The history and evaluation of design thinking as a product design research methodology

The discourse on the importance of separating scientific and design methods has existed since the 1960s, but design thinking gained popularity in the late 2000s, alongside theories proposed by IDEO and the Stanford Design School, in the context of leveraging the competitive advantage of innovation in Silicon Valley. Since then, design thinking has been applied to other fields outside design. However, over time, the effectiveness of designers came to be questioned. To untangle the confusion, this paper explains the history of design thinking and classifies existing methodologies into methodologies based on: (1) humans; (2) prototypes; (3) combination; (4) new product development, then evaluates them based on their implementation in students’ final projects. In choosing the right methodology, designers must consider: (1) the greater weight between problem or concept exploration; (2) the main focus of innovation lenses; and (3) theoretical approaches that can leverage the design’s added value.

This is the English translation of an article published in “Jurnal Desain Indonesia” https://doi.org/10.52265/jdi.v7i02.367

INTRODUCTION

The term design thinking (DT) is often associated with the design agency IDEO, which in the late 2000s succeeded in popularising a method for applying designers’ working patterns to fields outside design. For example, DT identified a problem that had not previously been considered during the transformation of hospital services, namely, ineffective nurse shift handovers [1].

Although discussions concerning the differences between design research methodology and scientific research had existed since 40 years earlier [2; 3], the framework formulated by IDEO [4] was highly applicable, so its dissemination and uptake were relatively rapid [5; 6]. To date, DT remains a popular method and is widely taught in various training sessions to convey that it is not only useful for creative workers but also for all members of society who wish to solve problems [7].

In line with the rapid development of DT, several institutions have developed new frameworks, including the Stanford Design School [8; 9] and the Design Council [10; 11]. Thus, when someone says that they use DT in the process of creating a design work, additional questions arise: Which DT method? Should designers not already be applying this method automatically? Why do designers still have to think like designers?

Not infrequently, DT, which was originally intended to help understand something abstract, is instead seen as a fixed system that is then forced to fit the research objectives [12]. As a result, criticisms of the DT method have recently begun to emerge. In addition, confusion sometimes occurs because this method overlaps with other methods, such as lean and agile [13], not to mention the many terms that have emerged, such as sprint, kanban, and backlog.

This paper aims to disentangle the abundance of information that can cause confusion by examining the emergence and development of DT over time, including the pros and cons that have arisen. The various existing methodologies will be presented from a more comprehensive perspective through clustering.

In general, these methodologies can be grouped into three main approaches, namely: human, prototype, and a combination of the two. The engineering science approach in the series of new product development (NPD) [14; 15] is also used as a reference to illustrate the importance of a research & development (R&D)-based approach in addition to the human-based approach that is more often highlighted at present.

Rather than critiquing DT, this paper will evaluate the mistakes designers may make when applying it, leading to ineffective product development. The results of this evaluation are expected to serve as a reference for developing a more optimal design research methodology framework and to help critically reinterpret the DT process.

METHODOLOGY

The research was conducted through a literature review, focusing on historical and comparative analyses of the various existing design methodologies. This process was followed by clustering approaches that can serve as a reference for selecting the most appropriate design research methodology.

To analyse the variety of research methodologies in greater depth and evaluate their effectiveness in the process of creating design works based on problem-solving, content analysis was conducted on 18 final project reports by Product Design Innovation (PDI) students at Universitas Prasetiya Mulya (UPM) completed in 2023–2024. The products produced had varied problem backgrounds and final solutions, ranging from furniture and applications to assistive technology. In the reports, the methodology was described to ensure that the students’ design process not only prioritised creativity but also carried research weight and offered novelty.

RESULT

Scientific method and design method

The initial idea of describing the design process in the context of scholarship or design science was first put forward by Buckminster Fuller in 1963 [16], and later explored by S. A. Gregory at the conference “The Design Method” in 1965 [17; 18]. This conference was followed by the publication of a book with the same title, a compilation of writings by several authors discussing variations in design approaches across human, technological, social, and material aspects.

The urgency of the discourse on design as a field of knowledge stemmed from the fundamental differences between science-based research—particularly that related to the exact sciences—and design-based research, as illustrated in Table 1.

	Source	Scientific Method	Design Method
1	Gregory, 1967 [19]	Analytic	Constructive
2	Gregory, 1967 [19]	Understanding and describing existing patterns	Creating non-existent things
3	Nadler, 1967 [3]	Inductive	Deductive
4	Nadler, 1967 [3]	Tend to be linear	Tend to be iterative
5	Simon, 1969 [20]	How things are	How things ought to be
6	Cross et al., 1981 [21]	Problem → Observation	Observation → Problem
7		Answering “what” as an absolute truth	Answering “how” a non-absolute truth
8		Seeking to disprove a theory	Seeking to confirm a solution
9	Brown, 2010 [6]	Producing knowledge by proving hypotheses and validating theory	Producing applicable and usable insights

The most striking difference is the placement of the problem in the research flow. In the scientific method, which is analytical and inductive in nature, the main aim is to understand and describe existing patterns as they are, so the problem-definition process precedes observation. Science also recognises hypotheses that must be proven either absolutely true or absolutely false. This definite research flow produces a process that tends to be linear.

Meanwhile, in the design method, observation is carried out first, so its nature is constructive and deductive. Problem definition can be carried out after a fairly long process of information gathering, so that the researcher can find the most appropriate solution to the problem and produce something that does not yet exist. The character of appropriate and less appropriate differs from that of absolutely true and false, and thus, this process tends to be more iterative, producing a more complex research flow. Indeed, design methodology is closer to engineering methodology than to science (Simon, 1988), given that both aim to create something new and tangible.

Although no specific framework had yet been found to describe design methodology, the growth of design scholarship accelerated in the 1980s [21] with the publication of design journals such as “Design Studies” in 1979 and “Design Issues” in 1984. Another influential event was the emergence of the need for objective and rational solutions after the Second World War [22]. Design was no longer synonymous with beauty alone. Function became the main consideration in creating design works based on real problems, which became the forerunner of DT.

The Birth and Critique of Design Thinking

Initially, the main target of DT was not only product ideation, but problem-solving in general. In this context, what DT offered was that a designer’s mindset could serve as “fresh eyes” and a catalyst for generating new ideas not commonly thought of within that scope. Designers also do not work alone; they invite workers to become “designers” together, thereby indicating active participation.

Although Peter Rowe introduced the concept of DT in 1987 [23], the method reached popularity when Tim Brown [5] successfully applied it in hospitals and Steven Bell [1] in libraries in 2008. Both were areas that had previously rarely been touched by design, apart from the design of the objects that filled the spaces, such as furniture and medical equipment.

It turned out that the mindset of designers—which seeks to understand the abstract, is empathetic, collaborative, and oriented towards finding as many creative solutions as possible—could be used to solve an organisation’s problems [12]. Design no longer had to be based on formed objects (tangible objects), but on solutions that could even be formless (intangible), in the form of new systems, services, and experiences. For example, IDEO succeeded in redesigning the school cafeteria system [24] and applying DT at the governmental scale to create public-friendly urban planning [25].

Bell [1] also emphasised the importance of the problem-definition process before proceeding to the problem-solution process. Indeed, the problem-definition process often takes longer than the design process. This is in line with Rittel & Weber [26; 27] and Buchanan [28], who stated that the design method is highly appropriate for solving wicked problems, namely, complex problems that lack a fixed formulation. Accordingly, there is no single correct solution; rather, the most appropriate one. Therefore, the stage of exploring the problem should have greater weight.

Meanwhile, to evaluate the feasibility of problem solutions, Brown [5] proposed the theory of the innovation lens, whereby innovation is regarded as an intersection of the achievement of three aspects, namely: (1) desirability from the user’s perspective; (2) feasibility from the technological perspective; and (3) viability from the business or financial success perspective. At present, a fourth aspect is often added to the diagram: sustainability [29].

The popularity of DT soared further with Silicon Valley’s development. The demand to innovate in accordance with user needs (demand pull) to accelerate the penetration of technology into society has become increasingly high [30]. Two prominent design agencies, Frog Design and IDEO, greatly assisted various global companies, thereby strengthening designers’ role as a pillar of innovation in creating competitive advantage amid intense market competition [31].

However, after several years, criticism of DT began to emerge as its popularity was not accompanied by a deep understanding of the theory’s foundations. Indeed, Bruce Nussbaum [32] explicitly referred to it as a failed experiment that instead reduced the process of creating work to a rigid, closed methodology. Many companies adopted DT to solve problems that were too simple, such as improving product features or even to justify an executive decision. This is contrary to the creative intelligence that designers should possess.

A similar comment was made by Don Norman [33], who regarded DT as a myth no different from creative thinking. Instead, he argued that it led practitioners to spend more time thinking than creating (design doing). In other words, these companies were considered to be using the wrong tool. The benefits of DT can, in fact, be felt most fully outside the design team itself [6; 34].

DT is also considered to minimise the role of intensive R&D and technology-related research because it is overly market-oriented and often serves only to fulfil commercial interests [35; 36; 12]. Data collection focuses on users rather than experts, even though users may not understand the problems or be able to provide realistic, immediately implementable solutions [24].

Unfortunately, many design students still cannot think like designers, even though this should arise naturally [37]. Reflecting on this, Don Norman revised his criticism. He stated that DT is still needed, on the condition that it returns to its main purpose: finding as many alternative solutions as possible rather than seeking justification for a solution. Now, the relevance of DT remains in question following IDEO’s reputational decline after it dismissed the majority of its employees at the end of 2023 [38].

Human-Based Methodology (Appropriateness)

DT is an empowering process because it is based on togetherness and user involvement [5; 6]. Its process is considered close to the ethnographic methodology commonly used by anthropologists [36], because researchers likewise go directly into the field and define problems in a bottom-up manner. Nevertheless, in DT, the process is condensed (condensed ethnography) with the main aim of achieving the appropriateness of the solution.

There are several popular human-based methodologies, as shown in Table 2. IDEO itself has two methods, namely DeepDive [1] and the more popular three-stage sequence: (1) inspiration, (2) implementation, and (3) ideation [5].

	IDEO’s Deep Dive [1]	IDEO’s Three Stages [5]	Stanford Design School [8]	Design Sprint [40]	Design Sprint [41]	Double Diamond [10; 11]
1	Understand	Inspiration	Empathize	Understand	Map	Discover
2	Observe	Implementation	Define	Diverge	Sketch	Define
3	Visualize	Ideation	Ideate	Converge	Decide	Develop
4	Evaluate		Prototype	Prototype	Prototype	Deliver
5	Implement		Test	Test	Test

Beyond this, the mindset applied during data collection, such as the “beginner’s mind” [5], namely having no assumptions and behaving as though one knows nothing, is the main idea that underpins DT, contrasting the old approach that assumes the company must know more than users. Another technique is framing problems using the “how might we…” framework, which is considered capable of sparking the spirit of creation and collaboration because of its open-ended nature [4].

One of the founders of IDEO, David Kelley, is also the founder of the Stanford Design School. Therefore, it is not surprising that another fairly popular DT theory comes from the university. This theory divides DT into five stages: (1) empathise, (2) define, (3) ideate, (3) prototype, and (4) test [8; 9]. Here, the importance of creating a prototype stage to test the resulting solution is emphasised. In addition, DT also emphasises the importance of data visualisation in the brainstorming process [39], leading DT to be synonymous with the use of canvases such as the empathy map, user journey map, and persona.

The Stanford theory was then adapted again into a rapid process called sprint [40; 41]. This approach was initially often used by Google to achieve maximum results in a rapid ideation process. The DT process, which should be carried out over a long period, was increasingly condensed into just a few days. This makes the sprint process more suitable for solving company problems internally, where employees act both as sources and as innovators [42].

A somewhat different method is put forward in the double diamond (DD) theory from the Design Council [10; 11]. DD returns the DT process to its roots, which consist of two main parts, namely problem definition and problem solution, represented by two diamonds. The diamond was chosen to depict the divergent/broadening and convergent/narrowing processes that must be carried out alternately. The first diamond on the left consists of the discover and define stages, with the final result being a problem statement. In contrast, the second diamond on the right consists of the develop and deliver stages, with the final result being a prototype solution. It should be noted that this does not mean that designers make only one prototype. The process carried out in the second diamond must be iterative.

The popularity of the DD method has led several researchers to develop it into the triple diamond method [43; 44]. The process remains the same, except that the three milestones are considered to better represent the design process undertaken. For example, the three endpoints of each diamond may contain the problem definition, initial prototype, and final prototype.

Prototype-Based Methodology (Speed)

Prototype-based methodologies include lean and agile methodologies, which are, in fact, not quite accurate to be considered as DT methodologies, as they have separate histories. The main idea of DT is problem-solving through an intensive and comprehensive process of understanding humans, whereas this method emphasises iteration and rapid prototyping.

Inspired by the software development process, which requires an approach different from the manufacturing process of other physical products, 17 practitioners gathered in 2001 to declare the “agile manifesto”, consisting of 12 principles considered able to resolve challenges in software development [45]. One of the main challenges is projects that are highly time-sensitive and frequently changing. This translates into agile principles that encourage designers to be agile, avoid wasting time, and remain open to the possibility of iteration, even at the final stage [45]. At first glance, this already appears to contradict the DT stage, which focuses more on exploring problems before finally reaching the prototype stage. Empathy and condensed ethnography are no longer the main objectives of this methodology; rather, rapid prototyping is.

The agile approach has several techniques, such as lean development, eXtreme programming, crystal methods [46], and the most popular one, which has been widely adapted into a book in its own right, SCRUM [47; 48]. This process was introduced by Jeff Sutherland at IBM in 1995. Although this idea originated in America, SCRUM was inspired by the success of lean product development in Japanese companies [49] such as Toyota [50] and Honda [51].

SCRUM leaves behind the product development process with a massive and detailed final objective and replaces it with smaller objectives, making the design process more flexible. The main overview of the project remains important as a boundary, but ample room is provided for change. These smaller objectives are integrated into the sprint backlog and product backlog [48]. This process can be visualised and integrated into a kanban board [52].

Although the agile manifesto does not formulate a product creation flow and rather offers a working approach, its main ideas have been widely adapted into other methods, such as lean startup [53], lean UX [54], and lean product [55], in line with the growth of the digital application industry. Eric Reis [53] is one of the pioneers who helped popularise the rapid Build-Measure-Learn iteration concept and convey the importance of creating a minimum viable product (MVP) for idea visualisation. For example, in application development, designers do not have to do coding straight away. Prototype testing can be carried out from the outset in a simpler way, such as by making paper wireframes. Agile excels in product development with high interaction and complex interfaces because, in the DT process, prototype creation only occurs at the very end. Thus, the process of reaching a prototype is slower, and if it is accelerated through a sprint system, it will not be in-depth [13

	Lean Startup [53]	Lean UX [54]	Lean Product [55]
1	Build	Declare assumptions	Determine target customers
2	Measure	Create an MVP	Identify underserved needs
3	Learn	Run an experiment	Define value proposition
4		Feedback and research	Specify MVP feature set
5			Create MVP prototype
6			Test MVP

Both human-based and prototype-based methodologies have their own advantages, so in practice, several designers combine the two approaches into a single methodology [56; 57; 58; 59]. For example, Zorzetti et al. [59] apply a software engineering process that uses the DD framework but fills each diamond with the build-measure-learn loop from the Lean Startup method. In addition to the two loops inside the two diamonds, the designer also adds four other loops to illustrate the work carried out by other teams.

New Product Development Methodology

The closeness of the design field to engineering became increasingly apparent with the development of the electronics industry. Thus, the book Product Design & Development [15; 16] emerged to bridge the two fields, with material and case studies centred on the development of highly complex tools, such as bicycle forklifts, printers, and car seat belts.

New product development (NPD) is a technology-based R&D process that can balance DT, which is often considered overly user-focused. In this approach, interviews are often conducted with sales consultants, distributors, and repair shops. Exploring expert opinions is considered important because users sometimes do not understand the complexity of technology-based products, so they cannot provide sharp ideas.

The product development process begins with creating a mission statement that will serve as the basis for the concept. The interview stage is conducted at the beginning, followed by the creation of measurable product specifications, benchmarking of similar products, and concept testing, also measurable with a scoring system, before finally moving on to the creation and testing of prototypes.

	Product Development [15; 16]	Concept Development [15; 16]
0		Mission statement
1	Planning (mission approval)	Identify customer needs (needs statement)
2	Concept development (concept review)	Establish target specifications
3	System-level design (specifications review)	Generate product concepts
4	Detail design (critical design review)	Select product concepts (concept scoring)
5	Testing & refinement (production approval)	Test product concepts
6	Production ramp-up	Set final specifications
7		Plan downstream development (development plan)

In other words, the process focuses on concept exploration rather than problem exploration, given its use within companies. This technique reflects the reality of the field, where designers do not always create 100% new products when market demand, brand identity, and production capacity are already clear. Problem exploration usually involves improving the quality of existing products rather than pursuing something radically new, and the study’s main focus is identifying trade-offs. For example, the more sophisticated a laptop is, the more difficult it will be to maintain a portable design, so designers must choose which needs to be prioritised. Another example is the adjustment between material quality and an appropriate selling price.

DISCUSSIONS: APPLICATION OF METHODOLOGY IN PRODUCT DESIGN RESEARCH

To understand the use of DT as a design research methodology, the analysis compared 18 processes for producing works and final project reports by PDI UPM students. Table 5 summarises the students’ initials, research themes, the methodologies and approaches used, and the main focus classified based on three innovation lenses, namely: (1) human, (2) business, and (3) technology, although several projects also have a fairly high sustainability weighting.

To access Table 5, please download the manuscript in the link at the end of this page.

In practice, each project certainly has all four of these aspects, but the emphasis on each varies. There are several projects that at first glance appear to be heavy on technology, such as the enclosure design for an AC-saving device and a foot pedal for a computer. Still, these projects place greater emphasis on the user experience when operating the product. Beyond that, the main focus of technology can be further broken down into product assembly, materials, and computation.

In general, the final project work can represent the various methodologies mentioned in the previous sub-chapter, ranging from IDEO DT, Stanford, DD, lean startup, lean product, lean UX, and NPD. Indeed, some also apply a combined methodology. Meanwhile, independent refers to a methodology that is completely new and specially prepared for the final project.

To complement the methodology, there are several approaches used in accordance with the product being developed, namely: design for assembly [60], design for manufacturing [61], universal design [62; 63], accessible design [64], interaction design [65; 66; 67], IMO [68], and ATUMICS [69; 70]. The selection of an approach is no less important, as it serves as a conceptual framework that defines the research angle and will therefore become the main source for the literature study. At present, the most widely used approach focuses on interaction and user experience.

Based on Table 5, DD is the most frequently chosen method. The reason is familiarity. Although students are introduced to all DT methodologies, DD is used as the syllabus framework in the Creative & Design Thinking course in semester 2, which the author taught. In fact, several topics would be better served by a method based on rapid prototyping.

So far, no student has comprehensively analysed an existing product in a precedent study. AHT did compare several car-cover products, but this was limited to desk research, and he did not conduct testing to measure differences in the usability of existing products. This approach can be taken if the product to be created is already widely available on the market, with a clear target market as well, so the more important discovery is the gap in existing products rather than the user gap. Other products whose development would be more optimal with this approach are working pods and furniture with special functions.

In another case, ASY found extremely abundant data from the sources, namely, people with chronic illnesses. The needs conveyed were also too varied and difficult to narrow down. In fact, if ASY had compared several furniture products specifically for sick people, looked for shortcomings, and observed how the sources interacted with the products, the potential problems might have been found more quickly.

Thus, when selecting a research methodology, several factors need to be considered.

First, determine which takes more time: problem exploration or concept exploration. Human-based methodology will excel at problem exploration, especially when complexity is high, and a lot of data is needed before making a decision. Ideally, designers also do not know what product will be made at the end. In this context, only KMK addressed the general problem of indoor parking and then offered several solutions, ranging from an application to a parking signboard integrated with the parking area’s lighting system. Meanwhile, the average other student already had an idea of the final product from the initial stage. This may be appropriate or less appropriate, depending on the subsequent process. If the weighting is substantial for concept exploration, that makes sense.

Designers are often hesitant to pursue concept exploration, whereas in practice, designers do not always develop products from scratch. Indeed, many of the products are quality improvements over previously produced items and can already be regarded as novelties. Creating added value is also considered sufficient, such as the addition of the peer-review feature implemented by FJS on a learning system management website.

To properly carry out human-based methodology, many products can be pulled back into more general questions. For example, “how can we create a seamless experience for event organisers who frequently move locations” so that the product produced may not be a backdrop (WLS), or “how can we make the experience of using a computer disability-friendly” so that it opens up many options beyond a foot pedal (AJG).

If the product has already been determined at the outset, then it is better to carry out concept exploration. This can still be done intensively with users, but the questions will focus more on the required features than on understanding user characteristics in general. Concept exploration can also be pursued using computer-aided engineering, as PLN did to determine the strength of recycled plastic and the load that its cat house can bear.

Second, pay attention to the choice of the three main focuses and determine which has the greatest weighting. If it is necessary to explore business opportunities or solve fairly complex problems, more time can be spent on the problem definition section through interviews, observations, and other methods. Human-based methods will be optimal if the designer can adopt a “fresh eyes” approach, such as IKC, who immersed himself in the low-vision community, and MAH, who collaborated with a coffee company to process coffee-ground waste. In this way, designers can minimise the most common bias, namely determining the solution from the outset and only then adjusting the design process.

If the product is used intensively by users, such as in applications or wearables, it would be good to integrate lean methods, as designers need a rapid iteration process. Interviews can still be conducted in depth, but based on prototypes rather than on information that is less relevant to be translated into a product, especially given that the work process lasts only one semester, while product development in companies can take years.

Regarding precedent studies and experiments on technology-based products, it should be noted that the word “empathise” is used to refer to the first stage of Stanford DT, whereas DD uses the word “discover”. Empathy will always be identical with humans, but ‘discover’ has a broad meaning. Designers can use any method, even those not directly related to humans. For example, the “discover” stage carried out by MAH involved extensive searches for data on coffee waste processing, and during the ‘develop’ stage, material exploration was conducted to identify the most suitable composition among coffee grounds, husks, and resin.

Designers must also not become trapped by existing design methods. Several students have successfully used methods they designed themselves, while several others used combined methods. FJS integrated the lean system into the second diamond of DD, and PLN transformed it into a triple diamond. In the visualisation process, the size of the diamonds also does not have to be the same. One of the diamonds can be made larger, flatter, and longer to represent the amount of time spent in this stage.

NPD material is taught in the Product Design & Development course in semester 3, which the author taught, but only BSD used it because they worked directly with a furniture company. In addition, this approach is better suited to technology-based products, whereas several students did not advance to the engineering stage. Even so, many students integrated the concept scoring stage, which is part of this methodology, into the concept selection process.

Third, choose an approach that can be used as a guide in the research process, where related theories will be very useful as a literature study. For example, in product design based on Yogyakarta and Batak culture, IKD and CPL used the IMO and ATUMICS methods, respectively, with a fairly strong integration of semiotic aspects. Meanwhile, to ensure their flatpack shoe rack could be assembled easily, quickly, and with minimal waste, MAM used the design-for-assembly approach.

In addition to those already mentioned in Table 5, there are many other approaches, such as design for the environment, design for ergonomics, design for maintainability, emotional design, and speculative design. Students’ ability to determine an approach will greatly influence the research weight of the project undertaken. Because in design research, creating a new product alone is not enough; there must also be a separate study that justifies the design process carried out. In this way, designers work not only with aspects of creativity [71] and visual elements [72] but also with problem-solving.

CONCLUSION

Beginning with a study of the urgency of separating the scientific method from the design method, the process carried out by designers was then translated into the DT method, which can be applied across various fields and serves as the basis for creating a competitive advantage for a company. However, within the field of design itself, DT can become an obstacle if designers do not choose the appropriate variation, thereby giving rise to opposing views about DT’s failure.

This paper groups several existing methods into human-based, prototype-based, mixed, and new product development categories, enabling designers to choose the appropriate product design research methodology. In the selection process, designers need to consider three things, namely: (1) the greater weighting between problem or concept exploration; (2) the main focus of the innovation lens; and (3) the theoretical approach that can increase the added value of the design. The main task of designers is to explain and justify the process they have chosen, including whether they have to create a new method based on existing methods.

In this way, designers are expected not merely to mention DT as a methodology, because that is where the main confusion lies: it is impossible for a designer not to carry out a design thinking process. If not, then they are doing it for themselves and not as a designer, just as not everyone who can cook is a chef.

FULL PAPER

Download link here (Bahasa Indonesia)

HOW TO CITE

Wiradarmo, A. A. (2025). The history and evaluation of design thinking as product design research methodology. Jurnal Desain Indonesia, 7(02), 98-133.

REFERENCES

In the full paper.