Shneiderman’s Usability Attributes
Ben Shneiderman (1998) identified five attributes of usability that are capable of being operationalised into testable usability requirements. Other attributes have been identified, e.g., efficiency, effectiveness, utility (Rogers et. al 2011) that are less straightforward to test, though aspect of them map on to Shneiderman’s.
1: Learnability (Time to learn)
The [mean, maximum, minimum, etc.] [lapsed time, time on-task, number of episodes, etc.] to achieve a specified level of performance [speed, error rate, etc.] for a specified task, by a specified user.
2: Speed of performance
The [mean, maximum, minimum, etc.] time to complete a specified task, by a specified user.
3: Accuracy (Rate of errors by users)
The proportion of erroneous transactions enacted, by specified users
4: Memorability (Retention over time)
The level of performance [speed, error rate] on a specified task, by a specified user, after a specified period since use.
5: Subjective satisfaction
The users’ subjective assessment of the overall acceptability of the software. It is usually measured by questionnaire, but behavioural measures are also used. May be specified as the proportion rating the system a particular level. Standardised questionnaires have been produced commercially (e.g. SUMI, QUIS).
Whilst it might appear desirable to specify the maximum rating for all of these attributes, they actually interact. In some cases they support each other but in others they oppose. The ten possible pair interactions are described below.
Learnability vs Satisfaction
It is what is usually meant by the naive requirement for the interface to be “user friendly” and is often the dominating usability attribute perceived by initial users. However as their experience of use grows its significance may wane in comparison with other attributes.
Satisfaction may support learnability
If you like a piece of software, does it make it easier to learn to operate? It may increase the motivation to learn. If the utility provided by a difficult to learn system is significantly high, once learned with its quirks and work-rounds mastered, its difficulty may actually be valued and its mastery worn as a badge of honour.
Learnability vs Memorability
These two attributes rely on related human capabilities; learning and memory. It may be they rely on different aspects of human memory; short-term (learning) and long-term (recall at the later date).
Learnability supports memorability
Memorability supports learnability
Less re-learning required
Learnability vs Accuracy
Learnability supports accuracy
If something is easy to learn less errors in performance are likely.
Accuracy opposes learnability
Accuracy is enhanced by redundancy, which increases the detail to be learned.
Learnability vs Speed
Interaction can be speeded up by short-cuts and special key manipulations, but these take longer to learn. However an easy to learn system will reduce the rate of navigation errors.
Speed vs Satisfaction
Completing a task rapidly is usually desirable.
However when speed is enforced by the system it may oppose satisfaction (time-out on data entry etc.).
Speed vs Memorability
System specific short-cuts and key manipulations may be forgotten
Memorability supports speed
Remembered detail does not have to be looked up or relearned
Speed vs Accuracy
Redundancy increases accuracy but reduces speed
Accuracy vs Satisfaction
Accuracy vs Memorability
Redundancy increases the detail to be remembered
Memorability may support accuracy
Well remembered details lessen errors
Memorability vs Satisfaction
Memorability supports satisfaction
Does dissatisfaction support memorability? There may be some evidence from the advertising industry: The Financial Brand ‘Do Irritating Ads Work Better Than Funny, Entertaining Ones?’ and RedC ‘The Art of Irritation in Advertising’.
Not a strategy likely to be intentionally employed in interaction design!
The Full Complex
For a particular piece of software it is necessary to find the ‘sweet-spot’ that optimally balances the usability attributes.
Bear in mind that ‘satisfaction’ will be affected by how well the other attributes are balanced against each-other together with attitudes towards other aspects of the design (e.g. graphic design, help and training, etc.)