COMP6471 WINTER 2003 User-Centered Design Instructor: Shahriar Ameri, Ph.D. Student: Pedro Maroun Eid, ID# 5041872. Date of Submission: Monday, March 10, 2003. (Week 9)
Outline Outline... 2 ABSTRACT...3 Introduction... 4 Usable Systems... 6 Principles... 7 Process... 9 Conclusion... 14 References... 16 2
ABSTRACT Since systems are intended to be run and used by users, then what a user wants is a system that does what he wants. He surely wants to find it easy and straightforward to use the system being developed. Simply, it should satisfy his needs. This paper will discuss the various matters that concern User-Centered Design, denoted by UCD, with respect to other software engineering approaches. It will talk about the User-Centered Design which is much more efficient than other software engineering methods in terms of process, benefits, and advantages. It will describe the cons of UCD and will state this revolutionary as of the best developed to date. 3
Introduction What s a system that does a lot of things, or is powerful, but can only be operated by the engineer that constructed it because he knows how it operates? Even for him, it might be confusing in doing a particular task with his system. Therefore, in order to accomplish this task, systems are best built upon the intended users needs. This whole process of including the user in the design process is referred as User-Centered Design. This method forces to put a user in the development team and to always consider him as a key ingredient to the software product. In doing so, the intended user could provide user tasks and what is referred by the ways of interaction with that program. Another benefit in including the user in the design of a product is that he will feel a kind of ownership of the product and will be able to interact with it freely and independently. He will be happy to contribute to his future system which he will already know before he even gets hands on it. UCD, like any other software design, has a sequence of steps that should be followed for ensuring success. Actually, UCD is integrated in other software models that show high reliability. This new model is made to follow some conditions in order not to miss any critical issues that form UCD. The best model in software development is probably the iteration model. An example is the prototype model: This model is circular, it uses iterations to complete parts of a system and then finishing them one by one will result in the final product. 4 main parts are included in this model and most other models as well and they are: Analysis, Design, Build, and Test. To include UCD in this process, we integrate it in the parts. The Analysis is made to include user opinions, tasks, and 4
choices. These make the developers more aware of what the users want and not what the system should provide. The Design should include a prototype for the users to use, test, and report opinions. The Build should be iterative and the collected information from the users should be dealt with in the next iteration. And finally in testing, the developers should test the product on a representative set of users and should always stay in contact for support and changes required after delivery. Usability Studies are performed in building and testing, and Maintenance. They are studies performed on the usability of a certain prototype or a first version release. It is highly important to perform these studies in order to result in the actual user-desired system. A user-centered design and system is best complete with a good user interface where the user decides its look and behavior. These issues are more explained in the process part of this paper. UCD mainly concentrates on knowing the user. 5
Usable Systems Usable systems is an important aspect that has been studied and which makes usercentered design more looked at and distinguished from other software engineering approaches. ISO 13407 relates to usable systems and differentiates between a usable computer system and another one. It puts guidelines to attain a usable system. The standard has these following required rules: Getting a clear understanding of user and task requirements Employing persons from a variety of disciplines and roles in the process (e.g. user interface designers, marketing, end users, technical authors, etc...) Practicing iterative design processes Evaluating designs against requirements These summarize the requirements for a usable system and they are all included in UCD as basic rules. Thus UCD is surely considered a good usable system in contrast to other approaches which may or may not enforce these rules. Poor usability very much contributes to a failure in a software system. Imagine an Automated Banking Machine that u cannot find your way through because it does not have a logical task for the withdrawal option that you are trying to perform. A lot of software actually fails because of that usability problem. UCD is a method that embeds user involvement in the software process thus making the software tailored specifically according to the users needs and requirements. It involves the user in all the phases of the development unlike conventional methods which gather the users feedback when the software is delivered. Real time user involvement allows the development team to make decisions based on information from real users. 6
Principles The principles of user-centered design are probably familiar and easy to perform. They concentrate on the user requirements and try to get information from the user as much as possible. For a good UCD design, some principles should be always followed and taken consideration of. The principles as defined by the IBM UCD development community, a leading research that focuses on UCD, are composed of six. The First is to set the business goals which includes all the business aspects of the software. It should consider the market in which the system will be used, the users that will operate this system denoted by intended users, and the competition on the system from other similar systems. This principle may be considered in other approaches of software engineering but not very much enforced. The second principle is to understand the user. This principle focuses on listening to the users needs and requests. To do this, u must include him in the design process and as said before he should be a key ingredient for designing the whole system. User tasks and User Stories are used to collect information on what and how the user uses the system for. This principle is barely considered in other approaches since it requires effort, money, time, and, most important of all, knowledge. The Third principle is Assessing competitiveness. It focuses on testing the obtained user tasks on your design and having a look on the competitor s way of doing things in order to result in a better system. The fourth principle is designing the total user experience or making the user happy with the product and proud to have it. It focuses on internal and external aspects like documentation, upgrade, support, installation, advertising, etc... 7
These are meant to be considered by the design team and are important to give a good impression of the product to the customer or the user, resulting in his satisfaction. The fifth principle is Evaluating the design. First, an incomplete design is given to the users to test. This is called the software prototype. Prototypes and releases are always given upon iterations and the feedback received from the users after an iteration is used to ameliorate the design and development of the next iteration or release. Finally, the sixth principle of UCD is Maintenance and support. This focuses on the aspects to consider after the product has been delivered. We all know that there is no such thing as a perfect software; this is why maintenance and support should be there along with documentation in order to solve any unexpected problem. This should also include listening to the users and watching them and using their feedback to make a better system. 8
Process User-Centered Design, as said before, could be integrated in any software process. However, it best works with the prototype model because of it concept of iteration and prototype releases. This figure presents the iteration cycle of the prototype model. It starts with Analysis which is collecting information about the project and having Design in mind what the project is in addition to a small outline for the project to proceed with. Next is design which is designing Analysis Build or Revise the collected information and data, along with performing the design decisions for User Test build the first release of the product. In this step, there could be a non-working prototype that is given to the users to test and give feedback. After design is the build or code state in which coding and design is physically applied. This state turns to the revise state starting from the next iteration. User test is then done by giving the release to the user to test and ask for corrections. So the whole process waits till the last step to involve the user. Whereas if we integrate UCD in the previous design, we would change the process slightly to include the user from the first step and along all the steps of the iteration throughout the whole process. In analysis, collect information about the user and about his way of doing things. Letting him say the user tasks he would want to perform contributes in a good design from the first step. User Stories are small stories written by the user that explain a task that he needs to be done. Some initial questions would be: 9
What do you want the product to do for you? What are your priorities when using the software? Which functions will you use most often? In the design phase, the task is developing and testing the conceptual model which will be the probable system after implementation. During design, you should consider the interface design (layout, color, etc...). This makes a step ahead for implementation. It s not good to implement the user interface directly without designing it first. Upon coding or building, you design the human computer interactions according to the technology available and what the user requests. Finally, the user is also included in the final step by performing usability studies on the product you have come up with. This is one iteration; this is done over and over again until the final product has reached maturity and is ready to be delivered. As described by IBM, there are six stages to perform good UCD software. Analysis which includes market definition, task analysis, and competitive evaluation, Design which uses the information from Analysis taking into consideration what the user expects and wants to have, and finally Coding and Testing which include both always evaluating the design with actual users and observing their experiences with it and benchmark assessment against competitors which verifies that the product has met its requirements (users evaluate your product with other available products). This is also an iteration which repeats until the final product is reached. The last step always uncovers new ideas and changes to be considered in the next step or iteration. This method is clearly similar to the one I described previously; it s only adding some competition requirements to have an advantage over other products. They both focus on involving the user every step of the way in order to get the best product that could be made. The RESPECT method also describes the UCD embedded in the spiral model. The figure is explained and states the 4 10
design stages: Context, Requirements, Design, and Test. You understand the user in the first step, the collect the requirements from the user as the second step, design and develop the operational prototype which should be evaluated by the users, and finally test the prototype to ensure that the organizational requirements are met. The spiral method differs in the prototype method in that it continues from where it stopped in the last phase rather than applying the changes to the whole process. All processes show the differences between the original approach and the usercentered approach that is integrated in the original approach. It is assumed that the discussed approaches are familiar. 11
Cost Benefits Why is UCD the best design practice that could be performed? A study done by the Software Engineering Institute in 1995 as listed in User Centered Design showed that 1 in every 3 software systems is cancelled. On average, projects are 50% over schedule which means that they are too fast. And that 70% of the total projects performed are failures that don t function or aren t used. You may be asking why, but the answer is clear, we will come to it later. Another Report called the Chaos Report, performed by the Standish Group in 1995, also had interesting results about software use. The report goes on to say that US companies spend more than $250 billion each year on IT application development of approximately 175,000 projects. The average cost of a development project, broken down by size of company is: Large company $2,322,000 Medium company $1,331,000 Small company $434,000 The percentages of software products life and costs are given as follows: 31.1% of projects will be canceled before they are completed. 52.7% of projects will cost 189% of their original estimates. 16.2% of all software projects are completed on-time and on-budget. In large companies, only 9% of all software projects are completed on-time and on budget. In large companies, only 42% of the software products contain the originallyproposed features and functions. The major factors that cause software projects to fail are also the key ingredients that can reduce project failures. The top 10 major factors, ranked by importance are: 1. User Involvement (19) 2. Management Support (16) 3. Clear Requirements (15) 12
4. Proper Planning (11) 5. Realistic Expectations (10) 6. Smaller Project Milestones (9) 7. Competent Staff (8) 8. Ownership (6) 9. Clear Vision and Objectives (3) 10. Hard-Working, Focused Staff (3) Tom Landauer, in his book, The Trouble with Computers says if software were more intuitive and easy to use, people could spend more time doing their jobs. He predicts that if every software program were design for usability, productivity would rise by 4% to 9% annually. The average user interface has 40 flaws. Correcting the easiest 20 yields an average improvement of 50%. The big win occurs when usability is factored in from the beginning, yielding efficiency improvements of over 700%. In her book, Cost-Justifying Usability, Deborah Mayhew said consider a simple transaction: data entry clerks filling in entries on a form. Given 20 users working 250 days a year, performing 80 transactions a day = 368,000 transactions per year. If you can reduce the time to complete the transaction by 10 seconds, you can save 1022 hours, or 25.5 person-weeks. The estimates given by Tom Landauer are a bit exaggerated because it s not possible to yield in such an improvement. However, Deborah s argument is logical and, anyways, since there exists an improvement, then it s worth the try. All the above information is obtained from User Centered Design, unknown author. All these arguments clearly show the need to use User-Centered Design. Since it relieves most of these problems and consumes better then it should be used in software engineering. I m sure economy will grow better! 13
Conclusion Therefore, UCD is based on these important aspects which are not or are barely available in other software engineering approaches: 1. Know thy users. 2. Understand the tasks the user is trying to accomplish. 3. Involve the user in iterative design throughout the development process. 4. Practice usability testing and evaluation throughout the development process. User-Centered Design is surely a great advance in the world of software engineering and information processing. It has some interesting advantages. It enables the development of easy-to-use products, satisfies customers, decreases expenditures, and makes business better and easier. However, many organizations still don t use UCD. Many developers assume they know what the user wants and likes to have. These assumptions make software dependant on technology rather than user needs. In Susan Dray s article The Importance of Designing Usable Systems. published in the January, 1995 edition of interactions as described by the User Centered Design unknown author listed these benefits: Reduced errors Lower support costs Lower initial training costs, and greatly reduced retraining Less productivity loss when the system is introduced, and more rapid recovery More focus on tasks to be done, rather than on the technology tool Lower turnover and better morale Reduced rework to meet user requirements High transfer of skills across applications, further reducing training needs Fuller utilization of system functionality Higher service quality Higher customer satisfaction Increased usability Greater user acceptance early on 14
Detection of issues earlier Reduced documentation and support Increased productivity Reduced overall costs Greater sense of accomplishment for designers Success is not always guaranteed. Every project is different. Every failure and every success contain lessons to be learned and applied on the next project. Each project and each user is different, thus making the software engineer experience the most important factor to be competent. Nobody notices good things but everyone notices bad things. And as my Human Computer Interface professor says, we re now at an advanced level of technology that systems really do depend on the users comfort and enjoyment. We should provide that for him in order to sell our products. Let s make software products a bit better. 15
References IBM s website for UCD development: o http://www-3.ibm.com/ibm/easy/eou_ext.nsf/publish/570 User Interface Design: o http://www.chesco.com/~cmarion/design/uidesign.html User-Centered Product Creation in Interactive Electronic Publishing: o http://www.vnet5.org/pub/approach/design.html User Centered Design o File://Usig-ucd1.pdf User Centered Requirements Handbook o http://www.ejeisa.com/nectar/respect/5.3/a.htm Software Engineering Lectures: o http://www.it.lut.fi/opetus/99-00/010758000/lectures/lecture4.html Software Engineering: A Practitioner s Approach. Roger S. PRESSMAN. Fifth Edition, McGraw-Hill International Edition, Computer Science Series, 2001. 16