Radical Design in HCI

From TEILab
Jump to navigation Jump to search

Project Team

Previous Participants

  • Yao Wang PhD student
  • Dr. Rex Hartson Emeritus Professor, Center for HCI

Purpose of the study

According to Norman and Verganti [1], there are mainly two types of innovative designs. One is the incremental design innovation, which is successfully led by human-centered design method so far. The other one is radical design innovation, which has no magic solution at this point to design a successful system. However, radical design plays a more important role than incremental design in world revolution, which can bring new things to change the way people live. Currently, few research prototypes are considered as radical designs since radical design is considered as difficult due to the lack of formal methodology. And even worse very few of them are successful. It usually took a long time for people to accept the surprising design ideas and become familiar with the new systems since it was quite different from what they used to use. Because very few research has been done to explore the methods for radical design and no users could provide requirements to this type of design, existing radical designs were all created based on designers’ intuitions and their own experiences.

As a result, the main purpose of this study is to explore at least a radical design pattern which can be used to create a successful prototype. We create a model named non-embarrassing version one(NEVO) to represent the radical design(see Figure 1), since most radical design prototypes were embarrassing at the beginning. So we would like to examine the design methods and design processes used by system builders who build innovative systems and design expertise who have lots of design experience, in order to find the similarity between the two groups and eventually provide a radical design pattern which at least can give birth to non-embarrassing version one. We are particularly interested in the two aspects of this type design, the first one is that how a designer could move from the seed idea to a fully acceptable research prototype and the second one is that how users could accept the new design and switch from what they usually used to what they are not familiar at all.

NEVO conceptual model

Figure 1.NEVO model. In any given domain, there is a space for possible designs where sits a lot of existing systems that users are familiar with, while the NEVO research prototype is different from what exists. It is designed to conduct research, like demonstrating research contribution or studying a phenomenon, rather than solving a problem or implementing a commercial product. This kind of research is framed by a seed idea such as a theory or a technique, and all of the design requirements are generated based on the concept of this idea or the research questions posed from this idea, not from user data collections. As a result, this type of research prototype is far away from existing designs, users may need to be taught how to use this type of prototypes and it might take a long time to be accepted by the users.

Study Participants

14 design researchers(four females and ten males) coming from academic research laboratory and industry research laboratory participated this study, eight of them are system builders, which was considered as group SB, and six of them are design gurus, which was considered as group D. We recruited eight system builders whose prototypes fitted in our NEVO model, from a six year sample of last decade conferences on Human Factors in Computing Systems, including CHI 2002,CHI 2003,CHI 2005,CHI 2007,CHI 2008 and CHI 2012, since the published systems are considered as acceptable. The eight systems were well distributed into different domains in Human-Computer Interaction including web, system, embodiment, eye tracking, social interaction, input device and interaction design. We recruited six design expertise from well-known design institutions all over the world. The six designers were considered as making noticeable contributions to the design area. All of the 14 researchers had some experience designing different kinds of innovative systems.


A 1-hour interview of questions and answers was conducted for each researcher. According to researchers’ preferences, 11 of them were interviewed through a commercial video teleconference tool, two of them were interviewed by telephone and one of them was interviewed face to face in the research lab.

For system builders, a semi-structured questionnaire was designed for the interview,which comprised a mixture of closed and open questions. The questions for system builders were made up of five survey questions, three rating questions and 24 open-ended questions.The first 31 questions were associated to the specific system published in CHI conference including the four sections: research description,system description, design process and methods as well as system evaluation. All of these questions are designed trying to find the effective and ineffective research methods in designing new systems they usually involve based on their empirical experiences. The last question was a general one in the design domain, which was about the meaning of design in Human Computer Interaction to see how those system builders perceive the design concept from their own perspective.

For designers, the interview was divided into two sections. Firstly, we started with the introduction of NEVO model using a powerpoint by sharing our screen.(For those who were interviewed by phone, we sent the powerpoint slides at the beginning of the interview). Then an unstructured questionnaire was designed according to the NEVO model,the questions were composed of two sets of open-ended questions. The first set of questions was related to the design process around how a designer could successfully move from the seed idea to a full research prototype. The second set of questions was about the role of user testing in this kind of research prototype. The particular questions within each set were not exactly the same as different designers because of their answers and their own experiences. The second section was to ask the designers to propose a system that fitted in this NEVO model. 20 questions(five survey questions, three rating questions and 13 open-ended questions) were asked specifically to the system they proposed, which was a subset of the 32 questions for system builders. There was one designer who couldn’t think of a system immediately, we asked the questions based on his general research interests.

Data Collection

Two voice recorders were used to record the interview sessions with researchers’ permissions. Both qualitative and quantitative data were collected from the interview questions. As a result, except for the one who did not provide the system, we had 13 sets of answers in the system builders group, and we had six sets of answers in the design group.


[1] Norman,D and Verganti,R. Incremental and radical innovation: design research versus technology and meaning change.Design Issues,2012.