How do we design digital interfaces? Broadly speaking, this involves an understanding of various factors related to psychology, social science, technology, and ethics that are put into practice using various design methods.
Understanding and defining user needs forms a crucial part of any design practice. It is figuratively sacrilegious to not consider user needs before designing the interface. As mentioned by Jennifer Tidwell in Designing Interfaces, "Good interface design doesn’t begin with pictures. It begins with an understanding of people: what they’re like, why they use a given piece of software, and how they might interact with it." By conventional wisdom , UX without user research is not UX.
Given the importance of user and business needs, we can say that, pragmatically, digital design is about translating user and business needs to an interface. In this essay, the aim is to establish an abstract structure between needs and interfaces using basic mathematical symbols and concepts that can be used to examine the synergy between needs and interfaces. We will answer the following questions with this structure:
- How do we evolve interface with needs?
- Why are enterprise applications complex? What can we do about it?
- What is redesign? When do we redesign?
- How do we account for innovation?
An important point to note here is that we are loosely using certain mathematical concepts like cardinality to convey the idea more abstractly. It is advised to not judge the essay on mathematical rigour.
The Structure of Needs and Interface
Let us consider an initial set of needs \( N1 \) consisting of various user and business needs with a cardinality \( |N1| \). (Cardinality is the number of elements in a set)
Let us also consider that our interface \( I1 \) comprises of various interface elements with a cardinality \( |I1| \).
By our definition, let us assume that we achieve the mapping of needs \( N1 \) to interface \( I1 \).
\[ N1 \longmapsto I1 \] (read as "needs map to interface")
The word mapping here does not imply a bijection between the sets. The symbolic representation of mapping conveys the idea that we have been able to achieve a translation from needs to an interface. This mapping is also inspired by the definition of morphism in category theory, where the association is defined from source to target. We can consider Needs as the source and Interface as the target.
A detailed explanation of this mapping would be covered in my upcoming book Needs and Interfaces.
We can begin with non-evolutionary cases in which there's no change in \( N1 \) and \( I1 \). In such cases, the value is in fulfillment of basic functional needs \( N1 \) that do not demand any changes in the interface \( I1 \) . Though such interfaces will look primitive with time, they will continue to serve the basic needs of the user. Some of the examples include Craigslist and Hacker News.
In most cases, there will be new needs as people start using the interface. These new needs will emerge from:
- using the interface itself. An example can be a usability need for restructuring of certain interface elements, need for better interface controls or providing adequate signifiers for an interaction affordance. Let us denote a set of all such incremental needs by \( n_i \). We will also analyse some needs later in the essay that require complete change in the interface. We denote all such acute needs by \( N_i \).
- from a change in business process, policies, infrastructure, socio-economic status, and technology. For example, there can be needs for compliance, inclusivity, and privacy not considered in the past. Various factors drive the quanta of such needs that vary with time. Some of these needs will be incremental while others will be substantial. We denote such incremental needs by \( n_j \) and substantial ones by \( N_j \).
Let us consider a shorter time-span \( t \) during which we have incremental needs \( n_i \) and \( n_j \) that do change \( N1 \) but not significantly. These incremental changes in needs can lead to incremental changes in the set of interface elements \( i_n \) leading to a change in \( |I1| \). Consequently, this will affect the interface \( I1 \). We can represent it as follows:
We need to consider some factors here:
- It is possible that a change in \( N1 \) might not lead to a change in interface I1. For example, there can be a need for a faster loading website. This might not have a direct impact on interface elements.
- There can be cases in which \( n_i \) can lead to a decrease in \( |I1| \). The user might give the feedback that certain interface elements can be reduced for better experience. Hence, a change in \( |I1| \) constitutes both its increase and decrease.
Over a more extended period \( T \) (spanning years), there will be significant changes in processes, policies, infrastructure, socio-economic status and interface trends. We can also have novel technologies that provide new interaction paradigms not available before. Such significant changes constitute the substantial needs \( N_j \) that we outlined above in the essay. This will drive an entirely new set of needs that we denote as \( N2 \) (It can include both old and new needs. However, we give it a new set representation for proper structuring.) Accordingly, the interface will also evolve to satisfy the new set of needs. Let us call it \( I2 \) . An example can be the evolution of Netflix, with substantial changes in business and technology \( N_j \) driving the changes in the interface. Netflix was a website-based movie rental service in 1997 that allowed people to rent DVDs online. Over an extended period, Netflix introduced a streaming model in 2010 that we all know today. We now see incremental changes in the interface \( i_n \) driven by various user and business needs ( \( n_i \) and \( n_j \)).
In this case, our updated representation will be:
(Note that there can be cases where the needs and interface do not evolve even for an extended period of time. In such cases, the structure will either continue to be in a state represented in Figure 1 or will follow the non-evolutionary state.)
In an ideal world, this process will continue ,and we will uphold this structure to maintain the relationship between needs and interface. As an example, we can see how the user interface has evolved over years.
However, as we don't live in an ideal world, let us consider various possibilities surrounding needs and interfaces.
The Enterprise Issue
Enterprise products are infamous for being complicated and unintuitive to use. One of the reasons for this complication is the temporal decoupling of needs and interface. Consider the situation when needs elevate to \( N2 \), but the interface does not change to \( I2 \). As the existing interface will continue to accommodate the new set of needs, it will lead to a substantial increase in \( |I1| \) ,thereby resulting in a bloated and complicated interface \( I1 \). In enterprise lingo, this is referred as the legacy UI. This can happen because of multiple reasons including business process complexity, short-term thinking, and huge technical debts.
At a surface level, the solution seems simple - evolve to \( I2 \) and maintain the symbiotic relationship between needs and interface. However, this is easier said than done. Enterprise applications come with an inherent complexity because \( |N1| \) is always large from the beginning. These needs are high because of multiple logical, situational factors and diverse sources of information leading to complicated task landscapes. For most cases, \( |N1| >> |N_c| \) where \( |N_c| \) can be considered the cardinality for consumer application needs. A large \( |N1| \) with different pattern of inquiries and multiple tasks can result in a large \( |I1| \) thereby causing inherent complexity. However, this does not mean that enterprise applications should be "too complex to use".
Focusing on core activities within task landscapes can mitigate this inherent complexity. A good understanding of needs and appropriate mapping to interface elements is crucial in making a usable enterprise application. One of the methods for appropriate mapping is to annotate interface elements with needs and then simplifying with deletion, organising and progressive disclosure.
The wrong turn
What happens when we do not have the correct mapping between needs and interface? It can happen because of three reasons:
- Misinterpretation of user research due to biases and preconceived notions leading to poor synthesis
- (Corollary) Poor implementation of the interface from an inadequate or incorrect understanding of needs. For example, adding low-value features can add to interface clutter.
- Misplaced priorities - When certain interface elements are excluded for reaching short-term goals despite a clear need from the users.
So, we have an interface that does not match the needs because of incorrect mapping.
\[ N1 \neg I1 \] (read as "needs do not map to interface")
It leads to an increase in need elements \( n_i \) as the interface does not conform to user and business needs. As a result, we no longer have just incremental interface needs, but an entirely new set \( N_i \) of unfulfilled, acute needs. Consequently, we will have a substantial increase in the cardinality of Needs that would make the current interface obsolete as incremental interface changes will no longer suffice. We can represent it as follows:
At this stage, the only viable option is to redesign the interface to \( I2 \) to match the needs \( N1 \)
We will define redesign as an activity to map the interface to unfulfilled needs. A re-design, by definition, should be done to meet needs rather than for vanity purposes. Note that the evolution of the interface as represented in Figure 2 should not be confused with a redesign. Evolution caters to a new set of needs, whereas redesign deals with the original set of needs.
Once in a while, we get a breakthrough interface and interaction that completely redefines our experience. However, this does not mean that there was no need for it. The need might be latent, but it's a need nonetheless.
That said, such an experience slingshots us into the future with a positive temporal decoupling between needs and interface. The latent becomes a new set of needs where competitors have to adopt this new interface and interaction paradigm to stay relevant in the market. It's represented as follows:
Companies would like to have this representation to create new markets and lead the way into the future.
Words alone can express the dynamics of needs and interfaces. However, this abstract structure gives an esoteric but a novel way to think about various scenarios by combinations of variables. Ultimately, it can be seen as an approach rather than a method to understand the interplay between needs and interfaces.
Notes1 IDEO gives a list of such design methods - https://www.designkit.org/methods
5 The word translating has been used colloquially and not formally.
6 From the book - Designing user interfaces - "An effective interface generates positive feelings of success, competence, and mastery in the user community. The users have a clear mental model of the interface that enables them to confidently predict what will happen in response to their actions."
7It's worth pondering about a bijection between needs and interface elements.
8 We can consider the evolution of browsers from an address bar and browsing window to tabbed interface, developer tools, plugins and many other elements to meet the new needs with evolving technologies and infrastructure - http://www.evolutionoftheweb.com/