Objective assessment of user interfaces

Meet Cømpass the AI that measures visual complexity

Cognitive resources are limited in users. They should be used to solve tasks, not to operate interfaces.

Cømpass uses computer vision to measure dozens of factors that contribute complexity to an interface, thus depleting cognitive resources and hindering user performance.

Analyze GUI components
Cømpass identifies individual UI components. It can then analyse them to understand what they mean and how difficult they are to process by humans.
Evaluate visual hierarchy in the GUI
Cømpass determines relationships between individual components to identify groups of components. It then evaluates the visual hierarchy in the layout.
Measure total visual complexity
The tool uses dozens of factors to evaluate the extent to which the interface is visually complex and thus difficult to process and understand by people.
Estimate extraneous cognitive load
Cognitive load is burden of information processing that people experience. Extraneous load is due to how the interface has been designed, and not due to the task itself.
Predict user performance
All these factors converge to hinder users from becoming proficient with a tool, prevent them from being productive, and induce them to make errors.
Develop designs that work for users
By using Cømpass to guide decision making in design, product owners can steer the development of user interfaces towards better outcomes for users.

Computer vision detects what people can't

Computer vision can analyse details and relationships that far exceed the boundary of human perception. Research in cognitive science has identified dozens of factors that predict cognitive load and correlate with performance in the use of digital interfaces.

f	Measure	Description	Cues	GUI Type	Dimension
f1	Number of elements	Total number of all UI components	124	Web, Mobile, Embedded	Visual complexity
f2	Alignment	Arrangement of components on the UI	17	Web, Mobile, Embedded	Structural
f3	Density	Ratio of total area covered by the UI components to the entire screen	86	Web, Mobile, Embedded	Visual complexity
f4	Element connectedness	Closeness of UI components in layout	54	Web, Mobile, Embedded	Perception
f5	Common region	Clearness of the gap between groups of components	23	Web, Mobile, Embedded	Perception
f6	Grouping	Number of objects that are distinguishable	243	Web, Mobile, Embedded	Visual complexity
f7	Sequence	Organization of components for easy perception	15	Embedded, Mobile	Structural
f8	Cohesion	Equality level of UI components’ aspect ratio with the display’s AR	41	Embedded	Visual complexity
f9	Size	The size of UI component	8	Web, Mobile, Embedded	Perception
f10	Balance	Layout smoothness of UI components	87	Web, Mobile, Embedded	Perception
f11	Equilibrium	Degree of overlap of the center of all components with the display’s center	25	Embedded	Structural
f12	Symmetry	Degree of uniform placement of UI components on an axis	26	Embedded	Structural
f13	Colour variety	The amount of displayed colours	124	Web, Mobile, Embedded	Visual complexity
f14	Clusters of colour	The number of clusters of RGB values	17	Web, Mobile, Embedded	Visual complexity
f15	Entropy	Shannon’s information entropy of UI image	86	Web, Mobile, Embedded	Visual complexity
f16	Luminance	Standard deviation of pixel luminance values	54	Web, Mobile, Embedded	Perception
f17	Figure-ground contrast	Separation of UI components from the background	23	Web, Mobile, Embedded	Perception
f18	White space	Ratio of the area without any UI element to the entire UI area	243	Web, Mobile, Embedded	Perception

Cømpass is the objective foundation for better digital products.

Without Cømpass, even the best design methods fall short of a paradigm where design decisions can be compared on a common basis.Cømpass is being developed to make user performance predictions more reliable.This is made possible by the objective measurement of factors that are proven to affect how humans process visual user interfaces.

*The impact of interface design on user performance in recurring digital tasks: a comparative study. Manuscript in preparation.

Cømpass is currently under development, with a focus on GUIs used by professionals in high stakes industries.

Common questions

How is Cømpass different from other tools in the field of design?
Other tools simulate or measure where a user is looking on the screen. While the data is informative, this insight is limited in how it can be applied to make design decisions. Cømpass tells designers if a particular design is visually complex and thus difficult to process and understand for users. Thus Cømpass makes a direct connection between a design and the expected performance of users.
Is insight from an AI tool better than direct feedback from users?
User feedback is useful and remains valuable when used along side Cømpass. However, the AI can offer insight that is difficult to obtain from user research studies alone. It is more precise than people reflecting on their experience, acting as an objective benchmark that allows comparisons between designs. It can detect issues that people are not conscious of. And of course, it is fast.
How can Cømpass be used in design practice?
Cømpass is a special measurement device because it correlates objective data with user performance. That's exactly what designers have been missing. With Cømpass, designers can test and compare design alternatives to guide the evolution of a GUI towards better user outcomes.
When will Cømpass be commercially available?
Currently under development. The tool is a joint effort by experts in computer vision and cognitive science. To be the first to hear about the latest updates, sing up for the Beta list.