Heuristic evaluation

Sources of usability criteria Organizational goals Pre-existing, general: heuristics and guidelines –Research –Convention –Consensus of experts –Legal requirements (e.g., accessibility) Empirical: –User and task analysis –Competitor sites (assumption: their goals apply to your site, your users, also)

Heuristics Defined: –Rules of thumb –“providing aid or direction in the solution of a problem but otherwise unjustified or incapable of justification” – Webster’s 3rd Method: –Find/develop set of criteria –Apply to software/website/etc Which are violated How severe

Heuristic evaluation Uses Methods Sources of heuristics, criteria –Existing sets of guidelines, heuristics –Customization for specific applications Applying the heuristics Advantages, disadvantages of heuristic evaluation

Uses of heuristics and guidelines Competitive testing Design guidance Evaluation criteria Summarize lessons learned for future design guidance The process of developing and applying heuristics helps design/evaluation group(s) to define and agree on goals, priorities,evaluation criteria

Heuristic evaluation: Method Multiple evaluators Experts (in heuristic evaluation) Representative users How many? –Nielsen’s rule of thumb: 3 to 5

Nielsen’s basis - # of evaluators

Empirical Evaluate for each major kind of use/user Keep going until you don’t find much that’s new and critical

Method Develop/identify heuristics [weight by importance] Apply heuristics –Individually –Collective debriefing Determine severity Make recommendations for improvement Inter-rater reliability: consistency x (similar) raters

Three Sets of Heuristics General –Nielsen –Other guidelines –accessibility For this kind of application (web, homepages, e-commerce sites…) For this specific application

Severity ratings e.g., Nielsen: Based on frequency, impact, persistence Possible severity rating scale: 0 not a problem 1Cosmetic – need not be fixed unless time available 2Minor – low priority 3Major – high priority 4Catastrophe – fix before release

Nielsen’s heuristics Visibility of system status Match between system and the real world User control and freedom Consistency and standards Error prevention Recognition rather than recall Flexibility and efficiency of use Aesthetic and minimalist design Help and documentation

Bruce Tognazzini’s principles Anticipation Autonomy Color Blindness Consistency Defaults Efficiency of the User Explorable Interfaces Fitts's Law – size and distance Human-Interface Objects Latency Reduction Learnability Limit Tradeoffs Metaphors Protect the User’s Work Readability Track State Visible Interfaces

Guidelines and checklists E.g. web guidelines from –Design guidelines Page length, layout; font; graphics –Design process guidelines Set performance and/or preference goals Create and evaluate prototypes –Usability guidelines Content/content organization; navigation; download time Accessibility

Web usability: some relevant characteristics of the web Uses, users may be hard to define –Multiple and varied users, contexts, technology, and so on –Uncontrolled –Must design for naïve and experienced users The web itself as the context of use –Users may be infrequent users who readily move to another site –User expectations formed by other sites Users’ Web-specific concerns –Privacy, security –Users may not know who you are Demonstrating credibility

Nielsen’s113 Design Guidelines for Homepages Determining Homepage Content Vertical Industry Segments Communicating the Site's Purpose Communicating Information About Your Company Content Writing Revealing Content Through Examples Archives and Accessing Past Content Links Navigation Search Tools and Task Shortcuts Graphics and Animation Graphic Design UI Widgets Window Titles URLs News and Press Releases Popup Windows and Staging Pages Advertising Welcomes Communicating Technical Problems and Handling Emergencies Credits Page Reload and Refresh Customization Gathering Customer Data Fostering Community Dates and Times Stock Quotes and Displaying Numbers

Nielsen’s Homepage Design Statistics (what people are used to; what others have found useful) Download Time Basic Page Layout Page Width Liquid Versus Frozen Layout Page Length Frames Fundamental Page Design Elements Logo Search Navigation Footer Navigation Site Map Routing Pages Splash Pages Frequent Features Sign In, About Us, Contact Info, Privacy Policy, Job Openings, Help Graphics and Multimedia Pictures, ALT Text, Music, Animation Advertising Typography Body Text and Background Colors Link Formatting

Guidelines based on research about users Guideline: Design for monitors with a screen resolution of 800 x 600 pixels. Comments: There is a definite trend in monitor design to go from screen resolutions of 800 x 600 pixels to screen resolutions of 1024 x 768 pixels. Five studies of screen resolutions were reviewed. Two of the studies reported that the largest number of users (53%) were using screen resolutions of 800 x 600 pixels (27% were using 1024 x 768). However, three of the studies reported that the largest number of their users (43%) were using screen resolutions of 1024 x 768 pixels (only 24% were using 800 x 600 pixels). Only about 7% of users are using 640 x 480 pixels, and about 13% are using higher resolutions (1280 x 1024, 1600 x 1200, etc.)

Research II Guideline: Design for connection speeds of 56 kilobytes per second (kbps). Comments: Sixty percent of users use a 56 kbps connection speed or slower. The remaining users have faster connection speeds (ISDN, DSL, Cable, T1, etc.). Actual connection speeds are about 38% lower than modem speed capability. This means that users with a 56 kbps connection actually have a connection averaging about 35 kbps. If you have data indicating that most, if not all, of your users have slower or faster connection speeds than 56K, determine what is appropriate.

Specialized Heuristics: e-Commerce and Order Forms From Shows total cost Shows itemized costs Shows product names and/or descriptions Allows the user to change the quantity easily Provides an option to save an order and complete it later Provides details on any other charges on the order Provides details on shipping options and charges Provides shortcuts for repeat visitors to make transactions faster Allows users to easily move from the order form to shopping and back again Provides security information Provides users with an alternate offline way of ordering Allows users to view and/or change previous orders Does not require users to register before a purchase

E-commerce II Customer support: Supporting users before, during, and after a purchase. Trust: Establishing trustworthiness. Product Navigation: Enabling users to browse products easily. Product Information: Providing the product information that users want, need, and expect. Purchase transaction: Providing easy means for users to purchase products.

Customer Support (IBM guidelines) Provide contact information on every page Provide assistance when users have forgotten their passwords Provide clear and informative error messages Address users' frequently asked questions Provide simple definitions and explanations of important terms Provide product selection assistance Provide assistance to guide users through multiple step processes

Heuristic evaluation: plusses and minuses Benefits –Low resource requirements –Usually find many problems fairly quickly –Easy to repeat in iterative design –Easy to communicate –Usually easy to get agreement on a basic set of heuristics –Face validity Limits –Can be superficial –Focuses on easily-seen problems; harder to find more subtle problems associated with in-depth use, repeated use –Can be deceptive – assumption that evaluation has been more complete and thorough than it has been –How similar to users are experts? How expert are users? –How appropriate are the heuristics to THIS site? –Tends toward a short list of heuristics –Trade-offs among heuristics, the fixes needed?

Some key points We need to differentiate among official standards, how people generally do things, and expert opinion Usability in a changing environment: what people are used to, their technology and expectations, are continually evolving Heuristics need to be customized to goals, context Most guidelines are solutions to problems; have to ask what is the underlying rationale, goal Trade-offs among different goals, heuristics often have to be mad

Heuristics - observations Heuristics shape what we see What we think the problems may be help shape decisions about heuristics As a practical matter, tend to focus on problems, not what a site/system does well CONTENT is not addressed by most heuristics and guidelines

Kinds of usability criteria Usefulness: task-related –Functionality –Content –Integration with tasks, tools, activities User productivity –Speed –Ease of learning –Ease of use Cognitive effort; tasks, activities “cognitive friction”: resistance encountered by human intelligence when it engages with complex set of rules that change. Quality of user experience –perceptions, feelings, opinions

Types of measures Performance metrics –Speed of response, availability, errors (e.g. dead links) User assessment –opinion, perception, feeling Observable user behavior –Time spent, user errors, operations performed… –Need to be cautious about drawing inferences, e.g. time spent on a page

Performance Metrics “Only 5.3 percent of visitors could load the NYTimes' home page within 30 seconds” on Nov 12, after the plane crashed on Long Island. “MSNBC.com had 63 percent availability and response times of 26 seconds for those who could get a page during the hour following the crash.... “New York Times Digital served almost 13 million page views on Nov. 12, about 3 million more than its daily average. Following the crash, the company pulled ads and rearranged its page content so that text and Web links would load first. Prior to Nov. 12, it added 50 percent more server capacity following its brownout of Sept. 11, and reconfigured its load balancers for better efficiency.”

Evaluating usability Identifying values on the measures/criteria (how much?) Setting goals for levels of performance, interpreting values (how good?) “It's difficult to believe the Times would find it OK if 95 percent of its print subscribers didn't get their newspapers in a timely way, or that MSNBC would conclude it had served customers well if 37 percent couldn't receive a clear broadcast signal.”