7.8 Directional Innovation in Smartphones

Marcos Antonio de Lima Filho, PhD.

As previously discussed, the touchscreen slate emerged as a dominant smartphone design (Section 6.4). Innovation scholars explain that a dominant design represents a milestone or transition point in the evolution of an industry (Suárez & Utterback, 1995). From this point onwards, future products are primarily driven by incremental refinement of the dominant design, as opposed to the introduction of completely novel architectures (Anderson & Tushman, 1990).

This normative effect of dominant designs is evidenced in the current dynamics of the smartphone industry. The diversity of smartphone form factors and keyboard types, prevalent before the emergence of the touchscreen slate, has practically disappeared (Figure 6.4.1). However, despite the near-universal adoption of this specific architecture, there is still a wide range of diversity within smartphones. Manufacturers continue to offer a variety of models that, though adhering to the dominant slate architecture, present their own unique set of incremental adaptations.

Smartphone manufacturers have differentiated and evolved their products across a multitude of dimensions, such as display sizes, CPU and graphic performance, camera quality, battery life, connectivity, and more (see Results 5.2). The resulting technological diversity can be visualised in the scatter plots exhibited throughout this thesis. For instance, each point plotted in Figure 7.8.2 corresponds to the CPU clock speed of a distinct mobile phone, totalling 10,255 models. The evident dispersion of these points demonstrates the varied range of options available for consumer selection.

Such incremental innovations are important because diversity is a necessary condition of evolution. Otherwise, if designs are perfectly homogeneous, there is nothing left to select from. Thankfully, products and services are hardly like undifferentiated commodities such as table salt. This extensive diversity in technological products, explains Basalla (1988), stems from the inherent variability in human values and preferences, which differ from culture to culture, from time to time, and from person to person.

Consequently, the management of incremental innovations should not be consigned to second-class status, as the generation of incremental products is a well-established aspect of industrial evolution. As noted previously, most successful products undergo continual incremental innovation to lower their costs and enhance their effectiveness (Norman & Verganti, 2014). However, despite the fact that the vast majority of new products are incremental improvements, radical innovation continues to receive the most attention from scholars and practitioners (Biskjaer et al., 2019).

The evolutionary trajectory of commercial aircraft and consumer electronics demonstrates the profound effect that incremental innovations can exert on cost, reliability, and overall performance. As emphasised in The Importance of Directional Innovations (Section 7.4), although individual changes may seem insignificant during incremental iterations, their cumulative impact often surpasses the original invention’s (Abernathy & Clark, 1985).

Consider early digital cameras, which were priced high and produced lower quality images compared to their 35 mm counterparts. Yet, as image quality steadily improved, digital cameras began to dominate the camera market (Dyer et al., 2018). Therefore, the incremental improvements that guided the maturity of this disruptive technology also contributed to the creative destruction of the industry. Then, the value of traditional photography assets, based on a century-old chemistry foundation, has been decimated (Magretta, 2012). Hence, the creative destruction of traditional film photography should be attributed equally to the groundbreaking advent of early digital cameras and the subsequent incremental improvements that guided this technology towards its maturity and widespread market adoption.

However, the window of opportunity for digital camera manufacturers to fully capitalise on this disruptive shift was relatively short-lived, as illustrated in Figure 7.8.1. According to the Camera and Imaging Products Association (CIPA), the sales of compact digital cameras peaked in 2008, with 110 million units sold. But by 2022, this figure had contracted to just 2 million units.

This drastic decline, representing a 98% drop in sales, was primarily driven by the widespread diffusion of smartphones. Equipped with increasingly sophisticated cameras, smartphones started serving the everyday photography needs of the mass market, thus reducing the demand for compact digital cameras. This rise and fall underscore the reality that no technology, however dominant, is immune to obsolescence when a new disruptive force emerges.

In response, the industry pivoted towards focusing on high-end products such as Digital Single Lens Reflex (DSLR) and mirrorless digital cameras. Despite the ongoing decrease since 2012, these high-end products remain appealing to professional photographers and photography enthusiasts.

This second wave of disruption, marked by the transition from digital cameras to smartphone cameras, was primarily driven by incremental innovation. Similar to the initial stages of digital cameras, the first smartphone cameras produced subpar image quality. Case in point, when the original iPhone was released in 2007, its 2-megapixel camera was not even considered a key feature:

In the beginning, the 2-megapixel camera that Apple tacked onto its original iPhone was hardly a pinnacle of innovation. Nor was it intended to be. “It was more like, every other phone has a camera, so we better have one too,” one senior member of the original iPhone team tells me. It’s not that Apple didn’t care about the camera; it’s just that resources were stretched thin, and it wasn’t really a priority. It certainly wasn’t considered a core feature by its founder; Jobs barely mentions it in the initial keynote (Merchant, 2017).

Over the years, camera and image quality have become major selling points for smartphones. Manufacturers responded with continuous improvements, thus gradually closing the quality gap with compact point-and-shoot digital cameras. This required advancements in both hardware and software, such as image sensors, optics, processing pipelines, and algorithms. Besides that, digital cameras could not offer the same convenience and ease of use. As smartphones became daily tools, compact digital cameras became redundant and obsolete for most consumers.

The concept of incremental innovation is an essential aspect of product development that managers and designers must not ignore due to its links with customer satisfaction. This potential correlation is further underscored by the Kano model (Section 9.4), a framework widely used in the field of quality management. This model identifies three distinct types of product attributes, namely basic, performance, and excitement factors, that significantly impact customer satisfaction.

As delineated by the Kano model, certain performance attributes exert a proportional influence on customer satisfaction. In other words, the more advanced or improved a feature is, the more satisfied a customer will likely be (Berger et al., 1993). Customers typically consider these so-called “one-dimensional attributes” when comparing different products. The stronger the attribute, the greater the customer satisfaction; conversely, weak values or the absence of a one-dimensional attribute will result in customer dissatisfaction (Horton & Goers, 2019).

For instance, in the automotive industry, gas mileage is a one-dimensional customer requirement. Greater gas mileage corresponds to increased customer satisfaction, while lower gas mileage typically yields dissatisfaction (Berger et al., 1993). This principle is similarly applicable to the range of an electric car (Horton & Goers, 2019). Within commercial aviation, factors like aerodynamic efficiency, fuel efficiency per passenger, reliability, and safety can also be categorised as “one-dimensional” attributes. Such advancements are unlikely to be ignored by customers.

Likewise, Kano’s model can provide valuable insights into the incremental evolution of smartphones. According to the model, an improvement in smartphone camera quality is likely to yield an increase in customer satisfaction, while a deterioration would have the opposite effect. While no specific study directly connects the Kano model with smartphone camera improvements and customer satisfaction, some indirect evidence suggests a correlation. These include the continued focus on camera improvements by smartphone manufacturers, along with customer reviews and surveys that prioritise camera quality, and general academic research supporting the Kano model's principles.

Alternatively, if customers showed apathy towards such technological progress, there would be no rationale for the observed thirteen-fold and forty-two-fold improvements in front and rear camera resolutions since 2010 (Figure 5.2.7).

---

The Force of Creative Destruction in the Mobile CPU Evolution

The evolution of mobile CPUs provides a prime illustration of Schumpeter’s concept of creative destruction. The continuous development of this technology has enabled the introduction of new and improved architectures that have displaced older, less efficient CPU architectures. This destructive effect demonstrates, once again, that constant incremental innovations can accumulate and result in major disruptions within the market.

Over the last two decades, mobile CPUs have become faster and more energy-efficient (Figure 7.8.2). However, this constant drive towards technological progress does not occur without a certain degree of collateral damage, embodied in the concept of “creative destruction”. This term, coined by economist Joseph Schumpeter, describes the process by which the relentless cycle of innovation and obsolescence forces industries to continuously reinvent themselves or face the prospects of stagnation and decline. Creative destruction is a paradoxical process: It’s a destructive process in that it renders previous investments redundant, but it’s also a creative force as it stimulates the development of new technologies, products, and processes.

In the context of the semiconductor industry, with the release of every new CPU generation, the capital invested in previous generations becomes redundant and obsolete. However, this destruction is not without its benefits. It paves the way for innovative companies to rise to the forefront, seizing the opportunity presented by market disequilibrium. These companies, by delivering advanced and superior CPUs, carve out a temporary competitive advantage for themselves. This advantage enables them to reap higher profits, at least for a while, until the next wave of innovation emerges. However, these temporary gains come at the cost of significant losses suffered by older, less efficient firms or industries that are forced out of business (Rothbard, 1987).

Lithography is the cornerstone of the semiconductor industry. This process allows the transference of circuit designs onto a silicon wafer. The continual refinement of this process is fundamental to consistently produce smaller, faster, and more energy-efficient chips. In short, the smaller the patterns, the more powerful and efficient the chip can be. Hence, the width of a transistor on a microchip has been the key measure of improvement in this field (Christensen et al., 2004).

Manufacturers of lithographic stepper tools, such as Nikon and ASML, engage in fierce competition to develop tools that offer the highest accuracy and resolution levels, as Adner and Snow (2010) pointed out. This is because their primary customers, like Samsung and Intel, require such tools to pack more circuits onto a given wafer, which ultimately enhances the performance of their products. As a result, the race to improve resolution is the key to maintaining competitive advantage and is therefore a top priority (Adner & Kapoor, 2016).

To make smaller transistors, light must be concentrated with greater precision to etch circuit patterns onto a silicon wafer. This is not only a meticulous process but also costly and technologically challenging. In the 1970s, industry experts postulated that the theoretical limit of this process was about one micron, or 1000 nanometers (nm). This was the smallest feature size they believed could be reliably produced through lithographic processes. By the mid-1980s, scientists believed the limit was 300 nm to 400 nm; by the 1990s, experts redefined the limit to 180 nm (Christensen et al., 2004).

As showcased in Figure 7.8.3, the semiconductor industry has consistently defied such theoretical projections. Within the period of 2000 to 2020, the industry has dramatically enhanced its lithographic processes, reducing process nodes from 180 nm down to just 5 nm.

Although the downsizing trend in the semiconductor industry may seem gradual and incremental, it is actually much more complex and disruptive. The implementation of more advanced lithographies often necessitates a complete redesign of circuit architectures and production methods. These innovations are rarely compatible with existing infrastructure, which can render previous process technologies obsolete. Thus, vast amounts of capital committed to older generations are frequently destroyed in semiconductor evolution.

This sort of creative destruction is not uncommon in the industry. Recent examples of this process include the transition from planar transistors to 3D FinFET designs starting in 2014, and the introduction of EUV (extreme ultraviolet) lithography in 2019. The production of 3D FinFET transistors required a considerable overhaul of existing manufacturing equipment and techniques, rendering previous investments in planar technology largely redundant.

Similarly, the adoption of EUV lithography has demanded heavy investment in new equipment and materials. As of 2023, ASML continues to be the sole producer of EUV equipment for chip manufacturing, focusing on cutting-edge 5 nm and 3 nm process nodes. EUV technology allows for greater miniaturisation of semiconductor circuits, but it also necessitates highly specialised, costly equipment that operates on entirely different principles compared to previous lithographic tools. This has led to a situation where older, non-EUV compatible equipment becomes obsolete, and the capital invested in these machines is effectively destroyed.

The semiconductor industry is also a good demonstration of how disruptive innovations initially introduced to high-end customers gradually spread to mainstream segments. For example, in 2017, both Samsung and Apple were the first to deploy 10 nm CPU chips in their high-end flagship products. The 10 nm fab process was later introduced in mid-range and then low-end smartphones. Similarly, in 2018, Apple’s A12 processor became the first 7 nm chip to go into commercial production. Built by TSMC, this high-end innovation subsequently spread to other foundries and smartphone manufacturers. By 2020, the 7 nm process had emerged as the dominant lithography, accounting for 39.5% of all smartphone models launched that year.

When a chip foundry phases out an outdated process technology from smartphone chip production, these matured processes often discover renewed relevance in other consumer electronics. These products may include digital TVs, streaming devices, IoT devices, and certain automotive applications. With different performance and cost requirements, these products can benefit from the more mature and cost-effective production methods associated with outdated process technology.

This process of top-down diffusion has contributed to the development of crucial smartphone components such as CPUs, memory chips, and image sensors. These advanced technologies initially appear as premium features in high-end devices, serving as significant differentiators that justify a higher price tag. As these technologies mature and diffuse to other manufacturers, the costs associated with them gradually decrease. This cost reduction enables these once cutting-edge features to be integrated into more affordable devices.

The dynamics described in the semiconductor industry seem to contradict Christensen’s concept of disruption. Instead of innovations emerging from lower-end markets and then moving upwards, in the semiconductor industry, this constant cycle of creative destruction is primarily financed by some leading manufacturers, who concentrate their efforts on early adopters and premium market segments. As these technologies mature and become less expensive to produce, they then cascade down to the more mainstream, lower-cost products.