Discount usability engineering methods

-- Jakob Nielsen

Involves a small team of evaluators to evaluate an interface based on recognized usability principles

Heuristics–”rules of thumb”

First introduced in 1990 by Nielsen & Molich

Quick, inexpensive, popular technique

~5 experts find 70-80% of problems n Based on 10 heuristics

Does not require working interface

What does this interface tell you?

• Use concepts, language and real-world conventions that are familiar to the user.
• Developers will need to understand the task from the point of view of users.
• Cultural issues relevant for the design of systems that are expected to be used globally.

A good match minimizes the extra knowledge required to use the system, simplying all task action mappings (re-expression of users’ intuitions into system concepts)

Would an icon with a red flag for new mail be appropriate in all cultures?

Users choose actions by mistake

Same words, situations, actions, should mean the same thing in similar situations; same things look the same, be located in the same place.

Different things should be different

• Both H2 (Match between system and the real world) and H4 related to user’s prior knowledge. The difference is
  • H2 is knowledge of world
  • H4 of knowledge others parts of application and other applications on the same platform.
• Consistency within an application and within a platform. Developers need to know platform conventions.
• Consistency with old interface

Consistency maximizes the user knowledge required to use the systems by letting users generalize from existing experience of the system to other systems

Evidence: Should include at least

• two inconsistent elements in the same interface, or
• an element that is inconsisten with a platform guideline

Explanation: What inconsistent element is and what it is inconsistent with

• Help users select among legal actions (e.g., greying out inappropirate buttons) rather than letting them select and then telling them that they have made an error (gotcha!).
• Subset of H1 (Visibility of system status) but so important it gets a separate heuristic.

Motivation: Errors are a main source of frustration, inefficiency and ineffectiveness during system usage (Lavery et al)

Explanation in terms of tasks and system details such as adjacency of function keys and menu options, discriminability of icons and labels.

• Classic examples:
  • command line interfaces (rm *)
  • Arrows on keys that people can’t map to functions
• Much easier for people to remember what to do if there are cues in the environment

Goes into working memory through perceptions

• Typing single keys is typically faster than continually switching the hand between the keyboard and the mouse and point to things on the screen.
• Skilled users develop plans of action, which they will want to execute frequently, so tailoring can capture these plans in the interface.

• Visual search--eyes must search through more. More (irrelevant info) interferes with Long Term Memory (LTM) retrieval of information that is relevant to task.
• Cluttered displays have the effect of increasing search times for commands or users missing features on the screen (Lavery et al)

Chartjunk (Tufte): "The interior decoration of graphics generates a lot of ink that does not tell the viewer anything new."

• Error messages in language user will understand
• Precisely indicate the problem
• Constructively suggest a solution

Allow search by gist--people do not remember exact system terms

Designer: Briefing (HE method, Domain, Scenario)

Evaluator:

• Two passes through interface (video in our case)
• Inspect flow
• Inspect each screen, one at a time against heuristics
• Fill out a Usability Action Report (we'll keep this simple in peer review)

UAR rather than “Problem report” because you report good aspects as well as problems -- you want to preserve them in the next iteration of the system!

• UAR Identifier (Type-Number) Problem or Good Aspect
• Describe: Succinct description of the usability aspect
• Heuristics: What heuristics are violated
• Evidence: support material for the aspect
• Explanation: your own interpretation
• Severity: your reasoning about importance
• Solution: if the aspect is a problem, include a possible solution and potential trade-offs
• Relationships: to other usability aspects (if any)

Should be A PROBLEM, not a solution

Don't be misleading (e.g., “User couldn’t find state in the list” when the state wasn’t in the list)

Don't be overly narrow (e.g., “PA not listed” when there is nothing special about PA and other states are not listed)

Don't be too broad, not distinctive (e.g., “User can’t find item”)

Ok to list more than one

This is subjective. Use your best judgement

• Make a claim about factors and support it with reasons
• Consider Frequency (e.g., All users would probably experience this problem because…)
• Consider Impact Will it be easy or hard to overcome. NOT is the task the user is doing important (put importance of task in explanation and in justification of weighting, if relevant).
• Consider Persistence Once the problem is known, is it a one-time problem or will the user be continually bothered? NOT low persistence because the user abandons goal (that is impact--can’t overcome it. If can’t detect and can’t overcome, problem persists).

0 - Not a problem at all (or a good feature)

1 - Cosmetic problem only

2 - Minor usability problem (fix with low priority)

3 - Major usability problem (fix with high priority)

4 - Usability catastrophe (imperative to fix before release)

“Discount usability engineering”

Intimidation low

Don’t need to identify tasks, activities

Can identify some fairly obvious fixes

Can expose problems user testing doesn’t expose

Provides a language for justifying usability recommendations

Un-validated

Unreliable

Should use usability experts

Problems unconnected with tasks

Heuristics may be hard to apply to new technology

Coordination costs

Heuristic Evaluation can be used to evaluate & improve user interfaces

10 heuristics

Heuristic Evaluation process

Individual: Flow & screens

Group: Consensus report, severity

Usability Aspect Reports

Structured way to record good & bad