In a previous post, I talked about how UX heuristics are getting less and less adapted to the modern digital landscape, relegated to being survivors from a time where the pinnacle of music was Ace of Base’s “All That She Wants”.
Today, I want to walk you through a new set of heuristics, crafted in 2012 by Lee and Kozar, which will help you build great experiences wherever content relevance and trust are important.
Lee and Kozar: heuristics for e-business, press or information-heavy websites
In 2012, Younghwa Lee and Kenneth Kozar, professors in Northern Iowa and Colorado, wanted to establish guidelines for e-business websites. They first reviewed 27 academic journals and conducted focus groups with usability experts, resulting in a set of 10 usability criteria. Then, to validate their findings, they asked 689 participants to fill a survey following a purchase experience on Amazon.com. With such a high number of participants, they were able to prove that the dimensions are well discriminated and consistent.
Lee and Kozar usability criteria are well differentiated, univocal and consistent across testers.
In technical terms, their criteria satisfy unidimensionality and convergent validity (factor loadings higher than 0.7 while significant at p=0.001) and that the dimensions were discriminant (Cronbach’s alpha higher than 0.83).
Screens show a consistent design throughout the system. Components and styles are similar everywhere: what behave the same, looks the same.
Navigating between screens is easy, and users are not disoriented. Features are easy to identify; the structure is clear and so is the hierarchy of available options.
Users can easily remember how to access pages and features, which are easy to understand (with respect to the complexity of the task).
The system provides interactions, multimedia features and micro-interactions that are appropriate and in sufficient number.
Users have a sense of their impact on the website or application and can sense the presence of other contributors and users.
The system displays a simple and minimal structure. Users can quickly grasp the components of the pages or screens. The components are not redundant.
The system provides a good balance between empty space and content. Components are well delineated and the screens are easy to scan. Graphic design facilitates legibility and understanding.
Information is up to date, covering a range of topics and presented in a level of details pertinent to the users and the context.
Distinguishing between content types is easy, especially when differentiating ads from content.
Help and support are clearly and easily accessible. Messages from the system are clear, reassuring and in appropriate number.
Users always know the system state.
Users feel safe and can trust the system’s handling of personal information. They know which information in required and why.
Content is perceived as true and trustworthy. Limitations of free or trial versions are known as well as possible future costs.
When should I use them?
With their focus on trust, telepresence and content relevance, Lee and Kozar criteria are perfect for e-business or informative web-site, such as magazines.
Even if Lee and Kozar built their set of criteria while considering purchases in an e-business context only, they are generic enough to be used in almost any situation. I found them particularly relevant when information quality and pertinence is central, would it be for a newspaper application, a bank or insurance website or even a university web ecosystem.
In my experience, they are more balanced than their older brothers, Nielsen’s and Bastien-Scapin’s criteria. The different dimensions are well-balanced and adequately separate the perceived experience between interaction, content, and trust issues.
The only situation where I found them wanting is when evaluating business applications aimed at performing highly complex tasks. Luckily, there is another set of heuristics that can deal with such context: Colombo and Pasch criteria, based on flow theory, which I will unpack in my next post.