify related content¹¹. Another specific and common method is used to display telephone numbers. Since a long number can be difficult to remember, displaying the number in separated chunks allows for more efficient memory storage.
Miller’s law, and its consequent findings should always be considered within experience design, promoting a product that compliments user memory capacities and capabilities. We can achieve this by paying attention to the amount of data we display, the type of data and the positioning or order of that data.
Hick’s Law
We have defined that our short-term memory has limitations in processing data, particularly in regards to the amount of data. Psychologists William Edmund Hick and Ray Hyman found in 1952 that the time taken to make a decision increases with the number of choices offered, and the complexity of those choices¹². Therefore we must also factor in time taken in addition to the amount of data presented, as well as context.
It was determined that the more choices offered, the longer it takes for us to think and comprehend these choices, therefore making it difficult to make a decision. These decisions become more difficult as more choices are offered, increasing cognitive pressure, consequently resulting in poor decision making or failure to make a decision; a negative experience. As we consider each choice, option or chunk of data, the way we process these can be individual and unpredictable; our minds associate, become distracted and trail off, almost to a fan-effect. This can result in memory overload as we process significant information outside the choices offered, and possibly failing to store data for the appropriate amount of time, ie as a STM, or even LTM.
We can determine from Hick’s law that we should aim as UX designers to present only the most necessary information where possible without overwhelming, as this can interfere with efficient memory storage. We should remember that users will struggle or fail to process data and/or convert to an appropriate memory type if we provide too much data at one time. There will be occasions where we must present a high volume of data, and we can prevent cognitive overload by using design methods that still give us access to this data, but via intuitive design.
Methods
With our understanding of the science behind how human memory works, particularly in interface design, as well as some of the foundation theories gathered, we can determine some specific design principles and methods to include in our designs. Of course these are not strict guidelines and must be used with caution, with memory understanding in mind, and with clarified understanding of the purpose, as well as context.
A key consideration to bear in mind is that we should primarily, but not strictly aim to design for the short-term memory. If we want our users to store data long-term, we must remember that this is achieved by converting a STM to a LTM.
As well as the above considerations and methods, other ways to positively work with human memory in UX include:
Optimise Response Times
It has been proven that a slower experience results in poor memory performance. As digital interfaces developed and the performance speed increased, William Doherty noticed an improved experience. He put this theory to the test in 1979 within a study, of which he discovered that the longer it would take for the computer to respond, the longer the user took to think about the next step. Even milliseconds made a significant difference¹³.
Doherty believed that users were storing the series of steps they wanted to take in working memory, and the long response times interrupted their train of thought, preventing the storage of that data. This resulted in a decrease in productivity and possibly cause errors.
Aside from any technical ways to reduce response times, any designed experience should consider potential time taken for a goal to be achieved. This could be simple waiting times or steps between start to finish. Either may disrupt the user’s goals by causing a failure to create a clear memory.
Utilise Help and User Assistance Features
As discussed, memorising steps can be disrupted by a poorly designed experience, resulting in a lack of productivity or errors. We can not just prevent this by improving response times, but also by guiding the user effectively. For example, if a user gets stuck during their step by step process, a helping hand could get them back on track. This can be achieved through various forms of help or support¹⁴.
However, this should always be as in-line as possible. Help and user assistance should be offered where needed in context as users will not need to travel away from their current flow, which could potentially disturb their train of thought, disrupting memory storage, and again, poor productivity or errors.
Define Patterns
Its important to remember that users become familiar with consistent design patterns, both in-product and in other products (Jacob’s Law). Many of these position things like dialog buttons in predictable places, of which affects our spatial memory¹⁵. Our long term spatial memory assists and influences us across many different products given our historical experience and mental models, while short term spatial memory assists and influences us within our current session; a single product. So if we create a familiar, and also consistent design pattern within our product, we promote less strain on memory. Conversely, by rearranging these patterns, for example changing the position of a CTA from one page to the next, may disrupt our spatial memory strength and result in poor data storage. Of course there are cases where this isn’t always possible, plus additional items need to be added on top of existing patterns, which means defining patterns should not be strict.
In addition to design patterns, we can also provide visual cues to promote relativity which can then ease strain on our memory. This could be icons in addition or in replacement of text or buttons.
Providing familiarity and recognisable patterns or cues can enable correct memory processing, and avoid creating new, and possibly unfamiliar data which could be processed with error or inefficiency.
Considering Emotions
Human emotions can also impact our memory data storage, both positively and negatively. A positive emotional experience is likely to promote efficient memory storage and good working memory capacity, whilst a negative experience has a poorer chance of storage, strains our working memory, and of course creates a memory we would prefer to forget. So if we create an experience that makes us happy or content, our users are likely to process data in our product effectively, and provoke a comfortable recurrence. Conversely, if our users become unhappy or uncomfortable, their capacity to process data decreases, and are not so likely to return¹⁶.
As designers there are many ways we can create a positive emotional experience, including more than mentioned here. This could be via aesthetics, appropriate casual language or satisfaction and gratification.
Conclusion
There are many other concepts of improving the UX of interactive digital products to promote efficient memory processing, much of which can be dependent on resources and context. However with our understanding of the importance of memory consideration in design and its effects, we can confidently build experiences that positively influence one of our users’ core human mechanisms.
With this evidence, we can now understand the importance of memory in design. Considering the history of interfaces and the introduction and intuitive enhancements of GUI systems, designers have more freedom to produce customised experiences that align effectively with our users’ memory capacities and abilities, therefore producing better product experiences.
References
- Cowan N. (2008). What are the differences between long-term, short-term, and working memory? Prog. Brain Res. 169 323–338.
- Robertson, L.T. (2002). Memory and the Brain. Journal of Dental Education, 66(1), pp.30–42.
- Miller, G. (1994). The magical number seven, plus or minus two: Some limits on our capacity for processing information. Psychological Review, 101(2), pp.343–352.
- Kandel, E. (2012). The molecular biology of memory: cAMP, PKA, CRE, CREB-1, CREB-2, and CPEB. Molecular Brain, 5(1).
- Johnson, J. (2010). Designing with the mind in mind. Burlington, Mass.: Morgan Kaufmann Publishers.
- Kelley, P. and Whatson, T. (2013). Making long-term memories in minutes: a spaced learning pattern from memory research in education. Frontiers in Human Neuroscience, 7.
- Farrington, J. (2011). Seven plus or minus two. Performance Improvement Quarterly, 23(4), pp.113–116.
- Miller, N. and Campbell, D. (1959). Recency and primacy in persuasion as a function of the timing of speeches and measurements. The Journal of Abnormal and Social Psychology, 59(1), pp.1–9.
- Shiffrin, R. and Nosofsky, R. (1994). Seven plus or minus two: A commentary on capacity limitations. Psychological Review, 101(2), pp.357–361.
- Nielsen Norman Group. (2018). How Chunking Helps Content Processing. [online] Available at: https://www.nngroup.com/articles/chunking/ [Accessed 6 Nov. 2018].
- Nielsen Norman Group. (2018). Closeness of Actions and Objects in GUI Design. [online] Available at: https://www.nngroup.com/articles/closeness-of-actions-and-objects-gui/ [Accessed 6 Nov. 2018].
- Hick, W. (1952). On the Rate of Gain of Information. Quarterly Journal of Experimental Psychology, 4(1), pp.11–26.
- Doherty, W.J., Thadhani, A.J. (1982). The economic value of rapid response time. IBM Report.
- Nielsen Norman Group. (2018). Short-Term Memory and Web Usability. [online] Available at: https://www.nngroup.com/articles/short-term-memory-and-web-usability/ [Accessed 6 Nov. 2018].
- Tidwell, J. (2011). Designing interfaces. 2nd ed. Farnham: O’Reilly, p.17.
- UX Planet. (2018). Emotions, Working Memory, & Design — UX Planet. [online] Available at: https://uxplanet.org/emotions-working-memory-design-ae9ecfa3249b [Accessed 6 Nov. 2018].
Further Reading
https://headrush.typepad.com/creating_passionate_users/2005/12/_but_is_it_memo.html
https://www.nngroup.com/articles/website-response-times/
https://uxknowledgebase.com/ux-human-memory-a306a7dadcb8
http://uxbert.com/ux-psychology-principles-design-ux/#Memory_Limitations
