Published Date September 2016, Vol.46:95–124,doi:10.1016/j.destud.2016.05.001 Open Access, Creative Commons license Author
Seda Yilmaz
College of Design, Iowa State University, Ames, IA 50010, USA
Shanna R. Daly
Colleen M. Seifert,
Richard Gonzalez
University of Michigan, Ann Arbor, MI, USA
Available online 7 June 2016.
Highlights
3457 concepts created by industrial and engineering designers are combined across four studies.
•
Observations go beyond products to consider multiple concepts generated for a given problem.
•
A master set of 77 design heuristics is extracted from a variety of design problems and designers.
•
The observed design heuristics capture ways to introduce variations into candidate concepts.
•
Design heuristics can help designers explore alternative concepts in early conceptual design.
How do product designers create multiple concepts to consider? To address this question, we combine evidence from four empirical studies of design process and outcomes, including award-winning products, multiple concepts for a project by an experienced industrial designer, and concept sets from 48 industrial and engineering designers for a single design problem. This compilation of over 3450 design process outcomes is analyzed to extract concept variations evident across design problems and solutions. The resulting set of patterns, in the form of 77 Design Heuristics, catalog how designers appear to introduce intentional variation into conceptual product designs. These heuristics provide ‘cognitive shortcuts’ that can help designers generate more, and more varied, candidate concepts to consider in the early phases of design. Keywords
A number of approaches for facilitating idea generation during the early phases of conceptual design have been proposed (c.f. Clapham, 1997, Shah et al., 2003 and Smith, 1998). One approach distills knowledge about specific designs into an intermediate-level knowledge base by constructing composites from multiple examples. In Alexander's pattern language (Alexander, Ishikawa, & Silverstein, 1977), and Krippendorf's design discourses (2005), patterns common in successful design solutions are identified at a component level, linking the designer to a broad range of helpful guidance from past solutions in a refined form (Alexander et al., 1977). This composite knowledge about design has been referred to as heuristic knowledge (Fu, Yang, & Wood, 2015). Heuristics are described as ‘mental shortcuts’ that capture cognitive strategies that may lead to solutions (though not necessarily the best one) (Nisbett & Ross, 1980), and are ubiquitous in human reasoning (Goldstein et al., 2001). Heuristics capture important features of problem situations and solutions that tend to reoccur in experiences (Clancey, 1985).
In software design, Riel (1996) has described the heuristic approach as ‘specific experience-based guidelines’ that help developers make good decisions. Lawson (1979) observed architectural students solving puzzles through ‘trial and error’ heuristic approaches. Lawson (1980) concludes, ‘An examination of protocols obtained from such closely observed design sessions reveal that most designers adopt strategies which are heuristic in nature… Heuristic strategies do not so much rely upon theoretical first principles as on experience and rules of thumb’ (p. 132). When generating new concepts, designers appear at times to offer intuitive responses derived from ‘large pools of experience’ (Cross, 2011, p. 10) to make a ‘best guess’ at a new design. Consider the example in Figure 1, a desk chair that reclines to allow the user to lie beneath (rather than in front of) a computer screen.
Figure 1. A design released by Altwork (http://altwork.com) positions the user under the workstation
In comparing this novel design to prototypical chairs, it is evident that the designer changed the user's direction of access. By moving the access point from in front of the screen to below it, an innovative design results. Further, this strategy, ‘change direction of access,’ may be a useful heuristic to apply in generating designs for other products. For example, applying the ‘change direction of access’ heuristic to a trackball controller may suggest side rather than top access, and accommodate thumb control rather than palm movements (see Figure 2). Design heuristics like this one may help designers create more, and more diverse, concepts, thereby increasing the likelihood that an innovative concept will result. Understanding how cognitive processes can be stimulated to generate design ideas may lead to more effective methods and tools to support conceptual design (Jin & Benami, 2010).
Figure 2. The original version of the Kensington Expert Mouse (www.kensington.com) used a center ball as a trackball, while the newer design by Logitech (www.logitech.com) positions the ball on the right side, under the thumb
In this paper, we examine evidence for design heuristics in the creation of multiple design concepts. First, we summarize prior research where design heuristics were derived from evidence in the field of product design, including approaches based on analysis of existing products and patents (e.g., Altshuller, 1984 and Skiles et al., 2006). Next, we compile results across four research studies to identify a distinct set of heuristics evident in a diverse sample of design solutions. These solutions include an analysis of award-winning products created by many different designers. Uniquely, the present analysis examines design concepts from a professional designer working on a single design problem. In addition, two think-aloud protocol studies of industrial and engineering designers working on a novel design problem are included. These samples add value because they include multiple concepts generated for the same design problem. By considering alternative concepts, it is possible to observe how heuristics are used in the idea generation process, and how they facilitate exploring the space of concepts for a design problem. Compiling patterns observed across varied products, design tasks, and design processes, we identify a new set of 77 design heuristics. Each heuristic is presented with a written description and an example of its application in an existing consumer product. Finally, we discuss issues of the granularity of heuristic descriptions, and the use of heuristics as a concept generation tool for product designers.
1 Heuristics in product design
How can we identify possible heuristics used in product design? Heuristics are learned from experience within a domain, and tend to be implicit and difficult to verbalize (Nisbett & Ross, 1980). The use of heuristics without conscious access has been documented in studies of experts including firefighters (Klein, 1993), scientists (Baker & Dunbar, 2000) and designers (Yilmaz & Seifert, 2011). However, this tacit knowledge about how to create designs may be observable by comparing designers' proposed solutions (Matthews et al., 2000 and Yilmaz et al., 2016). Several existing heuristic approaches to idea generation have drawn conclusions based on empirical studies of product concepts (Perez, Linsey, Tsenn, & Glier, 2011) and design patents (Altshuller, 1984).
The theory of ‘inventive problem solving’ (known as TIPS or TRIZ) (Altshuller, 1984) involved identifying heuristics from successful patents in engineering. The TRIZ analysis focuses on identifying technical contradictions in mechanical engineering designs. For example, Ogot & Okudan (2007) describe a design tradeoff when ‘increasing the stiffness of an airplane's wings to reduce vibration during flight (good) increases the weight of the plane (bad)’ (p. 111). Altshuller (1984) analyzed thousands of engineering patents and abstracted forty principles, and noted that certain contradictions lend themselves to particular solutions. These were compiled into a contradiction matrix of system features (e.g., speed, weight, measurement accuracy) crossed with typical undesired results to index relevant design principles (Altshuller and Rodman, 1999, Altshuller, 1997, Altshuller, 2005, Orloff, 2003 and Savransky, 2000). However, because TRIZ analysis requires the identification of technical tradeoffs first, it is most helpful for designs developed to the point of specific commitments to materials and mechanisms.
Learning to use the TRIZ system requires extensive training, effort and commitment (Ilevbare, Probert, & Phaal, 2013). The terminology and modeling methods are unique to TRIZ, and differ from those found in engineering design (Smith, 2003). However, in a classroom study with first-year engineering students, Ogot and Okudan (2007)trained teams of 4 students to use TRIZ to generate concepts while other teams used traditional idea generation methods. They found that teams using the TRIZ method produced more unique solutions compared to other teams, along with more feasible concepts. This was replicated in another engineering classroom study where the TRIZ method was found to result in more novelty compared to sketch methods. In a third classroom study, engineering students using TRIZ improved the novelty and variety of concepts generated (Hernandez et al., 2013 and Hernandez et al., 2014). Finally, an experimental study with graduate student and professional engineer teams found that TRIZ improved the novelty of solutions with only a ten minute training session (Chulvi, Gonzalez-Cruz, Mulet, & Aguilar-Zambrano, 2013).
Another approach to identifying design heuristics has examined existing products that ‘transform,’ or change into different configurations or states for use (Skiles et al., 2006). For example, a wooden chair may be designed to transform into a stepladder. Transformer products address each function set independently and at different times, while moving smoothly between states as needed (Weaver, Wood, Crawford, & Jensen, 2010). Based on analyses of 85 past patents, 40 analogies from nature, and 100 existing multistate products, three transformation design principles were extracted (expand/collapse, expose/cover, and fuse/divide) (Singh et al., 2007, Singh et al., 2009, Skiles et al., 2006, Weaver et al., 2008 and Weaver et al., 2010). A fourth principle, reorientation, was proposed in a later study (Haldaman & Parkinson, 2010). In addition, twenty subordinate ‘facilitators’ were extracted to support these principles. Example facilitators include using ‘generic connections’ to allow different modules to perform different functions; ‘segmentation,’ or dividing a single contiguous part into two or more parts; and ‘fold,’ or create relative motion between parts or surfaces by hinging, bending, or creasing. A study of engineering students found that encouraging the use of transformation principles and facilitators resulted in the generation of 25% more concepts (Weaver et al., 2009).
Several other studies have analyzed product designs to derive heuristics for idea generation. One study examined 197 award-winning innovative products, and organized the identified design features into categories (Saunders, Seepersad, & Hölttä-Otto, 2011). The thirteen ‘innovation characteristics’ identified in this analysis include ‘additional function,’ ‘modified size,’ ‘expanded usage environment,’ and ‘user interactions.’ Another study identified ‘consumer variation’ heuristics for designing for user differences (Cormier, Literman, & Lewis, 2011). Through an analysis of 31 product lines with 645 product models, 20 heuristics are identified and categorized into function, form, and information and control groups. Examples include, Utilize (re)configurability when the product architecture is specific to handedness, Use system (re)configurability facilitated by modules when desired functionality is decoupled, and Utilize materials which have built-in flexibility for aesthetic modification. Finally, a study of 46 bio-inspired products and systems resulted in six ‘scaling principles:’ change energy source, simplify system, change method, combine functions, directly transfer components, and change parameters (Perez et al., 2011).
In these different approaches, various design heuristics were identified based on the design evidence considered. These approaches differ in the observed designs, with a focus on transforming (dual function) products in Weaver et al. (2010), award-winning innovative products in Saunders et al. (2011), consumer variation product lines in Cormier et al. (2011), and products at varied scales (in Perez et al., 2011). TRIZ (Altshuller, 2005) stands out for the large number of patents analyzed. However, in all of these approaches, only a final ‘winning’ concept is considered. The present study also includes a large sample of designs for award-winning consumer products. But uniquely, the present study adds samples of multiple candidate concepts generated by designers for a single design problem. The opportunity to observe the set of candidate concepts generated by a designer for a given problem provides a richer sample of variations among concepts than is captured by final product designs. Observations from a long-term design project by a very experienced designer added hundreds of concepts for a single design problem. The observation of idea generation sessions (rather than solely the ‘winning,’ final product) provides more evidence about how designers introduce variations in their concept sets through what Lawson (2012) calls ‘knowing by doing.’ By consolidating results across four empirical studies of concept generation, with varied contexts and more concepts sampled, we hoped to detect a broad array of design heuristics.
2 Method
For the present study, we compiled a larger database from four prior empirical studies (described in Table 1). The goal was to create a larger, rich dataset of design concepts from three different contexts, multiple design problems and multiple designers. The four studies included diverse datasets: (1) award-winning products from a wide range of consumer domains, (2) an expert industrial designer's sequential concept sketches from a two-year solo design project, and (3) a protocol study of engineering designers where student and practicing designers' think-aloud protocols were recorded as they worked on a novel product design task. A fourth study (4) replicated the think-aloud protocol study with industrial designers in order to compare concepts from the two design disciplines.
Table 1. Separate empirical studies of design concepts included in the cumulative database
Study
Research question
Data collection
Source
Study 1. Product Analysis
What are the strategies that successful designers use to create novel products?
400 award-winning products from a diverse range of design domains.
How does an experienced designer add variation to concepts within a single long-term design problem?
218 sequential concepts created by an expert industrial designer over two years for a single design project (a universal access bath within an existing home).
How do different designers create concepts within a single novel design task?
Think-aloud protocols from 36 engineers at varying levels of expertise as they designed a novel product (a portable solar oven) in a 25-min session, with a total of 179 concepts generated.
How does Design Heuristic use differ among designers from different design disciplines?
Think-aloud protocols from 12 industrial designers at varying levels of expertise working with the problem (in Study 3) for a total of 68 concepts generated.
The process for extracting a design heuristic from award-winning product was as follows: For observed design concepts, major elements and key features of each concept were analyzed for functionality, form, and user-interaction features. A content analysis of the needs, design criteria, functions, and the design solution was performed for each concept. Then, potential heuristics were hypothesized and design criteria for their application were identified. Other concepts in the dataset with the same design features were compared in order to explore commonalities in candidate heuristics. Finally, a heuristic would be defined at a level of generality that applied to multiple products, but was still specific to the observed design solution. For example, one heuristic was described as the ‘hollowing out’ of material, such as a brush handle with its mass reduced by using a hollow cylinder for a handle. This kept the heuristic's description as close as possible to the observed concepts; for example, different heuristics captured reducing material through flattening or folding. This extraction approach catalogs more specific innovations while ensuring the heuristics are general enough to fit several different observed concepts. Singh and colleagues (2009) describe a similar extraction method in their analysis of transforming products.
The product images in Figure 3 illustrate the process of extracting a heuristic from two of the 400 award-winning products included in the study. The first image shows a new product – a paint roller – where a commonly used mechanism in ballpoint pens (the ink storage and roller) is applied in a new context to solve the problem of delivering wall paint touchups. This heuristic also appears in the second image as a brush repurposed as a desk organizer design. The heuristics extracted identify independent components of the design, and are not exhaustive, such that other features of these designs might serve to identify other possible heuristics. In the first image, a second heuristic is also observable; namely, Synthesize Functions, where both paint storage and applicator are combined in the design. In this way, observed concepts sometimes provided evidence of multiple heuristics.
This extraction method for identifying design heuristics in existing products was applied to the design concepts in the remaining three studies (Daly, Christian et al., 2012, Daly, Yilmaz et al., 2012, Yilmaz and Seifert, 2011 and Yilmaz et al., 2016). Study 2 provided 218 concepts created by a single, very experienced industrial designer over a two-year period (Yilmaz & Seifert, 2011). The design problem was to create a universal access bathroom to be installed in private homes. The designer worked on a large paper scroll to preserve his concepts as they were created. By examining sequential concepts, transitions between candidate concepts were evident. Across this set of designs, we observed that the same specific heuristics appeared repeatedly in this designer's work. For example, one heuristic addressed a change in how the functions of the product were controlled. In this example concept, the designer arranged components around the same central structure (a plumbing tube) (see Figure 4). This strategy was then observed in other designs, leading to a proposed heuristic, Align components around the center. This concept also suggests other heuristics, allowing the user to reorient the product according to their height, and repeat design elements.
Figure 4. Example concept combining the heuristics Align components around center, Allow user to reorient, and Repeat. (Courtesy of Allen Samuels, Industrial Designer.)
The concepts collected from Studies 3 and 4 involved a ‘think aloud’ protocol (Dorst and Cross, 2001 and Ericsson and Simon, 1993) of engineering and industrial designers' process while creating solutions for a novel product problem (the design of a solar oven for use in an outdoor setting). Forty-eight designers generated 247 different concepts for this single design problem. For example, one of the designers generated a concept for a portable backpack container that allowed cooking using sunlight (see Figure 5).
Figure 5. A concept for a solar oven generated by a designer using an Attach product to userheuristic, along with an Add functions heuristic. The industrial designer described a context in which the user was a hiker, and designed an integrated backpack with a heating element and pot attached to it. This would allow the user to warm food throughout the day while traveling
Next, three independent coders with advanced degrees (one with an M.F.A. in industrial design, one with a Ph.D. in engineering education, and one a senior student in mechanical engineering) worked as a team to examine each concept in the collected database. The coders considered each concept both individually and in its concept set sequence for evidence of heuristic use. The three coders worked collaboratively to refine heuristic definitions, and all decisions about identified heuristics were argued to consensus. Because the coders worked as a team during the extensive analysis, no measure of reliability was possible. The collaborative identification of heuristic use across these observed concepts occurred over a period of six weeks.
3 Results
The analysis of this combined sample of 3457 products and design concepts across four empirical studies resulted in the observation of 77 distinct design heuristics. Each of the identified heuristics was observed in at least four different concepts across the sample datasets. These heuristics addressed design goals such as adding functionality, using fewer resources, saving space, providing visual consistency, and forming new relationships among design elements. The 77 Design Heuristics are shown in Figure 6. This set of 77 Design Heuristics includes only those necessary to account for the data in these four studies. Each Design Heuristic is described, and illustrated with a commercial product where the heuristic is evident.
Figure 6. The 77 Design Heuristics identified across four studies of award-winning product designs, a solo professional design project, and protocol studies of engineers and industrial designers working on a novel problem. Each is illustrated with a description and an example consumer product where the Design Heuristic is evident. (Courtesy of Design Heuristics, Inc.)
The observations supporting this set of 77 Design Heuristics (capitalized when referring to heuristics from this set) are shown in Table 2. An important feature of this compilation of heuristics across studies is that each heuristic was observed multiple times (at least four) in different products and product concepts, and all were observed in solutions from more than one designer. The sole exception is expose interior, which was observed only one concept (in Study 4) but included because it is well known (e.g., watches or clocks) and may facilitate the goal of considering a variety of candidate concepts.
Table 2. Observations of heuristics observed across Studies 1–4, presented in alphabetical order. Seven heuristics originated from subdividing other observed heuristics
For further details log on website
http://www.sciencedirect.com/science/article/pii/S0142694X16300254
Published Date November 2016, Vol.47:47–72,doi:10.1016/j.destud.2016.08.003 Author
Elizabeth Starkey
Christine A. Toh
Department of Industrial & Manufacturing Engineering, Pennsylvania State University, University Park, PA 16802, USA
Scarlett R. Miller,
School of Engineering Design, Technology, and Professional Programs, Pennsylvania State University, University Park, PA 16803, USA
Available online 12 September 2016.
Highlights
Creativity in engineering education is impacted by the design task being explored.
•
Students discard novel ideas during concept selection for conventional alternatives.
•
Creativity during generation and selection does not predict final design creativity.
Creativity training has been widely integrated into engineering education as a means to prepare students to be an innovative force in design industry. However, much of this research has focused on training students to be creative idea generators, with limited attention to what happens after this generation. Thus, the current study was developed to understand how creative ideas are promoted or filtered throughout the design process in order to focus our educational efforts. In order to accomplish this, an 8-week study with 136 engineering students was conducted. Our results point to the reduction in creativity throughout the design process and student abandonment of novel concepts. We also expose the influence of the design task on student creativity.
Published Date Available online 9 November 2016,doi:10.1016/j.destud.2016.10.001 In Press, Corrected Proof —Note to users Author
Philip J. Cash,
Charlotte Gram Hartlev
Department of Management Engineering, Technical University of Denmark, DK-2800 Lyngby, Denmark
Christine Boysen Durazo
Department of Management Engineering, Technical University of Denmark, DK-2800 Lyngby, Denmark
KL.7, Nørregade 6, 1165 København K, Denmark
Available online 9 November 2016.
A case study surveying 20 behavioural design projects is reported.
•
Significant patterns in process progression are found in the surveyed projects.
•
A new design process is proposed that integrates behaviour change and design.
•
Key learnings for behaviour change through design are identified.
Nudge, persuasion, and the influencing of human behaviour through design are increasingly important topics in design research and in the wider public consciousness. However, current theoretical approaches to behaviour change have yet to be operationalized this in design process support. Specifically, there are few empirically grounded processes supporting designers in realising behaviour change projects. In response to this, 20 design projects from a case company are analysed in order to distil a core process for behavioural design. Results show a number of process stages and activities associated with project success, pointing to a new perspective on the traditional design process, and allowing designers to integrate key insights from behaviour change theory. Using this foundation we propose the Behavioural Design process.
Mule deer bucks move across Malheur National Wildlife Refuge, a wetland oasis in dry southeastern Oregon. Deer share the refuge seasonally with grazing livestock and hundreds of species of migrating birds. President Theodore Roosevelt, an avid hunter, established the refuge in 1908 to revive North American bird populations hard hit by plume hunters pursuing breeding feathers for fancy hats. Armed protesters occupied the refuge in January 2016, escalating tensions over the ownership and management of public lands in the United States. Credit, Barbara Wheeler/USFWSCC BY 2.0.
Americans are deeply divided about the proper uses of federal and private lands and the goods and services they supply us. Recent events, including the acquittal of the occupiers of the Malheur National Wildlife Refuge, protests of the proposed Dakota Access Pipeline, and the potential reopening of debates about the Keystone XL Pipeline, illustrate these value differences.
Scientists seek to bridge the divide with data and analysis. Today in Science magazine, U.S. Geological Survey ecologist Stephen Jackson and colleagues call for a national assessment of ecosystem trends. Citing examples of environmental change in which a national assessment could inform responses, such as drought in southwestern states, decline of native bees in Hawaii, and Gulf Coast restoration, the authors propose a public-private partnership, “between NGOs, philanthropic foundations, private-sector entities, agencies of the U.S. Government, and selected state, local and tribal agencies” to assess the status of the nation’s biodiversity and ecosystem services.
In an editorial in BioScience, Osvaldo Sala (2016) proposes a role for scientific measurement in resolving conflicts over land use, stating that
“Measuring ecosystem services and the demands from stakeholders may assist in conflict resolution by shifting the discussion from an emotion-based conflict to an evidence-based negotiation in which many novel options are in play. Then, it would be possible to assess whether there are trade-offs among the demands of different stakeholders.”
This optimistic view of the contributions of ecosystem service measurement to decision-making is echoed by Matthias Schröter et al. (2016):
“In order to safeguard biodiversity and the services that ecosystems provide to humans, national ecosystem assessments (NEAs) form an essential knowledge base to inform decisionmaking in policy and practice.”
Effective monitoring of biodiversity and ecosystem services is also key to meeting national conservation goals (Aycrigg et al. 2016) and responding to the continuing global impacts of climate change (Scheffers et al. 2016).
In response, Jackson et al. (2016) call for an ongoing national assessment of ecosystem functioning. They argue that
“A systematic national assessment program is needed to identify potential risks, determine priorities, identify trends, and evaluate decisions and actions . . . without the context and reference points provided by a comprehensive national assessment, it will be difficult to assess benefits of government decisions aimed at enhancing or protecting ecosystem services.”
Jackson et al., all members of ESA’s Steering Committee for the Intergovernmental Platform on Biodiversity and Ecosystem Services (IPBES), further note the U.S. government played a leading role in creating IPBES, an international organization that is conducting global and regional assessments of biodiversity and ecosystem services. Many U.S. scientists are contributing to these ongoing studies as well as to a global study of pollination services completed this year. While these assessments are important, they are at too large a scale to substitute for a national assessment. For example, the Americas regional assessment covers all of North, Central, and South America and the Caribbean. A national assessment would provide focused information of direct relevance to national decision makers and would reinforce U.S. leadership in environmental stewardship.
The election results last week may affect the direction and focus of U.S. environmental policy, but they do not alter the need for the United States to responsibly manage its public lands and natural resources. Now more than ever we need to place as much importance on monitoring biodiversity and ecosystem services as we do on monitoring the economic and public health of the nation.
References
J.L. Aycrigg, C. Groves, J.A. Hilty, J.M. Scott, P. Beier, D.A. Boyce, Jr., D. Figg, H. Hamilton, G. Machlis, K. Muller, K.V. Rosenberg, R.M. Sauvajot, M. Shaffer, and R. Wentworth. 2016. Completing the system: opportunities and challenges for a national habitat conservation system. BioScience 66:774-784. doi: 10.1093/biosci/biw090
Jackson, S.T., C.S. Duke, S.E. Hampton, K.L. Jacobs, L.N. Joppa, K.-A. S. Kassam, H.A. Mooney, L.A. Ogden, M. Ruckelshaus, and J.F. Shogren. 2016. Toward a national, sustained U.S. ecosystem assessment. Science 354: 838-837. DOI: 10.1126/science.aah5750
Sala, O. 2016. How scientists can help end the land-use conflict. BioScience 66:915. doi: 10.1093/biosci/biw145
Scheffers, B.R., L. De Meester, T.C.L. Bridge, A.A. Hoffmann, J.M. Pandolfi, R.T. Corlett, S.H.M. Butchart, P. Pearce-Kelly, K.M. Kovacs, D. Dudgeon, M. Pacifici, C. Rondinini, W.B. Foden, T.G. Martin, C. Mora, D. Bickford, and J.E.M. Watson. 2016. The broad footprint of climate change from genes to biomes to people. Science 354: 719. DOI: 10.1126/science.aaf7671
For further details log on website :
http://www.esa.org/esablog/guest-posts/usa-needs-national-ecosystem-assessment/
