The most widely used usability criteria sets (Jakob Nielsen’s 10 heuristicsand Bastien & Scapin’s 8 ergonomics criteria) just turned 23 years old. Now that they are of drinking age, it may be time to ask ourselves if they may need a face lift.
The Web has changed since the 90s
Jacob Nielsen have crafted his 10 Usability Heuristics for User Interface Design in 1994, the year of creation of Altavista, one year after the founding of Yahoo!, at a time where softwares and websites looked like this:
Designers have learned a lot in the last 20 years. Frameworks, usability, simplicity have infused our daily practice. With game changers like Google, Facebook, new hardware like the iPhone, new design frameworks like Material design, the majority of designers today know better than to fall into the same traps (or are protected by the frameworks they use).
20th century UX heuristics have served their purpose, their lessons have been learned, now is the time to move forward.
Traditional heuristics fall short when evaluating today’s interfaces
There are several difficulties with using the “good ol’ heuristics” to evaluate a 2016 interface or user experience.
Some categories are too broad, some too narrow
While performing an expert review, the majority “UX problems” fall down in 1 or 2 categories and some other dimensions are just missing. Bastien & Scapin’s dimension of “Guidance” can account to up to 70% of the usability problems encountered, from navigability issues to visibility of system status or visibility of possible actions. On the other hand, the “Signifiance” criterium exclusively deals with “understanding the labels and texts”. While this is important to usability, does it really deserve to be a whole criterium in itself?
Some categories are missing
Some dimensions are just missing from the heuristics. For instance, Nielsen has none related to readability. Try to categorise a lack of contrast between text and background for instance: it is not a problem of “Minimal design”, nor really of “Consistency and compliance to standard”…
Another example could be the navigation issues. Could we attribute a bad IA or a bad main navigation to the “Visibility of system status” dimension? Or is it a problem of “ User control and freedom”?
They are focused on old problematics
Nielsen criteria have three categories related to errors (“Error prevention”, “Help and documentation”, “Help user diagnose and recover from errors”). In the world of the 90s, with interfaces based on text mainframes, getting the software engineers to understand how to communicate errors to the users and help them recover was one of the biggest challenges.
Nowadays a lot of guidelines have been issued, and frameworks include default message formatting that mitigate that need. While it is still of importance to ensure error prevention and good communication, it doesn’t require a whole set of three dimensions anymore, only a single category “Error prevention and help” could suffise.
They are skewed toward performance and usability, not user experience as whole
This is in fact my main concern with the heuristics of old 1990s. They come from a world where usability was focused on “getting system to be simpler”. The main challenge was to get the job done in the easiest, fastest and/or most reliable way. Unfortunately this task oriented point of view doesn’t include dimensions such as Trust, Skill/Challenge balance, Fun, Motivation, Engagement… that are an integral part of what we think as User eXperience.
A modern definition of User experience could be the following.
User eXperience is a whole, hollistic view of the interactions that a person (called user) has with his or her surrounding environment, in order to attain a goal (conscious or unconscious).
As the system we are designing happens to be part of this environment the user interact with, we can contribute to a better experience for the person in his or her pursue of his or her goal.
In that sense, heuristics should also include “softer” dimensions, linked to the feelings that the system induces for the user. Having pragmatic dimensions for usability is good, but nowadays it is not enough to ensure a good experience.
But don’t throw the baby out with the bath water. Heuristics and expert review as a tool are not doomed. There are new sets of heuristics out there, crafted in the 21th century, that aim to encompass those dimensions. We will focus on two of them (at least) in our next post.