Summary

Iterative prototyping is a lightweight process which allows the designer to improve weak points in previous versions. When a prototype is built and put to use on real participants, the designer can revise the design and re-test the new prototype. The first iterations provide most of the gains in design realisation, with progressive iterations producing smaller potential for improvements.

Why?

Any successful design generally includes some form of iteration and prototyping, because even the best designers cannot design error-free websites in a single attempt. For this reason, websites should be built around the concept of iteration. (Bury 1984; Buxton 1980; Neilsen 1992).

This iterative development method is a lightweight process model based on multiple successive iterations that allow the designer to '...find the weak points in the design of the previous iterations' (Sotirovski, 2001, p. 67). Problems with the website interface become apparent early in the design process, when people actually sit in front of the website and use it (Fowler 2003). This reduces the risk of costly design changes later in the design process.

A prototype can be as simple as a drawing on the back of a dinner napkin or as complete as a high-fidelity software interface. The purpose of a prototype is to give the audience an understanding of ideas without overwhelming them with unnecessary details (Arnowitz, Arent & Berger 2007).

Prototypes can be developed with the intention that they become part of the final design (evolutionary prototype) or they can be discarded (throwaway prototyping) after serving their purpose (Hardgrave, Wilson & Eastman, 1999). The collaboration between the website designer and end-user is at the heart of prototyping (West, 1986). Cooprider & Henderson report that '...The essence of prototyping is in the dialogue between builder and user, within the resource constraints established by management' (1990, p. 72).

The user's basic requirements are identified and developed into a working prototype. The prototype is then put to use, to identify any problems or contextual misfits. The designer may then revise the design and re-test the new prototype. Following this repetitive process, incremental progress is made until a solution can be described in terms of the original problem (Adams & Atman, 1999).

Within this exploratory process, the goal of any prototype design project should be to answer one main question:

  1. 'What features should be incorporated into the prototype to meet the contextual requirements of the end-user?'

Given the broad nature of this question, the first phase of the process should be one of knowledge discovery. The enquirer needs to consider issues of contextual design detailed in existing literature, which are then structured into broad guidelines that will underpin later design choices. A focus group can then identify general end-user requirements. Using both the literature review and the focus group discussion, a broad set of key design features are formulated, for the development of the initial prototype.

'Key design features' comprise of general design factors dealing with the look and feel of the site. It is not necessarily the intention of a prototype to capture all the functions or user tasks to be performed on the final website, nor is it to consider the strict technical limitations of the design?for example, network bandwidth or cross-platform browsing issues. The planned outcome of the prototype is not a fully functional and deployed website but rather a consensus about critical design factors that meet the contextual requirements of the end-user.

Nielsen (1993b) presented a graph on iterative design to eliminate usability issues. The graph has been adapted here (Figure 1.1) to show how a design is accomplished as a function of the number of iterations. While this does not always work in practice, each iteration should produce a design that is closer to a solution.
 

Figure 1.1: Design Realisation as a Function of Design Iterations

iterative design

(Adapted from: Nielsen 1993, p. 33)

The first iteration provides most of the gains in design realisation, with progressive iterations producing smaller potential for improvements. The reconceptualisation shown in Figure 1.1 is a direct result of user feedback.

Recruiting Participants

Although user involvement in the design process might appear to be 'complicated and time-consuming', involving users in all stages produces many benefits. Participants are the typical end-users of the website. In determining the number of participants to test, Jacob Nielsen (2012) has advocated a 'discount usability engineering' approach since way back in 1989. He believes '... if you want a single number, the answer is simple: test 5 users in a usability study. This lets you find almost as many usability problems as you'd find using many more test participants' (2012, n.p). He explains '... Discount usability often gives better results than deluxe usability because its methods drive an emphasis on early and rapid iteration with frequent usability input' (2009, n.p).

In a study by the Nielsen Norman Group (2003), 201 usability professionals were surveyed to investigate the recruitment practices of companies that do user testing. Due to costs of recruitment services, most companies (64%) recruited their own participants. Those using a recruitment agency spent an average of $107 to recruit one participant. Companies recruiting in-house spent an average of 1.15 staff hours recruiting one participant, with 24% spending more than two staff hours to recruit one participant.

Of internal company employees recruited as participants, only 10% were paid a monetary incentive in addition to their salary, with 35% being provided a non-monetary incentive, such as a book coupon or a free lunch. In contrast, 63% of participants recruited from outside the company received a cash incentive, 41% received a non-monetary incentive, and 9% received no incentive. Thirteen percent were given both monetary and non-monetary incentives. External participants were paid an average of $64 per test hour, with an average no-show rate of 11% (Nielsen Norman Group 2003).

Paying participants is one thing but finding them is another. For realistic results, we suggest looking for users in their natural context: If a website is to be developed for a chocolate shop, visit a chocolate shop and talk with customers. If the website is for tourists, speak to people at a local tourist destination. This approach works even when the typical users of a website have a very specialised profile. For example, users of a pregnancy healthcare website can be found by contacting a local maternal health clinic. Locating potential participants is not difficult when characteristics of the typical user are considered.

When considering user participation, Visser and Visser (2006) observed that even after a four-month interval, participants' sensitivity to their study had not been reduced. They further noted that by not reusing participants, the richness of their contributions would be lost in the later development phases. However, they did not want to generalise based on one case. To the contrary, Blume (in Sova and Neilsen 2003) reports that it takes more effort to get good data when reusing participants, claiming that such participants become expert testers and lose their perspective, wanting to do the redesign and not the study tasks. Sova and Neilsen (2003) also make the distinction between 'novice' and 'expert' participants, advising that participants should generally not be used more than twice a year.

Adams, R & Atman, C 1999, 'Cognitive processes in iterative design behaviour', Frontiers in Education Conference, vol. 1, pp. 11A6/13-18.

Arnowitz, J, Arent, M, & Berger, N 2007, Effective Prototyping for Software Makers, Elsevier, San Francisco.

Blume, 2003, 'Expert testers', cited in Sova, D & Nielsen, J 2003, 234 Tips and Tricks for Recruiting Users as Participants in Usability Studies, viewed 15 February 2012, <http://www.nngroup.com/reports/tips/recruiting/>

Bury, K.F. The iterative development of usable computer interfaces, Proceedings of IFIP INTERACT'84 International Conference on Human-Computer Interaction (London, U.K., 4-7 September 1984), 743-748.

Buxton, W., and Sniderman, R. Iteration in the design of the human-computer interface. In Proceedings of the 13th Annual Meeting of the Human Factors Association of Canada, 1980, pp. 72-81.

Cooprider, J & Henderson, J 1990, 'Technology-Process Fit: Perspectives on Achieving Prototyping Effectiveness', Journal of Management Information Systems, vol. 7, no. 3, Winter, 1990/1991, pp. 67-87.

Fowler, M 2003, The New Methodology, viewed 3 January 2012, <http://www.martinfowler.com/articles/newMethodology.html>.

Hardgrave, B, Wilson, R & Eastman, K 1999, 'Toward a Contingency Model for Selecting an Information System Prototyping Strategy', Journal of Management Information Systems, vol. 16, no. 2, pp. 113-136.

Proceedings of the 4th Nordic conference on Human-computer interaction: changing roles, October 14 - 18, Oslo, Norway.

Nielsen, J 1992, The usability engineering life cycle', IEEE Computer, vol. 25, no. 3, pp. 12-22.

Nielsen, J 1993b, 'Iterative user-interface design', Computer, vol. 26, no. 11, pp. 32-41.

Nielsen, J 2012, 'Discount Usability: 20 Years', Alertbox, June 4, viewed April 24 2013, <http://www.nngroup.com/articles/how-many-test-users/>.

Nielsen Norman Group 2003, How to Recruit Participants For Usability Studies, viewed April 27 2013, <http://bit.ly/Zgi4GZ>.

Sotirovski, D 2001, 'Heuristics for iterative software development', IEEE Software, vol. 18, no. 3, pp.66-73.

Sova, D & Nielsen, J 2003, 234 Tips and Tricks for Recruiting Users as Participants in Usability Studies, viewed 15 February 2012, <http://www.nngroup.com/reports/tips/recruiting/>

Visser, F & Visser, V 2006, 'Re-using users: co-create and co-evaluate', Personal Ubiquitous Computing, vol. 10, no. 2-3, January 2006, pp. 148-152.

West, M 1986, 'A taxonomy of prototyping - tools and methods for database, decision sup- port and transaction systems', in M Lipp (ed.), Prototyping: State of the Art Report, Pergamon Motech Limited, England, p. 105-122.