Study On The Usability Of A Website Computer Science Essay

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

Customers sometimes do not find products on the web systems even when the products exist. This may be due to poor layout, design, navigation and many other factors. Many providers of web applications understand that the usability of their web site is an important factor, sometimes the outstanding determinant for building acceptance and thus for the frequency of use. Because of this growing awareness, many organisations are outsourcing expert usability services from "traditional" web agencies.

Given the insufficient quality of many poorly designed websites, however, it seems reasonable to assume that user centred design (UCD) methods are not employed consequently and seriously enough. According to Peisner & Rose (2002) Lack of awareness and low budgets might still be reasons for these poor quality websites. But often, also the necessary competences in usability methods are missing.

Introduction / Goals of the research:


An information system should provide the necessary functions so that the users can accomplish their tasks. The determinants of system acceptability are functionality and usability. System functionality is the degree in which the system provides the functions the users need to do their tasks, while system usability is the system's ease of use. Complete functionality of the system cannot be achieved if the system is not usable.


Usability refers to the extent to which the user and the system can 'communicate' clearly and without misunderstanding through the interface. Nielsen (1996) defines usability as a measure of quality of the user experience when interacting with a web based or traditional software application, and specifies five characteristics: ease of learning, efficiency of use, memorability (ease of remembering), error frequency and subjective satisfaction. In the context of the World Wide Web, "usability refers to how easy it is to find, understand and use the information displayed on a Web site" (Keevil 1998: 271). Usability is also defined as the effectiveness, efficiency, and satisfaction with which users of an application are able to achieve specific goals (ISO 9241-110, 2006)

Usability Engineering:

Usability engineering is a subset of human factors that is specific to computer science and is concerned with the question of how to design software that is easy to use. It is one of the key concepts of Human computer interaction. In this paper, the focus of Usability engineering is on Web Usability. This is because the world wide web is more and more used for commercial purposes and applications and therefore require usable Web sites as "Usability relating to the Web; Simply stated, if the customer cannot find a product, then he or she will not buy it." (Nielsen 2000:9).This defines the problem of study in this paper, explains the interest in evaluation methods for Sainsbury's web site.


Krömker (1999: 25), defines usability engineering as a set of methods to design user-friendly products and a process in which users, developers, and technical writers work together in order to integrate the knowledge and experience of all participants and enhance the quality of the product.

Fang & Holsapple (2000) have identified five features as joint contributors to Web site usability: task features, user features, provider features, system features, and environment features. The system features are the most controllable. And to enhance its knowledge acquisition usability for users engaged in certain kinds of tasks and operating within a given environment, it is important to know system features to be designed into a Web site.

In this research therefore, alternative navigation structures that can be designed into a Web site are explored. This navigation considers usability of the system feature from the objective standpoint of user performance, as well as the standpoint of users' subjective perceptions.

According to Nielsen (2008) web users scan lists by moving their eyes rapidly down the left edge. Menu items that are right-aligned make scanning more difficult. We know from eyetracking studies that users tend to rapidly move their eyes down the left-hand side of lists. People read the rest of a list item only if something catches their eyes in these left-most one or two words.

Aligning a navigation menu with the right margin might look cool, but the resulting ragged left margin severely reduces the speed with which users can scan the menu and select their preferred options. We have aligned our navigation menu to the left following this recommendation (see prototype).

Web Site Navigation - Definition

Web site navigation is the science and skill which you apply to a web site that helps visitors move from one page to another. Navigation on a web site is achieved by a collection of links (java applets / flash / images) that form the Web Site Navigation Menu or the Web Site Navigation Bar.

Text links in a web site navigation bar

The navigation bar can consist solely of text links. If the links are placed horizontally, they form the web site navigation bar [as I call it]. The links can also be placed vertically one after the other on the left side of a web page. The top horizontal row and the left side are two common places where you find the web site navigation system.

Navigation menu and images

Some websites have navigation menus which consists of a series of images which are linked to different pages. JavaScript codes are added to such images based navigation menu systems so that images change on mouse-over. This acts as a sort of visual guide - to let the visitors know which page they would go to if they clicked on the image.

Image based navigation systems can also consist of image maps, in which one large image is employed. Different sections of this image link to different web pages. A good example is the map of USA having sections for each state. On mouse over, you are shown by text the page relating to the particular state.

JavaScript Drop down dynamic menu systems

Dynamic drop down menu systems developed in JavaScript is another popular navigation type.. The reason for their popularity is their ease of creation and maintenance. Also cross-browser visually appealing drop downs can now be created with the help of programs.

Flash based web site navigation menu

Developers of websites sometimes wholly or partly develop sites in Flash. A flash menu has three advantages. Firstly, you can take use cool animation features of flash and provide your visitors a richer web experience. Thus, instead of simple image changes on mouse-over, the links on your Flash menu system can gradually fade in and out can play sounds on mouse-over. Secondly, if the navigation system has to be modified, it is easy to maintain the web site because you don't need to change each web page. The Flash navigation system is a single flash file. Replacing this file will change the navigation menu on all pages! Thirdly, with a little ActionScript you can create beautiful dynamic drop down navigation menus.

Flash menus however have two disadvantages. Users will require the Flash plugin to view the flash file and secondly, search engine indexing. Google has started indexing Flash files but does it follow links included in Flash remains a question.

Navigation using Java

Just like images and Flash, web site navigation systems can also be developed in java applets but this is not very popular.

Usability Engineering methods:

Usability engineering methods can be categorised into user focused and expert focused methods. A good example of expert focused method is heuristic evaluations. Nielsen (1997a, 1543) defines "heuristic evaluation as a way of finding usability problems in a design by contrasting it with a list of established usability principles". The established usability principles are listed in guidelines or checklists like Keevil's Usability Index (Keevil 1998) or Molich and Nielsen's nine principles for human-computer dialogue (Molich & Nielsen 1990) or the "Heuristics for Web Communications."

On the other hand user-focused methods rely on actual users to test the usability of a product. This process is called user testing, and Nielsen (1997a: 1543) says it "is the most fundamental usability method and is in some sense irreplaceable, since it provides direct information about how people use computers and what their exact problems are with the concrete interface being tested." The laboratory test with the thinking aloud method (Nielsen 1993: 195) is the most popular and most effective user testing method.

Usability Testing is the most powerful tool for Usability Engineering. It refers to the evaluation of information systems through the analysis of typical end users interacting with the system. Nielsen (1993, pp-17) shows that the more careful the method is, the more expensive it will be, often in terms of money and always in terms of required expertise. Simpler methods stand a better chance to actually being used in practical design situations. In this paper we have used usability testing because it is simple, cheap and does not require an expert. This is discussed further in the next chapter.

The phrase "discount usability engineering" was coined to describe the application of low cost methods for conducting such tests. The "discount usability engineering" Nielsen (1993) is based on the use of the following four techniques:

User and task observation


Simplified thinking aloud and

Heuristic evaluation

A.W. Kushniruk et al. (2001) discussed a new trend "virtual" usability engineering, involving remote testing. The basic strategy entails video recording all human-computer interaction (i.e., video recording all computer screens) and audio-recording all subject verbalizations as they interact with system prototypes. Such studies are being conducted using methods from experimental psychology, where subjects are asked to "think aloud" while interacting with the system. According to A.W. Kushniruk et al. (1997) the majority of usability problems and issues can be identified and summarised based on data collected from a representative sample of subjects, typically involving as few as 8-10 participants per study.

For example, subjects consisting of representative users of a system (e.g., physicians or nurses) are asked to interact with a system or prototype to perform a task, such as entering patient data. Subjects are also typically asked to verbalize their thoughts while doing the task. The complete audio and video recording of the interaction can then be analysed using methods involving the coding and classification of user problems.

J. Rubin (1994) explains that the information collected from such evaluation can be used to redesign a system. This information can include: (1) suggestions by users for improvements to both the user interface and system functionality, (2) identification of usability problems such as lack of consistency in interface operations, and (3) quantitative measurements including time to task completion and system response times.

The main objective in this paper is to:

collect suggestions by users for improvements to both the user interface and system functionality

identify usability problems such as lack of consistency in interface operations

carry out quantitative measurements including time to task completion and system response times

make recommendations for the improved system.

In system development especially where user requirements may be difficult to obtain using standard methods (such as interviews and questionnaires) and technical feasibility for some system functions may be unknown or uncertain, the appropriate approach is usability engineering.

Usability can be tested through different methods, which consist of observing actual users interacting with the site and collecting information in which the site is easy to or difficult for them to use, and the users' satisfaction with the system.

Approach and methods used

Some of the alternative names we have for usability include Usability Engineering, human factors, human performance engineering, human engineering and user experience design.

Performance and preference:

Performance relates to how quickly someone can do something, how accurately they do it, whether or not they are successful when doing it, and how long it takes someone to learn to do a task of some kind. These are all performance related metrics; speed, error, success, learnability

On the other hand we have preference related issues such as user acceptance or user satisfaction. User preference information is usually collected using questionnaires.

User centred approach:

If designers stay focused on the users and not the computer issues, they have a much better chance of constructing systems that have good usability.

Prototyping and usability testing:

The most useful testing method is "Human Performance Testing." This is test provides the best information because they have potential users performing real tasks using the web site.

Testers watch how they use the site, and when they have difficulties, the testers can assume that other users may have the same difficulties. These are valuable tests, requiring experienced testers and carefully chosen representative participants and conducted in a well controlled usability lab.

It is by creating the prototypes and testing them that we have the ability to iterate our design decisions. Therefore a prototype is created, tests conducted and changes made to the prototypes based on the outcome of the test, and then teat again. This is done over and over until the best outcome is achieved. The usable system is when the usability goals have all made.

Iterative design process gives us the ability to create prototypes and to conduct usability tests beginning early in the process.

The three important issues relating to designing highly usable systems. They include

focusing on users

effectively using guidelines and

doing iterative design

Extreme discount usability engineering:

We saw traditional user testing as is still often seen to be time- consuming and expensive, even usability advocates worry that unless an alternative to more formal methods is available, many organizations will continue to forgo user testing completely (Dumas, 2002). Researchers interested in discount usability engineering, therefore, focus their efforts on minimizing the cost and time required for user testing and maximizing the benefits accrued from those tests (Bias and Mayhew, 1994). These researchers have conducted many studies to answer questions about the relative effectiveness of various usability evaluation methods or the minimum number of user tests that will result in the maximize return on investment for usability evaluators (Jeffries et al., 1991; Desurvire, 1994). According to W. Gray and J Saltzman (1998), many doubt whether discount usability engineering techniques produce results that are worthwhile or even scientifically valid.

How many users should be tested and how long should it take?

In order to cope with both of these challenges, therefore, researchers working at the significant level of usability analysis require a new approach to usability assessment, one that does not depend on comparing the results from one user test to results from multiple tests, but on exploring how quickly usability evaluators transition from knowing absolutely nothing about an interface to making constructive recommendations for design improvements.

This mentality is very different from the typical concerns of usability evaluators, but it can be a very important one in situations where time and money put serious constraints on traditional user testing methods. A study of usability techniques capable of dealing with these extremes of tests and time helps us establish a lower bound on the ability of discount usability engineering to produce worthwhile results.

Background and Related Work

E-commerce has tremendously developed from "brick-and-mortar" retailing to "click-and mortar" environments raising considerations about the dynamics of this new business approach. To a great extent, what the electronic commerce and virtual retailing will offer depends upon interface design and the way that customers interact with the computer (Burke, 2002). Companies are changing their focus to strategic aspects involving how to best use this medium from creating web sites.

The first Web site, created by Tim Berners-Lee and Robert Cailliau at CERN (European Nuclear Research Center), consisted of a collection of documents with static content, encoded in the HyperText Markup Language (HTML). A screen shot taken from a NeXT computer running Tim Berners-Lee's original WorldWideWeb browser is shown below. The Web has evolved since then from an environment hosting simple and static hypermedia documents to an infrastructure for the execution of complex applications. There has been a lot of development in terms of web technologies and the Web has progressively become a multi-domain platform, that offers support for information delivery, and application execution. According to Casteleyn, Sven, et al. (2009), complex Web applications are almost omnipresent these days. Therefore design and development of complex web systems needs a discipline that deals all phases of realisation.

Screen-shot of the first line-mode browser

Tim Berners-Lee's original World Wide Web browser : from accessed on the 31st Oct 2010

The primary function of Web sites is to serve as a means for both the providers and the users of those sites to acquire knowledge (X. Fang, C.W. Holsapple, 2000). It is important therefore for the provider of a Web site to ensure that their websites are developed in a way that provides a high degree of usability to those who seek to acquire knowledge through it. The survey by GVU (1996, 1997, 1998), found that one of the main problems cited in using Web sites is that they are regarded as confusing and disorganised by their users. One contributor to such a problem may be a navigation structure that makes it awkward or difficult to traverse a Web site's pages when seeking knowledge.

Navigating complex websites:

Navigation systems are an important component that helps web users know where and what to find among available Items in a shopping website. The most significant issue that determines the success of a Website is a well-designed navigation system, such as global navigation and local navigation (Rosenfeld & Morville, 1998). Contextual navigation is another type of navigation system used ''to suggest related topics and content in other areas of the site'' (Farnum, 2002, p. 38).

When designing and organising information in our shopping website prototype, a hierarchal navigation was taken into account. Hierarchy is vital to Websites' design since most authors agree that it makes them highly accessible. The table of contents and the index provide users with better access to the navigation system on the Website pages (Rosenfeld & Morville, 1998).

Tarafdar and Zhana (2005) in their study of successful websites, point out that the navigation system is one of the critical Website characteristics. Nayak et al. (2006), also consider navigation systems as a compulsory component in the evaluation of any Website. According to Nielsen (2000) who adopts the navigation system in heuristic evaluation of Websites, one would tend to consider this as a sufficient evidence that navigation systems are a necessary condition for the successful performance of Websites with regard to the finding of and accessing information.

S.Eroglu, K. Machleit, and L.Davis (2001) Online shopping is growing quickly and this growth is fuelled by the advantages of online shopping (Eroglu, Machleit, and Davis, 2001). In light of this growth, firms increasingly rely on e-commerce solutions to reach their profit objectives (Freemantle, 2002).

Navigating a large and complex website is difficult. According to R. P Darken, & J.L Siebert, Navigation is then the process of determining a path to be travelled through the chosen environment. This means that navigation addresses the question of where do I want to go? To correctly answer this question, there are four basic questions one must ask himself when on the web; where am I? Where do I want to go? Am I on the right path? and Am I there yet?

William Horton et. al (1996) gives recommendations on structural wayfinding. They present a generic web page design and advocate using text and icon navigation buttons as well as path buttons to make life easier. Clicking on these buttons take you to a different place in the website's organisation or structure.

The web content are arranged in the website and the menu navigation represents the table of contents in a document. Hierarchical organisation is shown using indentation, and hyperlinks allow the user to go to the desired page directly from this table of content and the "Home" function takes the user to the user to the home page of the particular website.

Menu analysis:

The main usability concerns in Web Applications attributed to the ability to reach the intended target page include navigation and site structure. The purpose of navigation tools such as a sidebar with links is to give the user the required sense of orientation, not only identifying pages he can reach but also the current position within the Web site (Nielsen, 2000, pg. 188). And according to Y.L. Theng (2003) & R.H. Trigg (1988) without a solid concept of the structure of a Web site, the users will easily become lost due to the additional dimensions of movement as opposed to the forward and back in traditional text.

There are several types of Web site structures such as trees and tables, both of which are commonly encountered on the Web. While a tree-based navigation will lead users deeper into increasingly concrete categories, table-like structures define categories based on the values of the table's attributes. According to Nielsen (2000, pp. 15, 200-207) a linear structure similar to that of traditional media is considered to be useful only in exceptional cases. Regardless of the structure, a good home page is important and should give a good overview of the site as it is the starting point for most users [Nielsen, 2000, pg. 166]. And that the purpose of a Web Application should be immediately obvious to an incoming user [Nielsen, 2003 & Theng, 2003].

The webmaster decides on the web structure and the link are then created which emphasise this structure. Additionally, the design and realisation of individual links can also raise problems if not done properly. This means that the name of the link should be concise and hint at where it leads to (Borges et. Al, 1996), giving the application what Garzotto et al. call self-evidence.

There various methods to evaluate the quality of link descriptions. For instance, (Nielsen and Molich, 1990) shows that heuristic evaluation can involves a systematic inspection of the design of a web site, including link descriptions. Here the usability experts perform the analysis manually. Tullis (2007) methods are centred around the users of web sites, such as card sorting, and walk-through methods of Bias (1991), comprise of studies with representative groups of end-users.

There are several methods developed to improve web menu structure in terms of web usage data. A set of web log files are accessed and a link structure obtained and analysed to determine which links need to be added or removed. Usability can be increased by placing more frequently used items at higher positions in a menu or by adding links between pages that are often viewed in the same sessions. Some methods leave it open for the web designers to change them as they sees fit (e.g. Smyth and Cotter, 2003; Wu et al., 2005; Wang et al., 2006). In other researches, possible adaptations are recommended but left to the webmaster to decide which adaptations are implemented (e.g. Nakayama et al., 2000; Hollink et al., 2007). These methods have in common that they improve the topology of the link structure, but not the descriptions of the links.

For systems that add Links autonomously, provide description for these links using engineers' handmade rules. A good example is the m-Links system (Schilit et al., 2002) that creates navigation menus for web pages that are suitable for viewing on mobile phones. The links in the menus are labelled with texts from predefined sources, such as page titles and URLs. Brusilovsky (1996, 2001)talks about adaptive link annotation which is the process of attaching visual cues to links to help users select the most relevant links. Link annotations can provide specialised support as they are adapted to the personal needs of each user. A disadvantage of this approach is that elements are added to a page even when selecting the correct link is straightforward. This therefore means that the annotations unnecessarily increase the visual complexity of the page.

Both the structure and the description of the links can be optimised according to Nakayama et al. (2000). They suggest an algorithm that detects page pairs that are similar in content, but that are not frequently visited in the same session. If the pages are not yet linked, then they suggest to add a link, and if there is a link already, then they suggest a change of design in such a way that this link's visibility is improved. Some of the suggestions to change the layout include changing the link anchor or the text preceding the link. Nakayama et al. (2000) research has limitations as they can only detect problematic links between pages that are very similar in content. It is also important to find problematic links between pages that are related in terms of user relevance, but that have different contents.

Srikant and Yang (2001) propose a method to discover the location in a web site where users expect to find certain target pages. They assume that users follow links to the location where they believe a target is located and backtrack when they find out that no link to the target is present at the expected location. This method computes for each target page the positions where users frequently backtrack and recommends to add links to the target at these positions. Srikant and Yang method search for the end points of these paths (the backtrack points) therefore solving the problems at the end points by adding links to the target pages. In contrast, Hollink et al. (2009) determine the source of the problem, the point where users deviate from the optimal path and solves the problem at these points by improving the link descriptions that gave users incorrect expectations about the contents of the underlying pages.

Users who navigate unfamiliar web sites in search of information choose links on the basis of link anchors and the text surrounding the links. Landow (1989) refers to it as the rhetoric of departure which means the anchor and context of a link together form the description of the link. Users can make the right choices and reach their target information efficiently if the link descriptions are accurate, but if the link description is ambiguous, the users are forced to try out various other links to see if they can find what they want. In this paper, we will try and correct the link descriptions to help the users navigate the website as efficiently as possible.

Fitts's User Interface Applied to the Web

Fitts' Law states that the time to acquire a target is a function of the distance to and size of the target. This dictates that shorter mouse movements are better: it is always faster to click a target if it is closer to your starting position. Thus, placing the navigation rail next to the scroll bar will usually save users time over placing these two frequently-accessed areas on opposite sides of the window.

The basic idea in Fitts's Law is that any time a person uses a mouse to move the mouse pointer, certain characteristics of objects on the screen make them easy or hard to click on. The farther the person has to move the mouse to get to an object, the more effort it will take to get to. The smaller the object is, the harder it will be to click on. The menu buttons in the prototype are left aligned big enough. The page opens a page with more clickable icons that are well labelled and clickable.

Web site usability measurement:

Shackle (1991) describes usability as "a technology's capability to be used easily and effectively by the specified range of users, given specified training and user support, to fulfil the specified range of tasks, within the specified range of environmental scenarios". If websites are not as usable as they could be, Web site users can encounter various problems when trying to acquire knowledge from the Web sites and trying to use a Web sites' functionalities. This can lead to user's dissatisfaction and blockage of knowledge acquisition.

There is an extensive set of usability evaluation methods that can be used, with a variety of performance measures [e.g. R. Benbunan (1999, 2001), A. Gnisci, F. Papa, S. Spedaletti (1999), L. Van Wares (2000), M. Morris, J. Turner (2001)]. In this research, we gauge usability in terms of how well subjects perform a series of knowledge acquisition (KA) tasks with a particular navigation structure; this performance is measured by navigation speed and task correctness. We also gauge usability by directly asking subjects to reflect on and rate how usable the navigation structure with which they worked was for performing the series of KA tasks. In this paper we gauge usability by directly asking subjects to reflect on and rate how usable the navigation structure with which they worked was for performing the series of knowledge acquisition (KA) tasks.

System Design and Testing

Prototyping and usability testing:

The most useful testing method is chosen for this project is "Human Performance Testing." This test will have potential users performing real tasks using the web site, therefore providing the best information.

The testers watch how participants use the site, and when they have difficulties, the testers assume that other users may have the same difficulties. Because these tests are valuable, they'll require experienced testers and carefully chosen representative participants and conducted in a well controlled usability lab. More about this in the next chapter.

A prototype web system is created for this purpose and tested so that we have the ability to iterate our design decisions. Therefore a prototype is created, tests conducted and changes made to the prototype based on the outcome of the test, and then test again. This is done over and over until the best outcome is achieved. The usable system is when the usability goals have all met.

Iterative design process will give us the ability to create prototypes and to conduct usability tests beginning early in the process.

There are three important issues relating to designing highly usable systems. They include

focusing on users

effectively using guidelines and

doing iterative design

We will not discuss this as it is beyond the scope of this project.

The prototype:

We designed a prototype with a different navigation to the sainsbury's system. Though the navigation systems in the two systems were different, the site contents were the same in both the prototype and the sainsbury's web system. A brief description of each navigation approach is given below;

Sainsbury 1:

Expand/Collapse Menu. Clicking on a "closed" menu opens it. Clicking on an "open" menu closes it. Menus are not closed automatically, though "Open All" and "Close All" functionality is provided. Some people think this type of control is more suited to an online "book."

Sainsbury 2:

Yahoo Style Menu: Clicking on a heading opens the menu item to show its sub-items, and automatically closes the previously opened menu. We have seen usability issues in which the "auto-close" feature annoys some users.

How we conducted the study:

We wanted to conduct the study with a limited number of participants, therefore we approached undergraduate students in a business college in Manchester. We wrote an email to the college administration to allow us to conduct the experiment in their college.

Each participant was randomly picked and assigned the two navigation styles, and that the participant was to use both of the navigation styles complete a set of 6 tasks. Each task required the participant to navigate through the prototype to find a particular product he needs to purchase. The participant had to navigate to the product and add it to the shopping basket. The task ends when the item is in the shopping basket.

Here are samples of tasks:




For each task, we collected the following data:

1. How long it took the participant to complete the task.

2. A participant-provided rating of the ease/difficulty of each task.

3. The accuracy of the participant's answer.

At the end of the study, we asked for:

1. An overall rating of the effectiveness of the navigation technique that the participant used.

2. Comments about the navigation technique.