This document discusses exploring the limits of technology S-curves by examining their usefulness for managers in planning new technology development. It focuses on the disk drive industry as a case study. The author makes four key points: 1) S-curves accurately describe industry-level technology substitution patterns, 2) to improve products, managers must oversee both component and architectural technology development, 3) S-curves describe individual firm experiences with components but cannot prescribe strategy, and 4) attackers gain advantage in this industry through architectural, not just technological, innovation in new applications.
This document discusses the implementation of design for manufacturing (DFM) principles in mechanical engineering education. It provides examples of applying DFM to dimensional tolerancing using standards to ensure functionality and manufacturability. Advanced computer technologies like CAD/CAM and AI are also discussed in improving the link between design and manufacturing to better implement DFM concepts. The authors aim to integrate DFM training earlier in the manufacturing engineering curriculum.
This document summarizes research adapting the Supply Chain Operations Reference (SCOR) model to the construction industry. The SCOR model was originally developed for manufacturing but lacks consideration of construction industry characteristics. Through case studies of construction projects over five years, the researchers identified necessary adaptations. They developed the Builder's SCOR (BSCOR) model which retains SCOR's structure but adds new processes and metrics to capture material delivery to sites, on-site procurement, construction, and planning activities. The BSCOR model allows mapping and measuring performance of construction supply chains from suppliers to project completion.
Using data collected from 100 construction projects in Iraq, the study developed a multiple linear regression model to predict productivity of marble floor finishing work. The model uses 10 independent variables that influence productivity, including worker age, experience, crew size, floor height, tile size, security, worker health, weather, worksite conditions, and material availability. The regression model achieved an excellent correlation coefficient of 90.6% and average accuracy of 96.3% when validated, indicating a strong relationship between the independent and dependent variables.
Concurrent engineering an effective engineering management strategyIAEME Publication
The document discusses concurrent engineering as an effective engineering management strategy for product development. It defines concurrent engineering as an approach that integrates product and process development from the beginning to complete the product lifecycle faster at lower cost. The document outlines the history and evolution of concurrent engineering. It describes the key elements of a concurrent engineering approach including concept development, customer needs identification, and design for manufacturing. It emphasizes the importance of teamwork and provides guidelines for successful implementation through multidisciplinary teams and management support.
The applicant has extensive experience in the aeronautical field spanning over 30 years, including expertise in jet engine overhaul, maintenance, manufacturing, and testing. He seeks a new career opportunity utilizing his skills in leadership, project management, and process improvement. The applicant's background includes experience with Pratt & Whitney JT8D and IAE V2500 engines, aircraft manufacturing, and international customer service.
IRJET-Quality Benchmarking in Construction IndustryIRJET Journal
This document summarizes a research paper that studied differences in project planning quality across different industries. It found that construction and engineering organizations planned projects better than other industries like software, services, and production/maintenance. Construction and engineering organizations also had less cost and schedule overruns, indicating better project success. The study used a Project Management Planning Quality model to evaluate the use of 33 different planning processes and organizational support factors. It analyzed data from 201 project managers across four industries. Key findings were that construction/engineering planned projects best while production/maintenance planned projects worst, and construction/engineering projects succeeded best while production/maintenance projects succeeded worst.
A Methodology for Parametric Production Planning in Preliminary Aircraft Desi...Chandrashekar Sundaresan
This document introduces a methodology for integrating aircraft design and production planning during the preliminary design phase. The methodology consists of two main parts: 1) A parametric equipment and tooling model that estimates tooling costs, number of tools, and tooling capacity for different production scenarios. 2) A production planning optimization model that minimizes total cost to meet demand under constraints. Design of experiments and surrogate modeling are used to parameterize the models. The methodology is intended to enable trades between design alternatives and producibility during preliminary design. A case study on an advanced composite fighter wing box serves as a proof of concept.
Joseph Ragusa has over 30 years of experience in industrial engineering, project management, and supply chain roles in the aerospace industry. He has a proven track record of reducing costs through lean initiatives, sourcing externally, and transitioning production. Currently he is a Project Manager at Triumph Aerostructures where he develops supplier chargeback processes and transitions production of over 3,200 parts to external suppliers.
This document discusses the implementation of design for manufacturing (DFM) principles in mechanical engineering education. It provides examples of applying DFM to dimensional tolerancing using standards to ensure functionality and manufacturability. Advanced computer technologies like CAD/CAM and AI are also discussed in improving the link between design and manufacturing to better implement DFM concepts. The authors aim to integrate DFM training earlier in the manufacturing engineering curriculum.
This document summarizes research adapting the Supply Chain Operations Reference (SCOR) model to the construction industry. The SCOR model was originally developed for manufacturing but lacks consideration of construction industry characteristics. Through case studies of construction projects over five years, the researchers identified necessary adaptations. They developed the Builder's SCOR (BSCOR) model which retains SCOR's structure but adds new processes and metrics to capture material delivery to sites, on-site procurement, construction, and planning activities. The BSCOR model allows mapping and measuring performance of construction supply chains from suppliers to project completion.
Using data collected from 100 construction projects in Iraq, the study developed a multiple linear regression model to predict productivity of marble floor finishing work. The model uses 10 independent variables that influence productivity, including worker age, experience, crew size, floor height, tile size, security, worker health, weather, worksite conditions, and material availability. The regression model achieved an excellent correlation coefficient of 90.6% and average accuracy of 96.3% when validated, indicating a strong relationship between the independent and dependent variables.
Concurrent engineering an effective engineering management strategyIAEME Publication
The document discusses concurrent engineering as an effective engineering management strategy for product development. It defines concurrent engineering as an approach that integrates product and process development from the beginning to complete the product lifecycle faster at lower cost. The document outlines the history and evolution of concurrent engineering. It describes the key elements of a concurrent engineering approach including concept development, customer needs identification, and design for manufacturing. It emphasizes the importance of teamwork and provides guidelines for successful implementation through multidisciplinary teams and management support.
The applicant has extensive experience in the aeronautical field spanning over 30 years, including expertise in jet engine overhaul, maintenance, manufacturing, and testing. He seeks a new career opportunity utilizing his skills in leadership, project management, and process improvement. The applicant's background includes experience with Pratt & Whitney JT8D and IAE V2500 engines, aircraft manufacturing, and international customer service.
IRJET-Quality Benchmarking in Construction IndustryIRJET Journal
This document summarizes a research paper that studied differences in project planning quality across different industries. It found that construction and engineering organizations planned projects better than other industries like software, services, and production/maintenance. Construction and engineering organizations also had less cost and schedule overruns, indicating better project success. The study used a Project Management Planning Quality model to evaluate the use of 33 different planning processes and organizational support factors. It analyzed data from 201 project managers across four industries. Key findings were that construction/engineering planned projects best while production/maintenance planned projects worst, and construction/engineering projects succeeded best while production/maintenance projects succeeded worst.
A Methodology for Parametric Production Planning in Preliminary Aircraft Desi...Chandrashekar Sundaresan
This document introduces a methodology for integrating aircraft design and production planning during the preliminary design phase. The methodology consists of two main parts: 1) A parametric equipment and tooling model that estimates tooling costs, number of tools, and tooling capacity for different production scenarios. 2) A production planning optimization model that minimizes total cost to meet demand under constraints. Design of experiments and surrogate modeling are used to parameterize the models. The methodology is intended to enable trades between design alternatives and producibility during preliminary design. A case study on an advanced composite fighter wing box serves as a proof of concept.
Joseph Ragusa has over 30 years of experience in industrial engineering, project management, and supply chain roles in the aerospace industry. He has a proven track record of reducing costs through lean initiatives, sourcing externally, and transitioning production. Currently he is a Project Manager at Triumph Aerostructures where he develops supplier chargeback processes and transitions production of over 3,200 parts to external suppliers.
Dear students get fully solved assignments
Send your semester & Specialization name to our mail id :
help.mbaassignments@gmail.com
or
call us at : 08263069601
Dear students get fully solved assignments
Send your semester & Specialization name to our mail id :
help.mbaassignments@gmail.com
or
call us at : 08263069601
This document discusses different types and patterns of innovation. It describes the differences between product and process innovation, radical and incremental innovation, competence-enhancing and competence-destroying innovation, and architectural and component innovation. It also discusses how the rate of technological improvement and adoption often follows an S-curve pattern over time. However, using S-curves to precisely predict the limits of a technology or the timing of switching to a new technology can be misleading due to varying investment levels and comparison benchmarks.
Current Trends in Product Development during COVID-19vivatechijri
This document summarizes current trends in product development during the COVID-19 pandemic. It discusses how the pandemic has accelerated existing trends like distributed product development teams collaborating online, as well as new trends like increased use of virtual prototyping and simulation to speed processes while limiting physical contact. Product development is increasingly integrating different technologies and balancing both technical and user requirements. Sustainability and the entire product lifecycle from design to end-of-life are also major considerations in current product development practices.
The document discusses performance analysis of digital signal processing (DSP) technologies using the S curve model. It begins by establishing the theoretical foundations of the S curve model, noting its relationships to innovation diffusion models and life cycle models. The S curve model represents the evolution of technology performance over time. The document then presents a methodology for applying the S curve model to analyze the technology performance of DSPs. Finally, it discusses the possibilities and usefulness of the S curve model as a strategic analysis tool.
The document discusses performance analysis of digital signal processing (DSP) technologies using the S curve model. It begins by establishing the theoretical foundations of the S curve model, noting its relationships to innovation diffusion models and life cycle models. The S curve model represents the evolution of technology performance over time. The document then presents a methodology for applying the S curve model to analyze the technology performance of DSPs. Finally, it discusses the usefulness of the S curve model as a strategic analysis tool.
This document discusses different types and patterns of innovation. It describes product versus process innovation, competence-enhancing versus competence-destroying innovation, component versus architectural innovation. It also discusses how the performance of technologies often follows an S-curve pattern over time, and how the adoption of new technologies by the market also follows an S-curve pattern. Finally, it notes that while the S-curve model can be used as a predictive tool by managers, it also has some limitations as a prescriptive tool.
The previous chapters pointed out that technological innovation can come from many sources and take many forms. Different types of technological innovations offer different opportunities for organizations and society, and they pose different demands upon producers, users, and regulators. While there is no single agreed-upon taxonomy to describe different kinds of technological innovations, in this chapter we will review several dimensions that are often used to categorize technologies. These dimensions are useful for understanding some key ways that one innovation may differ from another. The path a technology follows through time is termed its technology trajectory. Technology trajectories are most often used to represent the technology’s rate of performance improvement or its rate of adoption in the marketplace. Though many factors can influence these technology trajectories (as discussed in both this chapter and the following chapters), some patterns have been consistently identified in technology trajectories across many industry contexts and over many periods. Understanding these patterns of technological innovation provides a useful foundation that we will build upon in the later chapters on formulating technology strategy. The chapter begins by reviewing the dimensions used to distinguish types of innovations. It then describes the s-curve patterns so often observed in both the rate of technology improvement and the rate of technology diffusion to the market. In the last section, the chapter describes research suggesting that technological innovation follows a cyclical pattern composed of distinct and reliably occurring phases.
.
Using s curves and trends of evolution in r&d strategy planningAshok Rangasamy
This document discusses using S-curves and trends of evolution to aid in R&D strategy planning. It examines the mature product optimization/innovation dilemma and how TRIZ tools can help determine the best strategic path forward, such as whether to continue optimizing an existing technology or innovating towards a new generation. The document analyzes four TRIZ metrics - performance, number of inventions, level of complexity, and profitability - that can indicate where a product lies along its evolutionary S-curve and help assess if further optimization or a new innovation is warranted. It applies this framework to analyzing the maturity of refrigerant compressor technology.
Chapter 3 Schilling 2017 Types and Patterns of Innovationahmdirvan
This document discusses different types and patterns of innovation. It describes the differences between product and process innovation, radical and incremental innovation, competence-enhancing and competence-destroying innovation, and architectural and component innovation. It also discusses how innovations can be analyzed using dimensions like radicalness and newness. Finally, it covers how the performance and adoption of technologies often follow S-curve patterns over time, and how S-curves can provide insights but also have limitations as a predictive tool.
Rapid Improvements with No Commercial Production: How do the Improvements Occ...Jeffrey Funk
This paper empirically examines 13 technologies in which significant cost and performance improvements occurred even while no commercial production occurred. Since the literature emphasizes cost reductions through increases in cumulative production, this paper explores cost and performance improvements from a new perspective. The results demonstrate that learning in these pre-commercial production cases arises through mechanisms utilized in deliberate R&D efforts. We identity three mechanisms - materials creation, process changes, and reductions in feature scale – that enable these improvements to occur and use them to extend models of learning and invention. These mechanisms can also apply during post commercial time periods and further research is needed to quantify the relative contributions of these three mechanisms and those of production-based learning in a variety of technologies.
This document discusses different types and patterns of innovation:
1) It describes product innovations versus process innovations and radical innovations versus incremental innovations.
2) It also discusses competence-enhancing innovations versus competence-destroying innovations and architectural innovations versus component innovations.
3) Technologies often follow an S-curve pattern, where performance and adoption rates improve slowly at first, then accelerate, before reaching saturation. However, using S-curves to precisely predict new phases can be misleading.
This document discusses different types and patterns of innovation:
1. It describes product innovation versus process innovation, and radical innovation versus incremental innovation. Radical innovations are new to the world and significantly different, while incremental innovations are marginally different.
2. Innovations can also be competence-enhancing or competence-destroying from the perspective of a firm. Competence-enhancing innovations build on a firm's existing knowledge, while competence-destroying innovations render existing competencies obsolete.
3. Architectural innovations change a system's overall design and interactions, while component innovations only change parts but not the overall architecture. Architectural innovations often have broader impacts.
4. Performance and adoption of
The document discusses technology trajectories and patterns of technological innovation. It describes how technologies typically follow an S-curve where performance improves slowly at first, then rapidly, before reaching diminishing returns. This S-curve model suggests that technological change proceeds in cycles, with each new technology initially causing turbulence before a period of rapid improvement, until it is displaced by another discontinuity. The document also categorizes different types of innovation as product vs process, radical vs incremental, competence-enhancing vs competence-destroying, and architectural vs component. Understanding these typical patterns of technological change and innovation provides a foundation for formulating technology strategy.
Technological innovations are often described using dimensions such as “radical” versus “incremental.” Different types of innovation require different kinds of underlying knowledge and have different impacts on the industry’s competitors and customers. Four of the dimensions most commonly used to categorize innovations are described here: product versus process innovation, radical versus incremental, competence enhancing versus competence destroying, and architectural versus component.
This document summarizes the strategic plan implemented by Pratt & Whitney to improve manufacturing efficiency at their Southington, Connecticut plant. The plan focused on implementing group technology concepts through physical changes like installing flowlines for disk and hub production and cellular manufacturing for drum rotors. It also included organizational changes like establishing business units responsible for specific part families. The plan required $173 million for new equipment and $50 million for facility upgrades over 5 years. It aimed to leverage technologies and focus efforts to dramatically improve quality, cost and delivery of gas turbine components.
This document discusses the application of statistical process control (SPC) in automotive manufacturing. It provides 4 case studies that demonstrate both basic and advanced applications of SPC, including using SPC with multivariate analysis and design of experiments. It also describes a case where SPC was ignored, leading to failed experiments. The case studies illustrate how SPC can be used to monitor and improve processes, reduce variation, and gain understanding of process capabilities.
Challenges in ramp-up process according to the findings in articles, Ishikawa analsis for the challenges in ramp-up. Strategies to overcome the mentioned challenges and appopriate ramp-up categories for the strategies.
References are given.
This document discusses technology S-curves and their implications for renewable energy alternatives. It finds that both wind and geothermal energy are poised to become more economical than fossil fuels relatively soon based on historical data. Government R&D funding has been lower for wind and geothermal compared to solar, and funding for fossil fuels may be excessive given their diminishing performance. Analyzing renewable technologies through an S-curve lens provides insights not found using other approaches and has implications for future government and industry investment decisions.
This document discusses technology S-curves and their implications for renewable energy alternatives. It finds that both wind and geothermal energy are poised to become more economical than fossil fuels relatively soon based on historical cost reductions from R&D investment. However, government R&D funding has been lower for wind and geothermal compared to solar energy. The analysis suggests governments may over-fund fossil fuel technologies given their diminishing performance improvements. Examining renewable technologies through an S-curve lens provides insights not found using traditional experience curves and indicates areas where industry and government could better support technology development.
Technology S-Curve Analysis (TSC) is a method to determine the relationship between investment in improving technology and corresponding market sales. The curve typically takes the shape of an S as the market develops through phases of launch, growth, maturity, and decline. TSC Analysis is particularly useful for launching new technologies and assessing the remaining lifespan of technologies to guide business strategy and shifting resources to next-generation technologies. TSC Analysis provides intelligence on the current technology lifespan, potential limits, economics of innovation, a company's position, and quantifying investment versus payoff.
The document discusses S curves, which plot cumulative project quantities like costs or hours against time. It describes different types of S curves, including baseline, target, and actual S curves. The document explains how S curves can be used to analyze a project's progress, growth, and slippage. It also provides details on generating S curves from project schedule data and interpreting S curve analyses.
More Related Content
Similar to Exploring the limits_of_the_technology_s-curve
Dear students get fully solved assignments
Send your semester & Specialization name to our mail id :
help.mbaassignments@gmail.com
or
call us at : 08263069601
Dear students get fully solved assignments
Send your semester & Specialization name to our mail id :
help.mbaassignments@gmail.com
or
call us at : 08263069601
This document discusses different types and patterns of innovation. It describes the differences between product and process innovation, radical and incremental innovation, competence-enhancing and competence-destroying innovation, and architectural and component innovation. It also discusses how the rate of technological improvement and adoption often follows an S-curve pattern over time. However, using S-curves to precisely predict the limits of a technology or the timing of switching to a new technology can be misleading due to varying investment levels and comparison benchmarks.
Current Trends in Product Development during COVID-19vivatechijri
This document summarizes current trends in product development during the COVID-19 pandemic. It discusses how the pandemic has accelerated existing trends like distributed product development teams collaborating online, as well as new trends like increased use of virtual prototyping and simulation to speed processes while limiting physical contact. Product development is increasingly integrating different technologies and balancing both technical and user requirements. Sustainability and the entire product lifecycle from design to end-of-life are also major considerations in current product development practices.
The document discusses performance analysis of digital signal processing (DSP) technologies using the S curve model. It begins by establishing the theoretical foundations of the S curve model, noting its relationships to innovation diffusion models and life cycle models. The S curve model represents the evolution of technology performance over time. The document then presents a methodology for applying the S curve model to analyze the technology performance of DSPs. Finally, it discusses the possibilities and usefulness of the S curve model as a strategic analysis tool.
The document discusses performance analysis of digital signal processing (DSP) technologies using the S curve model. It begins by establishing the theoretical foundations of the S curve model, noting its relationships to innovation diffusion models and life cycle models. The S curve model represents the evolution of technology performance over time. The document then presents a methodology for applying the S curve model to analyze the technology performance of DSPs. Finally, it discusses the usefulness of the S curve model as a strategic analysis tool.
This document discusses different types and patterns of innovation. It describes product versus process innovation, competence-enhancing versus competence-destroying innovation, component versus architectural innovation. It also discusses how the performance of technologies often follows an S-curve pattern over time, and how the adoption of new technologies by the market also follows an S-curve pattern. Finally, it notes that while the S-curve model can be used as a predictive tool by managers, it also has some limitations as a prescriptive tool.
The previous chapters pointed out that technological innovation can come from many sources and take many forms. Different types of technological innovations offer different opportunities for organizations and society, and they pose different demands upon producers, users, and regulators. While there is no single agreed-upon taxonomy to describe different kinds of technological innovations, in this chapter we will review several dimensions that are often used to categorize technologies. These dimensions are useful for understanding some key ways that one innovation may differ from another. The path a technology follows through time is termed its technology trajectory. Technology trajectories are most often used to represent the technology’s rate of performance improvement or its rate of adoption in the marketplace. Though many factors can influence these technology trajectories (as discussed in both this chapter and the following chapters), some patterns have been consistently identified in technology trajectories across many industry contexts and over many periods. Understanding these patterns of technological innovation provides a useful foundation that we will build upon in the later chapters on formulating technology strategy. The chapter begins by reviewing the dimensions used to distinguish types of innovations. It then describes the s-curve patterns so often observed in both the rate of technology improvement and the rate of technology diffusion to the market. In the last section, the chapter describes research suggesting that technological innovation follows a cyclical pattern composed of distinct and reliably occurring phases.
.
Using s curves and trends of evolution in r&d strategy planningAshok Rangasamy
This document discusses using S-curves and trends of evolution to aid in R&D strategy planning. It examines the mature product optimization/innovation dilemma and how TRIZ tools can help determine the best strategic path forward, such as whether to continue optimizing an existing technology or innovating towards a new generation. The document analyzes four TRIZ metrics - performance, number of inventions, level of complexity, and profitability - that can indicate where a product lies along its evolutionary S-curve and help assess if further optimization or a new innovation is warranted. It applies this framework to analyzing the maturity of refrigerant compressor technology.
Chapter 3 Schilling 2017 Types and Patterns of Innovationahmdirvan
This document discusses different types and patterns of innovation. It describes the differences between product and process innovation, radical and incremental innovation, competence-enhancing and competence-destroying innovation, and architectural and component innovation. It also discusses how innovations can be analyzed using dimensions like radicalness and newness. Finally, it covers how the performance and adoption of technologies often follow S-curve patterns over time, and how S-curves can provide insights but also have limitations as a predictive tool.
Rapid Improvements with No Commercial Production: How do the Improvements Occ...Jeffrey Funk
This paper empirically examines 13 technologies in which significant cost and performance improvements occurred even while no commercial production occurred. Since the literature emphasizes cost reductions through increases in cumulative production, this paper explores cost and performance improvements from a new perspective. The results demonstrate that learning in these pre-commercial production cases arises through mechanisms utilized in deliberate R&D efforts. We identity three mechanisms - materials creation, process changes, and reductions in feature scale – that enable these improvements to occur and use them to extend models of learning and invention. These mechanisms can also apply during post commercial time periods and further research is needed to quantify the relative contributions of these three mechanisms and those of production-based learning in a variety of technologies.
This document discusses different types and patterns of innovation:
1) It describes product innovations versus process innovations and radical innovations versus incremental innovations.
2) It also discusses competence-enhancing innovations versus competence-destroying innovations and architectural innovations versus component innovations.
3) Technologies often follow an S-curve pattern, where performance and adoption rates improve slowly at first, then accelerate, before reaching saturation. However, using S-curves to precisely predict new phases can be misleading.
This document discusses different types and patterns of innovation:
1. It describes product innovation versus process innovation, and radical innovation versus incremental innovation. Radical innovations are new to the world and significantly different, while incremental innovations are marginally different.
2. Innovations can also be competence-enhancing or competence-destroying from the perspective of a firm. Competence-enhancing innovations build on a firm's existing knowledge, while competence-destroying innovations render existing competencies obsolete.
3. Architectural innovations change a system's overall design and interactions, while component innovations only change parts but not the overall architecture. Architectural innovations often have broader impacts.
4. Performance and adoption of
The document discusses technology trajectories and patterns of technological innovation. It describes how technologies typically follow an S-curve where performance improves slowly at first, then rapidly, before reaching diminishing returns. This S-curve model suggests that technological change proceeds in cycles, with each new technology initially causing turbulence before a period of rapid improvement, until it is displaced by another discontinuity. The document also categorizes different types of innovation as product vs process, radical vs incremental, competence-enhancing vs competence-destroying, and architectural vs component. Understanding these typical patterns of technological change and innovation provides a foundation for formulating technology strategy.
Technological innovations are often described using dimensions such as “radical” versus “incremental.” Different types of innovation require different kinds of underlying knowledge and have different impacts on the industry’s competitors and customers. Four of the dimensions most commonly used to categorize innovations are described here: product versus process innovation, radical versus incremental, competence enhancing versus competence destroying, and architectural versus component.
This document summarizes the strategic plan implemented by Pratt & Whitney to improve manufacturing efficiency at their Southington, Connecticut plant. The plan focused on implementing group technology concepts through physical changes like installing flowlines for disk and hub production and cellular manufacturing for drum rotors. It also included organizational changes like establishing business units responsible for specific part families. The plan required $173 million for new equipment and $50 million for facility upgrades over 5 years. It aimed to leverage technologies and focus efforts to dramatically improve quality, cost and delivery of gas turbine components.
This document discusses the application of statistical process control (SPC) in automotive manufacturing. It provides 4 case studies that demonstrate both basic and advanced applications of SPC, including using SPC with multivariate analysis and design of experiments. It also describes a case where SPC was ignored, leading to failed experiments. The case studies illustrate how SPC can be used to monitor and improve processes, reduce variation, and gain understanding of process capabilities.
Challenges in ramp-up process according to the findings in articles, Ishikawa analsis for the challenges in ramp-up. Strategies to overcome the mentioned challenges and appopriate ramp-up categories for the strategies.
References are given.
This document discusses technology S-curves and their implications for renewable energy alternatives. It finds that both wind and geothermal energy are poised to become more economical than fossil fuels relatively soon based on historical data. Government R&D funding has been lower for wind and geothermal compared to solar, and funding for fossil fuels may be excessive given their diminishing performance. Analyzing renewable technologies through an S-curve lens provides insights not found using other approaches and has implications for future government and industry investment decisions.
This document discusses technology S-curves and their implications for renewable energy alternatives. It finds that both wind and geothermal energy are poised to become more economical than fossil fuels relatively soon based on historical cost reductions from R&D investment. However, government R&D funding has been lower for wind and geothermal compared to solar energy. The analysis suggests governments may over-fund fossil fuel technologies given their diminishing performance improvements. Examining renewable technologies through an S-curve lens provides insights not found using traditional experience curves and indicates areas where industry and government could better support technology development.
Similar to Exploring the limits_of_the_technology_s-curve (20)
Technology S-Curve Analysis (TSC) is a method to determine the relationship between investment in improving technology and corresponding market sales. The curve typically takes the shape of an S as the market develops through phases of launch, growth, maturity, and decline. TSC Analysis is particularly useful for launching new technologies and assessing the remaining lifespan of technologies to guide business strategy and shifting resources to next-generation technologies. TSC Analysis provides intelligence on the current technology lifespan, potential limits, economics of innovation, a company's position, and quantifying investment versus payoff.
The document discusses S curves, which plot cumulative project quantities like costs or hours against time. It describes different types of S curves, including baseline, target, and actual S curves. The document explains how S curves can be used to analyze a project's progress, growth, and slippage. It also provides details on generating S curves from project schedule data and interpreting S curve analyses.
The S curve is used to spread project costs over time in a bell-shaped curve. It is based on the sine and cosine waves, with the integral of the sine wave producing a leaning S-shaped cumulative cost curve. The document describes tweaking the basic S curve model by introducing variables to skew the timing of costs (make them front-loaded or back-loaded) and adjust the peakness to flatten or steepen the curve. Equations are provided to calculate cost at any point in time or rate of spending based on total project cost, time, skewness, and peakness variables.
The document provides safety instructions and an overview of the features and operation of the Dual 31 Band Equalizer. Key points include:
- The equalizer has 31 frequency bands covering 20Hz to 20kHz that can each be boosted or cut by up to 12dB or 6dB depending on the range setting.
- It has electronically balanced inputs and outputs, variable low cut filter, bypass switches, and LED VU meters.
- The document explains the controls on the front and rear panels and provides instructions for setting up and using the equalizer for different applications such as with a mixer, patchbay, or real time analyzer.
This document discusses the application of S-shaped curves, also known as logistic curves or S-curves, to model the evolution of systems over time. It provides background on the origin and development of S-curves as models of growth. S-curves have been widely used across many domains to describe trends like population growth, market penetration of new technologies, and diffusion of innovations. The document reviews several examples of S-curve applications and discusses their use in areas like technological forecasting and TRIZ problem solving. It argues that S-curves provide forecasting power because growth is ultimately limited by scarce resources based on mathematical concepts like Verhulst's logistic growth equation.
This document presents a case study analyzing the S-curve of a construction project. It describes the original planned S-curve, which projected slow initial progress accelerating in the middle before slowing at completion. Actual progress fell far below this, requiring three adjustments - adding scope, and extending the schedule by 9 months total - to get the project back on track. By the third adjustment, actual progress more closely aligned with the planned S-curve, allowing the project to progress through start-up, growth, and commissioning phases to completion. The case study demonstrates how S-curve analysis can identify issues and inform decisions to successfully manage a project over its lifecycle.
The document provides documentation and a tutorial for implementing an S-curve motion profile in a MyoStat Motion Control system. An S-curve allows for smooth acceleration and deceleration during motion to prevent damage. The K69 parameter controls the S-curve function, with higher values creating a more pronounced curve. The tutorial instructs the user to set parameters, collect speed data, and graph results at different K69 values to observe the S-curve motion profile.
The document discusses optimizing S-curve velocity profiles for motion control. An S-curve velocity profile is a smooth curve that is differentiable to the second order. The document describes decomposing 3D motion into 1D components and synthesizing physical constraints on jerk, acceleration, and velocity for each axis. It then presents calculations for generating an optimized S-curve velocity profile that satisfies the constraints and produces smooth, fast, and accurate motion between a start and end point.
The document discusses how to generate S-curves in Oracle Primavera P6 to analyze project progress and performance. S-curves show cumulative costs, labor hours, or other metrics plotted against time and typically have an S-shape. In Primavera P6, S-curves can be generated by activity or resource in the usage profile windows. Various analysis can then be done by comparing baseline, target, and actual S-curves to determine project growth, slippage, and progress. The S-curves can also be published from Primavera P6 as prints or embedded in webpages.
This document describes a methodology for modeling S-curves to forecast cost distribution over time for construction projects. It discusses three approaches to S-curve modeling: 1) Analyzing cost-time curves from literature, 2) Examining data from completed projects, and 3) Creating standard critical path models. The results from each approach are presented through mathematical expressions and diagrams. Finally, the results of the three approaches are integrated to develop final S-curves showing standard cost dynamics over time for different project types (buildings, tunnels, motorways). The proposed methodology can be used to forecast cost schedules during early project phases when detailed information is limited.
This document summarizes an optimization model for a contractor's S-curve developed using genetic algorithms. The model aims to minimize total construction costs while considering the tradeoff between different resource productivity and costs of resource mobilization/demobilization. An example project is optimized, resulting in a smoother resource allocation and lower total cost compared to the early and late schedules. The optimal S-curve developed from the model provides a baseline for measuring impacts of changes on construction costs.
This document discusses innovation lifecycles and how companies can leverage S-curves to drive breakthrough growth. It contains the following key points:
1. Products, services, and technologies progress along S-curves over time from emergence to maturity. Understanding where opportunities fall in their lifecycles can help companies innovate and avoid disruption.
2. Companies often fail to transition to new S-curves due to not focusing on or defending emerging technologies, cultural inertia, or lack of foresight.
3. Organizational strategies like leadership, structure, and metrics should evolve to support innovation or optimization depending on a business's point in the lifecycle. Tailoring the organization maximizes growth across the portfolio
This document discusses three frameworks used by Gartner Group to analyze information systems research: the technology maturity curve, adoption curve, and identification of strategic technologies. The maturity curve tracks how a technology matures over time through various stages from embryonic to obsolescence. The adoption curve shows how technologies are adopted cumulatively by organizations over time. Considering where technologies fall on these curves can provide insights into appropriate research questions and methodologies. Identifying strategic technologies may help determine promising areas for new research.
The document discusses an S-curve model that relates per-capita income to insurance penetration. It finds:
1) Estimating life and non-life insurance penetration globally yields an S-curve, where income elasticity starts and ends at 1 but exceeds 1 at intermediate income levels.
2) For life insurance, the income at maximum elasticity is $15,000, while for non-life it is $10,000.
3) Using purchasing power parities rather than market exchange rates increases estimated penetration levels and elasticities for developing countries.
This document describes proposed changes to improve a model for estimating project S-curves based on project attributes and conditions. The key changes are:
1. Changing the model outputs from the polynomial function's parameters (a, b) to the inflection point position (p) and slope (s), which better indicate schedule performance.
2. Adding two new input factors - project difficulty and participant competence - to capture schedule influences beyond basic attributes like cost and duration.
3. Using fuzzy inference systems instead of neural networks to build transparent input-output relationships from historical project data, applying fuzzy clustering and hybrid training.
The goal is to develop a model that more accurately predicts project progress curves by incorporating schedule performance indicators
The document discusses patterns of technological substitution that challenge the traditional S-curve model. It presents several historical examples that demonstrate more complex substitution patterns than the smooth S-curve, including: concatenated generations in steelmaking technologies; overlapping generations in IBM mainframe computers; and a case of long-term feedback reversing the substitution of DDT as an insecticide due to environmental concerns. The author argues that accounting for these complex real-world patterns requires broadening the theoretical framework for understanding technological substitutions.
This document discusses the history and development of semiconductors and integrated circuits. It describes how the transistor enabled electronics to be performed using silicon, leading to solid-state electronics like transistor radios. The integrated circuit was developed using the planar process to fabricate multiple transistors on silicon wafers. Moore's Law, proposed in 1965, predicted that the number of transistors on an integrated circuit would double every 18 months. This prediction has proven remarkably accurate and has driven innovation in the semiconductor industry for over 40 years. Continued shrinking of circuit elements has enabled faster processing speeds, higher functionality, and lower costs over time.
The document summarizes the key points from the book "Jumping the S-Curve" by Paul Nunes and Tim Breene. It discusses how companies can succeed by repeatedly climbing business S-curves and jumping to new S-curves before performance plateaus. To climb an S-curve, companies must identify a large market opportunity, build the necessary capabilities, and attract top talent. To jump to a new S-curve, companies must manage hidden S-curves, develop edge-centric strategies, reconstitute leadership teams early, and cultivate a talent pipeline. High performers are able to continually reinvent themselves through repeated S-curve climbing and jumping.
This document summarizes a research paper on algorithms for planning s-curve motion profiles.
The paper generalizes the model of polynomial s-curve motion profiles in a recursive form. It then proposes a general algorithm to design s-curve trajectories in a time-optimal manner. The algorithm calculates the time periods for connecting trajectory segments to generate a smooth path that meets velocity and acceleration limits. Experimental results on a linear motor system demonstrate the effectiveness of the algorithms in generating s-curve motion profiles.
1) The document discusses the application of S-shaped logistic growth curves to model and forecast technological trends over time.
2) It specifically fits a logistic curve to data on annual TRIZ publications from 1996-2006 to illustrate the three parameters (κ, α, β) of the simple logistic model.
3) The ceiling parameter κ represents the expected maximum number of future publications, the growth period parameters α and β specify the linear-like growth phase, and tm indicates the midpoint time of the symmetric S-curve.
ISO/IEC 27001, ISO/IEC 42001, and GDPR: Best Practices for Implementation and...PECB
Denis is a dynamic and results-driven Chief Information Officer (CIO) with a distinguished career spanning information systems analysis and technical project management. With a proven track record of spearheading the design and delivery of cutting-edge Information Management solutions, he has consistently elevated business operations, streamlined reporting functions, and maximized process efficiency.
Certified as an ISO/IEC 27001: Information Security Management Systems (ISMS) Lead Implementer, Data Protection Officer, and Cyber Risks Analyst, Denis brings a heightened focus on data security, privacy, and cyber resilience to every endeavor.
His expertise extends across a diverse spectrum of reporting, database, and web development applications, underpinned by an exceptional grasp of data storage and virtualization technologies. His proficiency in application testing, database administration, and data cleansing ensures seamless execution of complex projects.
What sets Denis apart is his comprehensive understanding of Business and Systems Analysis technologies, honed through involvement in all phases of the Software Development Lifecycle (SDLC). From meticulous requirements gathering to precise analysis, innovative design, rigorous development, thorough testing, and successful implementation, he has consistently delivered exceptional results.
Throughout his career, he has taken on multifaceted roles, from leading technical project management teams to owning solutions that drive operational excellence. His conscientious and proactive approach is unwavering, whether he is working independently or collaboratively within a team. His ability to connect with colleagues on a personal level underscores his commitment to fostering a harmonious and productive workplace environment.
Date: May 29, 2024
Tags: Information Security, ISO/IEC 27001, ISO/IEC 42001, Artificial Intelligence, GDPR
-------------------------------------------------------------------------------
Find out more about ISO training and certification services
Training: ISO/IEC 27001 Information Security Management System - EN | PECB
ISO/IEC 42001 Artificial Intelligence Management System - EN | PECB
General Data Protection Regulation (GDPR) - Training Courses - EN | PECB
Webinars: https://pecb.com/webinars
Article: https://pecb.com/article
-------------------------------------------------------------------------------
For more information about PECB:
Website: https://pecb.com/
LinkedIn: https://www.linkedin.com/company/pecb/
Facebook: https://www.facebook.com/PECBInternational/
Slideshare: http://www.slideshare.net/PECBCERTIFICATION
Thinking of getting a dog? Be aware that breeds like Pit Bulls, Rottweilers, and German Shepherds can be loyal and dangerous. Proper training and socialization are crucial to preventing aggressive behaviors. Ensure safety by understanding their needs and always supervising interactions. Stay safe, and enjoy your furry friends!
How to Manage Your Lost Opportunities in Odoo 17 CRMCeline George
Odoo 17 CRM allows us to track why we lose sales opportunities with "Lost Reasons." This helps analyze our sales process and identify areas for improvement. Here's how to configure lost reasons in Odoo 17 CRM
The simplified electron and muon model, Oscillating Spacetime: The Foundation...RitikBhardwaj56
Discover the Simplified Electron and Muon Model: A New Wave-Based Approach to Understanding Particles delves into a groundbreaking theory that presents electrons and muons as rotating soliton waves within oscillating spacetime. Geared towards students, researchers, and science buffs, this book breaks down complex ideas into simple explanations. It covers topics such as electron waves, temporal dynamics, and the implications of this model on particle physics. With clear illustrations and easy-to-follow explanations, readers will gain a new outlook on the universe's fundamental nature.
This slide is special for master students (MIBS & MIFB) in UUM. Also useful for readers who are interested in the topic of contemporary Islamic banking.
This presentation includes basic of PCOS their pathology and treatment and also Ayurveda correlation of PCOS and Ayurvedic line of treatment mentioned in classics.
Strategies for Effective Upskilling is a presentation by Chinwendu Peace in a Your Skill Boost Masterclass organisation by the Excellence Foundation for South Sudan on 08th and 09th June 2024 from 1 PM to 3 PM on each day.
Physiology and chemistry of skin and pigmentation, hairs, scalp, lips and nail, Cleansing cream, Lotions, Face powders, Face packs, Lipsticks, Bath products, soaps and baby product,
Preparation and standardization of the following : Tonic, Bleaches, Dentifrices and Mouth washes & Tooth Pastes, Cosmetics for Nails.
This presentation was provided by Steph Pollock of The American Psychological Association’s Journals Program, and Damita Snow, of The American Society of Civil Engineers (ASCE), for the initial session of NISO's 2024 Training Series "DEIA in the Scholarly Landscape." Session One: 'Setting Expectations: a DEIA Primer,' was held June 6, 2024.
Macroeconomics- Movie Location
This will be used as part of your Personal Professional Portfolio once graded.
Objective:
Prepare a presentation or a paper using research, basic comparative analysis, data organization and application of economic information. You will make an informed assessment of an economic climate outside of the United States to accomplish an entertainment industry objective.
Introduction to AI for Nonprofits with Tapp NetworkTechSoup
Dive into the world of AI! Experts Jon Hill and Tareq Monaur will guide you through AI's role in enhancing nonprofit websites and basic marketing strategies, making it easy to understand and apply.
How to Build a Module in Odoo 17 Using the Scaffold MethodCeline George
Odoo provides an option for creating a module by using a single line command. By using this command the user can make a whole structure of a module. It is very easy for a beginner to make a module. There is no need to make each file manually. This slide will show how to create a module using the scaffold method.
1. PRODUCTION AND OPERATIONS MANAGEMENT
Vol. I. No. 4. Fall 1992
Prinrcd in U.S.A.
EXPLORING THE LIMITS OF THE TECHNOLOGY S-
CURVE. PART I: COMPONENT TECHNOLOGIES *
CLAYTON M. CHRISTENSEN
Harvard University Graduate School of Business Administration,
Boston, Massachusetts 02 163
The technology S-curve is a useful framework describing the substitution of new for old
technologies at the industry level. In this paper I use information from the technological
history of the disk drive industry to examine the usefulness of the S-curve framework for
managers at thefirm level in planning for new technology development. Becauseimprove-
ments in over-all disk drive product performance result from the interaction of improved
component technologies and new architectural technologies, each of thesemust be monitored
and managed. This paper focuses on component technology S-curves, and a subsequent
paper, also published in this issue of the journal, examines architectural technology S-
curves. Improvement in individual components followed S-curve patterns, but I show that
the flattening of S-curves is a firm-specific, rather than uniform industry phenomenon. Lack
of progress in conventional technologies may be the result, rather than the stimulus, of a
forecast that the conventional technology is maturing, and some firms demonstrated the
ability to wring far greater levels of performance from existing component technologies
than other firms. Attacking entrant firms evidenced a distinct disadvantage versusincumbent
firms in developing and using new component technologies. Firms pursuing aggressive S-
curve switching strategies in component technology development gained no strategic ad-
vantage over firms whose strategies focused on extending the life of established component
technologies.
(INNOVATION; MANAGEMENT OF TECHNOLOGY, TECHNOLOGY S-CURVE;
DISK DRIVE INDUSTRY; COMPONENT TECHNOLOGIES)
The technology S-curve has become a centerpiece in thinking about technology
strategy. It represents an inductively derived theory of the potential for technological
improvement, which suggeststhat the magnitude of improvement in the performance
of a product or processoccurring in a given period of time or resulting from a given
amount of engineering effort differs astechnologies become more mature. The theory,
depicted in Figure 1,statesthat in a technology’s early stages,the rate of progress in
performance is relatively slow. As the technology becomes better understood, con-
trolled, and diffused, the rate of technological improvement increases(Sahal 1981).
But the theory posits that in its mature stages,the technology will asymptotically
approach a natural or physical limit, which requires that ever greater periods of time
or inputs of engineering effort be expended to achieve increments of performance
improvement.
* Received June 1991; revision received September 1992; accepted November 1992.
334
1059-1478/92/0104/0000$1.25
Copyri&t0 1992, Production and Operations Management Society
2. LIMITS OF THE TECHNOLOGY S-CURVE: I 335
Time or engineering Effort
FIGURE 1. The Technology S-Curve
Foster ( 1986) used S-curves to explain ageneral phenomenon of the sort observed
by Cooper and Schendel ( 1976) and Henderson ( 1988)-that radically new tech-
nologies are frequently developed and brought into an industry by entering firms,
rather than by the incumbent leaders. Foster cites the tendency of leading firms to
reinforce and refine maturing technological approaches and their failure to spot new,
successortechnologies in a timely way as a primary reason why leading firms lose
their positions of industry dominance.
The unit of analysis in most published studies of technology maturity and tech-
nology S-curves has been at the industry level. For instance, Roussel ( 1984) looked
at foam rubber; Constant ( 1980) examined aircraft engines; van Wyk, Haour, and
Japp ( 1991) studied permanent magnets; and Foster ( 1986) used examples from a
range of industries. My purpose in this paper is to summarize a body of theoretical
and empirical research, much of it relating to patterns of technological progress in
the disk drive industry, to enrich our understanding of the usesand limits of tech-
nology S-curve theory from thepoint of viewofa managerwithin a single$rm . Given
that S-curve phenomena have convincingly been shown to exist at more aggregate
levels, I explore in this paper and its companion paper (Christensen 1992b), which
is also published in this issue of the journal, the strengths and shortcomings of S-
curve theory when managers use it within individual firms to plan technology de-
velopment. I have summarized the data sources and the methodologies employed
to collect and analyze it in Appendixes A and B.
Drawing upon analyses of the disk drive industry, I offer four propositions about
the usefulness of technology S-curves to managers of technology development and
suggestthat the insights drawn from studying the disk drive industry may be arche-
typical of a broader range of industries whose products are complex assemblies of
components:
1. At the industry level, using a high-level measure of product performance-the
recording density of magnetic disk technology, in the case studied here-S-curves
can provide rather convincing explanations of why alternative technologies have
made or have failed to make substantial inroads against currently dominant tech-
nology.
2. To achieve improvements in the sorts of high-level measuresof system perfor-
mance mentioned in #1 above, managers must conceive and execute a sequence of
projects to improve the component technologies used in a product and to refine or
revamp the architectural system design within which the components operate. For
an engineering or research manager, therefore, technology S-curves will be opera-
3. 336 C. M. CHRISTENSEN
tionally useful if they aid in planning component and architectural technology de-
velopment programs.
3. When used to assesscomponent technologies’ improvement trajectories, S-
curves may be useful in describing an individual firm’s experience, but the framework
hasserious shortcomings if used in aprescriptivesenseto indicate the direction future
research programs ought to take. The levels at which individual firms perceived
component technologies to have plateaued differed across firms by nearly an order
of magnitude. The industry’s leading incumbent firms were generally the most ag-
gressivein switching to new component technology S-curves,but there is no evidence
that they gained any sort of strategic advantage over firms that stayed longer with
conventional componentry. If anything, a strategy of extending or “riding” the S-
curve of conventional technology and of switching component technology S-curves
behind the industry’s component technology leaders seemsto have led to greater
success.
4. In the disk drive industry, the technological changes in which attackers have
demonstrated strategic advantage (Foster 1986) have been architectural in nature.
Established firms find these technologies difficult to spot because alternative archi-
tectures are often initially deployed in historically unimportant commercial appli-
cations. Typical S-curve frameworks in which a new technology S-curve rises from
beneath and intersects the performance obtainable from mature technologies tend
to frame architectural innovation only in technologicalterms. In reality, architectural
technology change involves an intense degree of market innovation, in addition to
technological innovation. I propose an alternative S-curve framework for assessing
architectural change, one which embraces both aspectsof such technologies.
This paper supports the first three of these propositions; the fourth is discussed in
the subsequent paper (Christensen 1992b). This paper is divided into three principal
sections. In the first, I summarize key concepts treated in earlier studies of techno-
logical innovation and briefly outline the technological history of the disk drive in-
dustry. In the second, I evaluate the usefulness of S-curves in assessingthe potential
for performance improvement of magnetic recording technology versus other tech-
nologies at an industry level. In the third section, I examine the value and limits of
S-curve frameworks to managers in planning a sequence of projects to develop new
component technologies.
Context of this Study
I define technology for the purposes of this study asa process,technique, or meth-
odology-embodied in a product design or in a manufacturing or service process-
which transforms inputs of labor, capital, information, material, and energy into
outputs of greater value. Building upon the work of Sahal ( 1981)) I define a tech-
nological change as a change in one or more of such inputs, processes,techniques,
or methodologies that improves the measured levels of performance of a product or
process.Technology defined in this way is specific to particular products or processes.
As such, it is distinct from knowledge, whose value may not be unique to specific
products or processes.Definitions of the technical terms related to disk drives used
in the following discussion can be found in Appendix B.
The vertical axis of technology S-curves is constructed to measure an important
dimension of product or processperformance. Choice of the units measured on the
horizontal axis generally reflects the purpose of the author (O’Brien 1962). Scholars
whose objective is to measure the relative efficiency or potential productivity of
4. LIMITS OF THE TECHNOLOGY S-CURVE: I 337
development teams’ efforts generally measure engineering effort along the horizontal
axis (Foster 1986). Those attempting to assessthe impact of differences in techno-
logical maturity on product salesor competitive position often measure time on the
horizontal axis (Becker and Speltz 1983; Roussel 1983; Thomas 1984).
Although many of the researchers cited herein simply report observations of S-
curve phenomena, a few examine processesof technology maturation in considerable
depth. For example;Foster ( 1986) suggeststhat the leveling of a technology’s tra-
jectory of improvement is attributable to limits imposed by fundamental facts of
nature. Foster supports this explanation of maturity with examples from several
industries-one of which is the substitution of steam for wind-powered ships: he
shows that the speedof wind-powered vesselswas inherently limited by the physics
of wind and water. Constant ( 1980) explored a single industry and technology at
much greater depth, showing how the substitution of turbojet technology for piston
engine technology in the aircraft industry proceededthrough a seriesof asynchronous,
discontinuous improvements in the performance of individual materials and com-
ponents. Sahal ( 1981) essentially offers a theory of technology maturity: he posits
that the rate of performance improvement achievable within a given technological
approach declines becauseof scale phenomena (things either get impossibly large or
small) or because of system complexity. Because either of these problems makes
further progressmore difficult, Sahal suggeststhat the only way to maintain the pace
of progress is through radical system redefinition.
Rigid disk drives are an interesting product category to which S-curve analysis
might be applied. The industry has been characterized by a high degree of techno-
logical turbulence since IBM invented the first disk drive at its San Jose, California,
laboratories in 1956. In investigating how new technologies emerged and substituted
for maturing ones in this industry, I employ the typologies of technological change
proposed by Henderson and Clark ( 1990). Architectural change involves a rear-
rangement of the way in which components (whose fundamental technological basis
remains unchanged) relate to each other within a product’s system design. Modular
innovation is a fundamental change in the technological approach employed in a
component, where the product architecture is fundamentally left unchanged. Incre-
mental change refers to ( 1) improvements in component performance that build
upon the established technological concept or (2) refinements in system design that
involve no significant changes in the technical relationships among components.
Radical innovations involve both a new architecture and a new fundamental tech-
nological approach at the component level.
At the architectural level, seven distinctively different architectural technologies
captured a double-digit share of market units at some point between 1960and 1990.
And at the component level, there were innumerable incremental technological ad-
vances, aswell asseveral modular or “competency-destroying” ones (Tushman and
Anderson 1986) in the heads, disks, actuators, motors, and controller software or
firmware that constitute the drive. At the architectural and component levels, this
hasbeen an industry in which the strategic management of technology seemsto have
been an extraordinary challenge. Over 130firms entered the world disk drive industry
between 1960 and 1990-firms ranging from such vertically integrated computer
giants such as IBM and Fujitsu, to venture capital-backed start-ups. Leadership in
this industry has been tenuous: in the merchant or original equipment manufacturer
(OEM) disk drive market, an entrant company emerged to lead five of the seven
architecturally defined product generations.
5. 338 C. M. CHRISTENSEN
Rigid disk drives are comprised of one or more rotating disks-polished aluminum
platters coated with magnetic material-mounted on a central spindle. Data is re-
corded and read on concentric tracks on the surfaces of these disks. Read/write
heads, one each for the top and bottom surfaces of each disk on the spindle, are
aerodynamically designed to fly a fraction of a micron over the surface of the disk.
They generally rest on the disk’s surface when the drive is at rest; “take off” as the
drive begins to spin; and “land” again when the disks stop. The headsare positioned
over the proper track on the disk by an actuator motor, which moves the heads
across the tracks in a fashion similar to the arm on a phonograph. The head is
essentially atiny electromagnet whose polarity changeswhen the direction of electrical
current passing through it changes. Becauseopposite magnetic poles attract, changes
in polarity of the head orient the polarity of the magnetic domain on the disk’s
surface immediately beneath it, resulting in a sequence of positively and negatively
oriented domains. In this manner, data is written in binary code on the disk. To
read data, the drive uses changes in magnetic field on the disk as it spins beneath
the head to induce changes in current flow, essentially the reverse processof writing.
Disk drives also include electronic circuitry enabling computers to control and com-
municate with the drive.
As in other magnetic recording products, area1recording density (measured in
megabits per square inch of disk surface area or mbpsi) is the pervasive measure of
product performance in the disk drive industry. A drive’s total capacity is calculated
by multiplying the total available square inches on the top and bottom surfaces of
the disks mounted on the spindle of the drive by its area1recording density.
An Aggregate, Industry-Level View of Technological Maturity
in Magnetic Rigid Disk Drives
Figure 2a charts the average area1density of all disk drive models introduced for
sale by all manufacturers in the world between 1970 and 1989. The pace of im-
provement has been remarkably steady over this period, averaging 34% per year;
with time asthe horizontal metric, no S-curve pattern of progress is yet apparent.
In spite of this progress, radically different technologies such as bubble, optical,
and flash memory-which actually or potentially have offered greater recording den-
b
FIGURE2. Historical Improvements in the Area1 Density of New Disk Drives (Densities in Millions
of Bits per Square Inch)
6. LIMITS OF THE TECHNOLOGY S-CURVE: I 339
sity, speed, or reliability-have loomed threateningly on the horizon of economic
and technological competitiveness for years. The approach employed by S-curve
theorists, such as Foster ( 1986), offers an explanation for why magnetic recording
technology has held competing approaches at bay for solong. Foster notes that even
though it may be natural to think of technological improvement in terms of an
annual rate, it is engineering e&t, not time, that causestechnology to improve.
Foster therefore urgesthat the horizontal axis of S-curves should measure engineering
effort. Constructing an area1density S-curve with a proxy for engineering effort on
the horizontal axis, rather than time, can indeed help us understand why alternative
recording technologies have not yet significantly displaced magnetic disk memory
technology in spite of repeated forecasts of its demise (e.g., Drexel, Burnham, and
Lambert 1985).
Figure 2b shows that what appeared in Figure 2a as a relatively constant rate of
improvement overtime in areal density appears instead to be an increasing rate of
improvement per unit of engineering e@rt applied. (Because accurate measures of
industry engineering effort are unavailable in public sources, I have measured total
industry revenue on the horizontal axis of Figure 2b as a proxy for effort. The per-
centage of industry revenues devoted to research and engineering has not changed
significantly.) Foster ( 1986) contends that during such periods of increasing returns
to technology development effort as are shown in Figure 2b, the performance of
alternative techniques rarely surpassesthat of established technologies. S-curve theory,
when used in a descriptive or predictive mode at the industry-level, would suggest
that it is only after the industry’s technological productivity has reached its zenith
at the S-curve’s point of inflection, that magnetic disk recording technology might
begin to be vulnerable to alternative approaches.
As a descriptive or predictive theory, S-curves such as these can be helpful in
understanding more thoroughly the dynamics of technologically competitive envi-
ronments, at least at an aggregate,industry level. Similar analyses such asRoussel’s
( 1984) for foam rubber; Tchijov and Norov’s ( 1989) for computer-integrated man-
ufacturing (CIM ) technologies; and van Wyk, Haour, and Japp’s( 1991) for permanent
magnets seemto provide useful insights about the potential of alternative technologies
at an industry level.
After drawing upon the S-curve’s descriptive power at an aggregateindustry level,
a number of writers have advocated the use of S-curves as a firm-level prescriptive
guide in the strategic management of technology. Becker and Speltz ( 1983) and
Foster ( 1986)) in particular, seemto draw strong prescriptive implications for man-
agersfrom industry-level observations. Figure 3 shows the essenceof these prescrip-
tions. These authors urge strategists to identify when the S-curve of the technology
they currently employ has passedits point of inflection, to identify new approaches
that are rising from below at a more productive rate and that may in the future
intersect with the current technology, and to launch efforts to acquire or develop the
new technology in time to switch to it when its performance surpassesthe capabilities
of the present technology. In other words, prescriptive S-curve theory would have a
firm follow the dotted line in Figure 3.
Although this framework seemssensible, studies of technology maturity to date
have not empirically addressed how managers at the firm-level might use S-curve
analysis as a guide in the strategic management of technology development within
their individual firms. As I show in the following analysis, the view from the trenches
is more ambiguous than aggregateviews.
7. 340 C. M. CHRISTENSEN
Product
Performance
Time or Engineering Effort
FIGURE 3. Prescriptive S-Curve Strategy
Questions of whether and when a radically new technological approach such as
optical storage or flash memory will intersect with the magnetic recording S-curve
are important long-term strategic questions for disk drive company managers. On a
month-to-month and year-to-year basis,however, technology managers in disk drive
companies are not occupied with such high-level questions. The smooth performance
improvement trajectories mapped in Figure 2 are a summary manifestation of myriad
Figure 3-type technology extensions and substitutions at lower component and ar-
chitectural technology levels. Engineers manage improvements in over-all product
performance by interactively affecting the capabilities of componentsand by refining
or overhauling the product’s architectural design. These are the sorts of technology
planning decisions that dominate the technology manager’s planning calendar. To
keep up with the industry’s relentless pace of improvements in recording density,
technology managersmust monitor improvement trajectories of present and potential
architectural technologies and the extent to which individual component technologies
constitute an actual or potential bottleneck to the continued improvement in the
recording densities of their disk drives. They then must conceive a setand sequence
of architectural and component technology development programs that, when suc-
cessfully executed and integrated, will keeptheir firms in the competitive technological
race charted in Figure 2. When using the firm as a unit of analysis, therefore, one
must assessthe value of S-curvesin planning component and architectural technology
development.
As an example of how S-curves might be used asa framework for planning com-
ponent technology change, consider the role of read-write head technology change
in driving disk drive system performance improvement. In conceptual framework
of Figure 3, there were numerous incremental improvements to the original ferrite
head technology, which enabled manufacturers to grind the heads to smaller, more
precise dimensions. Such improvements were the drivers of performance along the
first technology curve. Thin film heads represented a second, modularly different
technology, which displaced ferrite heads in most models between 1979 and 1990-
analogous to the way in which the second curve intersects with the first in Figure 3.
Magneto-resistive heads have recently emerged, representing a third, fundamentally
different approach to head design, illustrated in concept by the third curve. This
sequence of technology substitution is typical of what has occurred with each com-
ponent and architectural technology over the past 30 years. The driver of the smooth
8. LIMITS OF THE TECHNOLOGY S-CURVE: I 341
progression of performance at the system level has been extensive technological tur-
moil at the component and architecture levels.
Analyzing technological maturity at the component and architectural levels is
important in firm-level analysessuch asthis not only becauseboth sorts of innovation
can be sources of system performance improvement but because component and
architecture are relative, not absolute concepts. For example, a read-write head can
be viewed at one level as a complex system architecture, comprising component
parts and materials that interact with each other within an architected system. At
the next level, the head is a component in a disk drive, which itself is a complex
architected system, composed of a variety of components. At a yet higher level, the
disk drive is a component in a computer, in which a central processing unit, semi-
conductor memory, rigid and floppy drives, and input-output peripherals interact
within a designed architecture. And finally, such a computer is itself a component
in an information processing system architecture, comprised of the computer, soft-
ware, operators, applications, sourcesand usesof data, and soforth. These constitute
asort of nestedsystemofarchitectures.System performance at any given level within
a nested system such asthis is generally driven not only by innovations at that level
but by improvements in component performance and architectural design at lower
levels in the system.
For these reasons, in this paper and its companion (Christensen 1992b), I focus
first on the use of S-curve theory in planning component technology development
and then examine the value of S-curve theory in guiding plans for architectural
technology development.
Using S-Curves to Prescribe Development of New Component Technologies
Whether S-curves can be used to guide the planning of component technology
development is important because, to borrow bank robber Willie Sutton’s phrase,
“That’s where the money is.” In the enterprise of disk drive researchand engineering,
asshown by Christensen ( 1992a), component technology development often begins
with fundamental research questions, passesthrough applied research and product
design and development, and endsin extensive processengineering. The development
of thin film headsat IBM alone took longer than a decadeand cost over $300 million,
and the industry spent well over $1 billion on thin film disk development. Devel-
opment of new product architectures, on the other hand, at most cost developers a
few million dollars and often consumed lessthan a year of calendar time.
Evidence from the substitution patterns of successive waves of new component
technologies in disk drives suggeststhat using S-curve analysis asa basisfor prescribing
new component technology development programs can be problematic at several
levels. In the discussion that follows I describe in detail the substitution of new-
technology thin-film read-write heads for ferrite heads and of thin film disks for
particulate oxide-coated disks as an example of the difficulties individual managers
may confront when managing the switch from one component technology to another.
Similar accounts could be constructed for every other significant component tech-
nology in the industry’s history. I then present evidence that despite thesedifficulties,
and contrary to Foster’s (1986) observation that attacking firms often seize the
advantage when new technologies invade a market, the incumbent firms quite
capably used new component technologies to pre-empt competitive attacks into
existing markets.
9. 342 C. M. CHRISTENSEN
TheEmergenceof Thin Film Head and Disk Technologies
The physical size of the head’s electromagnet is a critical factor affecting a drive’s
recording density. Heads were traditionally built by coiling fine copper wire around
tiny, precision-ground cores of ferrite. For reasons of cost, reliability, and certainty
of supply, engineers generally had strong incentives to continue using conventional
ferrite heads as long as possible. As engineers sensed they were approaching the
physical limits of how small ferrite cores could be machined, however, they began
efforts in the early 1970s to use thin-film photolithography-a process used in in-
tegrated circuit manufacturing-to create much smaller, more preciseelectromagnets
on the heads.
Disks historically were coated with microscopic particles of magnetic metal oxide.
Efforts to improve density within the particulate oxide approach involved making
the particles smaller and more uniform and dispersing them so that the maximum
possiblesurfaceareaon the disk wascoated with magnetic media. When disk engineers
felt they had reached the limits of fineness, uniformity, and dispersion, they too
turned to thin-film deposition technology, attempting to coat substrates with ex-
tremely thin, continuous coatings of metal.
There were greatdifferences in perceptions-within firms and acrossfirms-about
whether and when thin film heads and disks needed to be substituted for the estab-
lished ferrite and oxide technologies. Both types of ambiguity are illustrated in Figure
4, which charts the experiences of two of the industry’s leading competitors, Fujitsu
and Control Data Corporation (CDC), as they wrestled with the switch from ferrite
to thin film head technology and from particulate oxide to thin-film disk technology
asmeans for achieving greater area1density. Control Data was the largest American
supplier of disk drives to OEM computer manufacturers throughout the 1970s and
(Densities in Millions of Bits per SquareInch)(Densities in Millions of Bits per SquareInch)
Area1Area1
DensityDensity
40 -
10 - ,
I
8 - I
I
I
6. ,
,
5 -
4 - 'I -
/
3. /'
/
- Control Data
A---A Eujitsu
1
7676 77 78 79 80 81 82 83 84 85 86 87 8877 78 79 80 81 82 83 84 85 86 87 88
FIGURE4. S-Curves for Ferrite/Oxide Technologies at Fujitsu and Control Data Corporation
10. LIMITS OF THE TECHNOLOGY S-CURVE: I 343
1980s with a market share in several of those years exceeding 60%. Fujitsu was the
largest Japanesemarker of disk drives from 1977 to the present. Figure 4 castsboth
firms’ performance with ferrite heads and oxide disks in an S-curve format, where
the maximum area1density of models introduced in each year is measured on the
vertical axis and time is charted on the horizontal axis. (Since this analysis involves
comparing two firms’ technical progress over time, rather than assessingthe pro-
ductivity of engineering efforts targeted at two different technologies, I have charted
time rather than engineering effort on the horizontal axis.) Note that for each firm
there appear to have been two, not one, ferrite/oxide S-curves. What accounts for
the first plateau of ferrite/oxide technology and its subsequent second wind?
Apparently, according to the industry participants I interviewed for this study,
both firms launched development efforts for thin-film heads and/or disks just prior
to the onset of the plateau--cDc in about 1977 and Fujitsu in 1980. Both firms’
projects encountered a range of unforeseen problems, however, and neither could
introduce thesecomponents according to their original plans. With no technological
alternatives, their only choice was to wring additional performance from the ferrite/
oxide approach while they scrambled to get thin-film components ready. Both firms’
engineers met this challenge with astounding success,pushing area1densities with
ferrite/oxide technology to about triple the level at which each seems initially to
have planned to abandon ferrite/ oxide technology.
The proximate cause of the temporary plateaus in Figure 4 seemsto have been
that engineering resourceswere reallocated: in both instances, thesefirms scaledback
the engineering effort targeted at ferrite heads and oxide disks, betting that ferrite/
oxide technology was nearing its limit and that thin film was a key to future system
improvement. The time-measured plateaus in the area1density achieved with ferrite
heads may have been “induced” by the appearance of the alternative thin-film ap-
proach, which relieved the pressure-and usurped the resources-to push conven-
tional technology further. In other words, the very forecast that the conventional
technology was approaching its natural limit may in fact have been the proximate
cause of a leveling in the technology’s improvement trajectory, becauseof the impact
the forecast had on the allocation of engineering resources. Whether the 30 mbpsi
plateau Fujitsu achieved in 1987 represents the “real” natural limit of recording
density achievable with ferrite heads and oxide disks or is simply a self-fulfilling
prophecy that the future belongs to thin film we may never know.
Steele ( 1983) examined this phenomenon, noting that executives and engineers
alike often become enamored with radically new technologies-we might call them
technological long shots-as solutions to product performance plateaus. Steeleshows
that these long shots generally require far more time and money to develop than
originally believed and that most progressis achieved instead through the incremental,
steadyadvance of conventional technology. The forecastarrivals of such technological
long shots asgallium arsenide, optical disk memory, and ceramic engines have been
delayed or preempted by the steady cumulation of incremental improvements to
conventional technology. The casespresented here support the proposition that there
can be far more latent performance potential in a conventional technology than
individual firms or industry experts may perceive.
The innovations that enabled the second burst of performance improvement for
each of these firms were of the incremental sort defined by Henderson and Clark
( 1990). For example, three important incremental technologies advanced the per-
formance of ferrite heads. A modified barium-doped ferrite material wasdeveloped,
11. 344 C. M. CHRISTENSEN
which had the strength to beground to thinner dimensions. Lapping processescapable
of making smaller and more precise heads supplanted grinding. And finally, placing
a strip of metal in the gap separating the leading and trailing sections of the head
proved to strengthen the magnetic field created by the head. Similar incremental
improvements were made in the size, uniformity, and dispersion of oxide particles
that coated the disks. The cumulation of these developments in ferrite heads and
oxide disks pushed performance far beyond what initially had been viewed aslimits.
An explanation of why Fujitsu and CDC perceived limits to be at such different
levels is that nobody knows what the natural, physical performance limit is in complex
engineered products, such as disk drives and their components. Since engineers do
not know what they may discover or develop in the future, since the physical laws
(and the relationships between laws) governing performance are imperfectly under-
stood, and since possibilities for circumventing known physical limits cannot be well
foreseen, the natural or physical limits cited by scholars of technological maturity,
such as Foster ( 1986) and Twiss ( 1979), may in practice be moving targets rather
than immovable barriers. Foster ( 1986) cites sailing ships as an example in which
the physics of wind and water imposed a natural limit on the speedof sailing ships.
In retrospect, given what we now know, that is probably a true statement. But to the
designers of ships in the 1800s(and one might saythe 1990s)) the interactive physics
of wind and water were themselves being explored and defined by those practicing
the art of sailing ship design. Such limits are dynamic, relative, changing concepts
from the point of view of technology developers.
Even when designers confront an apparently immutable natural limit in a com-
ponent within a product system as complex as that of a disk drive, there may be
several engineering avenues for resolving the system-level performance problem-
there is often more than one way to skin the cat. Although one component’s per-
formance may be on a plateau-an actual or perceived physical limit-engineers
can continue to improve system performance by applying effort to lessmature ele-
ments of the systemdesign. For example, the discovery of run length-limited recording
codesin the mid- 1980scontributed to the resurgence of ferrite-oxide disk drive per-
formance, because it boosted the density of drives by 30%independently of which
disk and head technologies were employed in a model. In an even more dramatic
instance, Henderson ( 1993) has shown that the line resolution that could be achieved
with step-and-repeat photolithographic equipment by 1986 was substantially less
than the minimum possible resolution that had been calculated, in the early 1980s
to be constrained by the wavelength of light.
Becauseactual or perceived limits can be circumvented through advances in less
mature elements of a product’s design, when designs differ significantly acrossfirms,
perceptions of technological maturity may be highly firm specific.Whether aparticular
component technology is perceived to be a bottleneck to further improvement and
whether viable alternatives exist for circumventing such a bottleneck may depend
upon firm-specific characteristics of a product’s design. Such firm specificity is ap-
parent in Figure 4. Note that CDC’S initial “limit” for ferrite/oxide technology was
about 4 mbpsi-one third the density at which Fujitsu encountered its initial “limit.”
Figure 5 illustrates the extent of these across-firm differences in perceptions about
the limits of density achievable with ground ferrite heads and particulate oxide-
coated disks across a wider range of firms. On the horizontal dimension, it shows
that the timing of switching to thin-film technology differed among leading firms by
a decade-a very long time in such a turbulent industry. On the vertical axis, Figure
12. LIMITS OF THE TECHNOLOGY S-CURVE: I 345
Arcal Density (mbpbi)
.,OOO,OOOr
60,000
._--_. - - _
10,000I- :*,.I P#.9 fl
8,000
I-
.-I
*. Ix.. . . . . . . z4
Y
*:.
.:..
.*
6,000
t
1 1 1 1 I 1 1 I 1 I 1 I -
78 79 80 81 82 83 84 85 86 87 88 89 90
FIGURE5. Points at Which Thin-Film Technology was Adopted by Leading Manufacturers, Relative
to the Capabilities of Ferrite/Oxide Technology at the Time of the Switch
5 showsthat the densities to which the late-moving firms had pushed the conventional
technology were an order of magnitude beyond the levels achieved by the first movers.
The chart shows that there was a long, close race between the conventional and new
technologies before thin film finally triumphed.
The solid S-curve fit through the black dots in Figure 5 tracks the industry-average
area1density for drives using ferrite heads and oxide disks between 1975 and 1990.
The dashed line above the industry S-curve charts the highest density available in
ferrite-oxide drives in each year. Note that densities at this upper performance en-
velope were generally twice that of the industry average. The dotted line which is
just slightly above the ferrite/oxide envelope represents the thin film envelope-the
highest density among all models using thin-film technology.
13. 346 C. M. CHRISTENSEN
The paired open circles connected by solid lines denote the points at which the
industry’s leading firms started their switch to the new component technology S-
curve by introducing their first product employing either a thin-film head or a thin-
film disk. The first open circle in each pair denotes the highest density the company
had achieved in a model using ferrite/oxide technology prior to its introduction of
thin-film components. The second circle in eachpair is placed at the density achieved
in its initial thin-film model. The figure depicts each firm moving from its highest
density conventional model to itsfirst thin-film model. As such, it appearsthat most
of the innovators were above the industry average curve to begin with. Each firm,
however, had a range of models with a range of densities-some above and some
below the industry average. The highest-density conventional models of a few of the
firms-Rodime, Hewlett Packard, Quantum, Seagate,and DEC,were actually below
the industry average.
Several features in Figure 5 merit comment. Only 5 of the 15firms shown actually
leapt above the ferrite/oxide envelope with their first thin-film model. Although
most achieved higher density in thin film than they had in ferrite-oxide, they usually
ended up within the range achievable with conventional technology when they
switched S-curves. Second, thin-film technology eventually triumphed only after a
decade-long battle with ferrite/oxide. Key engineering managers involved in this
race indicated that the conventional technology progressedfur further than anyone
had anticipated when thin-film technology was first recognized as a technological
alternative. Third, different competitors switched S-curves at different points. IBM
moved to thin-film technology when its ferrite-oxide capability had reached 3,500,OOO
bpsi in 1979. Hitachi and Fujitsu rode the conventional S-curve far longer and had
achieved 27 and 30 million bpsi, respectively-over eight times the performance IBM
seemed to have identified as the limit of the ferrite-oxide approach-by the time
they switched to thin film.
Finally and possibly most important, there is little evidence that the firms that
switched component S-curvesearly-in this caseIBM, Memorex, StorageTechnology,
NEC,CDC, and Rodime-enjoyed sustained first-mover advantages. I have shown
this ordinally in Figure 6. The horizontal axis in that chart marks the order in which
the leading firms adopted thin-film technology-IBM being first, Fujitsu being #15,
and so on. The vertical axis ranks the firms according to the area1density of their
most advanced model in 1989. There seemsto be no correlation between order of
adoption-and presumably the deeperexperience with the technology that leadership
might entail-and the density each was ultimately able to achieve. In fact, the com-
bined shareof the total world market held by the early adopters of thin-film technology
fell from 60% in 1981to 37% in 1989. The firms that switched curves later-Priam,
Micropolis, Miniscribe, Seagate,Hewlett Packard, Quantum, Toshiba, Hitachi, DEC,
and Fujitsu-saw their combined world market share rise from 10%in 1981to 33%.
Christensen ( 1992a) shows that the industry’s leading incumbent firms were con-
sistently the leadersin developing and adopting new component technologies. Entrant
firms that pioneered the useof new component technologies asa vehicle for achieving
improved product performance were rarely successful:entrants enjoyed no attacker’s
advantage. Many factors affect the successand failure of firms, only one of which is
component technology strategy. The point, however, is that switching to new com-
ponent technology S-curves early does not seemto have been necessaryor sufficient
for competitive successin this industry. In contrast, I show in the companion article
14. LIMITS OF THE TECHNOLOGY S-CURVE: I 347
1
2
3
4
.e 5
aP13
a.$ 6
411 7
%2
a s
vd3
-5 8
.&a
(ju g
Id 10
11
12
13
14
15 L
Hcwim Packard-
IBM
(62)
0
Hitachi
w 0
Diitd Equipment
m 0
controlData
* (53)
Rodimc
(45) 0
Quantum
0 (44)
Prism
(45) 0
Memorex
e-0.
seagate
storageTeehnolW l tm
(25)I I , , . I I I I
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Order of Adoption of Thin Pii Techoology
(1 = first; 15 = last)
FIGURE6. Relationship Between Order of Adoption of Thin Film Technology and Area1Density of
Highest-Performance 1989 Model (Numbers in Parentheses Indicate Area1 Density in Millions of Bits
per Square Inch)
to this paper (Christensen 1992b) that those firms which led the industry in switching
to new architectural technology S-curves enjoyed powerful first-mover advantages.
This suggeststhat the S-curve switching mode of planning for component tech-
nology development prescribed in Figure 3 may not be the manager’s only option.
Since industry-level technological maturity curves are aggregatesof the performance
achieved by many firms and since a product’s performance results from the complex
interaction of many different components and system design alternatives, individual
managers may have substantial leeway for extending the performance of established
component technologies before undertaking the risk and expense of developing and
employing new componentry based upon fundamentally different technological ap-
proaches.
There may not always be such wide differences of opinion about S-curve exhaustion
within an industry asare illustrated here. Indeed, in some technical questions, such
as whether to switch from one type of material to another, the natural limits of
performance may be relatively unambiguous, broadly known, and uniformly un-
derstood. In instances such asthe one above, however, in which a technology’s per-
formance results from exploiting some combination of broadly understood physical
laws and firm-specific, experience-basedknow-how, the shapeof perceived technology
S-curves may be unique to individual firms rather than driven by absolute laws and
15. 348 C. M. CHRISTENSEN
physical relationships. Descriptions and predictions based upon industry-level ma-
turity curves, therefore, need not and possibly should not be taken as prescriptions
of firm-level strategy.
Do Systematic Differences Exist in How Firms Respond to Potential
Maturity in Component Technology?
Given this ambiguity in whether and when to switch to new component technology
S-curves, it is instructive to examine whether leading disk drive manufacturers re-
sponded to that ambiguity in any systematic way-to seewhether some firms’ tech-
nology strategies might emphasize switching component technology S-curves, while
the strategies of other firms might emphasize extending the S-curves of current com-
ponent technologies. To do this, I compared data on changesin the component and
architectural technologies employed in each firm’s disk drive models with changes
in the performance of these models over the 1976-1989 period, to trace the means
by which each firm achieved performance improvement. Based on this analysis, I
found that some firms, such as IBM, tended systematically to rely upon frequent
switches to more advanced component technologies as a primary driver of perfor-
mance improvement. Other firms, such as Hewlett Packard, tended to rely upon
incremental improvements in established component technologies and upon refine-
ments in system design to achieve competitive performance improvements. Most
hrms’ tendencies were consistent over time, essentially reflecting conscious or de
facto technology strategies.
These findings emerged from a regressionanalysis, in which I estimated coefficients
to a multivariate equation that describesthe log of area1recording density asa function
of ( 1) the year in which a model was first shipped and (2) technologies employed
in that model for the components that most directly determine area1density-actuator
motors, disks, heads, recording codes, interfaces, and the basic architectural tech-
nologies. Data used to estimate these coefficients were the detailed product specifi-
cations for every disk drive model announced in the world between 1979 and 1990.
I coded the technologies used in these models by dummy variables: use of the early,
dominant technology was coded as 0, and use of a new technology was coded as 1.
In some components, there was a progression of new technologies over the period
studied rather than a single new replacement, technology. For example, modified
frequency modulation (MFM) recording codes were replaced by the sequence of
2,7RLL; 1,7RLL; and partial response, maximum likelihood ( PRML) codes. The
dummy variable technique used here measures the impact on density of each of
these technologies relative to the original technology ( MFM in the caseof code tech-
nology), rather than measuring density relative to the immediately prior technology.
Table 1 describes the variables included in this exercise, along with the coefficients
and t-statistics, which were estimated using multiple least squares.
The interaction between head and disk technologies wascaptured in an interaction
term in this specification. Interviews with industry technical experts suggestedthat
from an engineering perspective, this was the primary interaction that needed to be
measured to isolate the impact of changes in individual component technologies on
improvements in recording density. To gaugethe stability of the measured coefficients
over time, I split the sample into two periods, 1979-1986 and 1987-I 990. The coef-
ficients of the equation estimated from thesetwo subsetsof the population arereported
in the right-most columns of Table 1. To better understand possible differences in
strategic approaches toward new component technology, I first estimated the coef-
17. 350 C. M. CHRISTENSEN
ficients from the entire industry data base,and then estimated coefficients using only
the models introduced by specific firms.
Note in Table 1that since the dependent variable is the log of area1density, if the
coefficients for each of the explanatory variables are exponentiated and significant
interactions are taken into account, the result is the percentage improvement in
density associatedwith the use of each new technology. The improvement in density
not attributable to.modular changes in component technology or system architec-
ture-and presumably due to the combined impact of incremental improvements
in established component technologies and refinements in systemdesign-is captured
by the coefficient of the TIME variable. Its value in the total-period equation of
0.163 (which exponentiated is 1.18) indicates that of the 34% average annual rate
of improvement in area1density over this period, slightly more than half-18%-is
attributable to incremental improvements that cannot be traced to specific new com-
ponent or architectural technologies. The remaining 16% of the industry-average
34%annual rate of improvement wasthe summary impact of the modular substitution
of new component technologies, made component by component, model by model,
year by year, and firm by firm.
A way to visualize these measures of the impact of incremental and modular
modes of component technology development on the improvement of industry-
average area1density is in the framework of Figure 3. On average for the industry,
slightly more than half (53%) of the total improvement came from progress along
established component or architectural technology S-curves and slightly less than
half came from switching S-curves. Note that in the split-sample analysis the B,
coefficient for the TIME variable, a proxy for the contribution of incremental in-
novation to over-all improvement, was similar in the two periods.
The middle section of Table 1 presents the coefficients for the sequence of new
architectural technologies. When compared to the density of 14-inch drives with
equivalent component technology in the total-period sample, the smaller the form
factor, the greater the density. This is because smaller drives have more rigid com-
ponents; the head-disk assembly weighs less, so that it can be positioned more ac-
curately, with lessinertia, over more finely-spaced tracks; and there is lessvibration.
The 8-inch architecture enabled a 10%density improvement over the 14-inch drives;
5.25-inch drives had 20% higher areal density than 14-inch drives with equivalent
componentry, and 3.5~inch products enabled an 37% density increase over 14-inch
products, holding component technology and vintage of models constant. Unlike
the coefficients for the TIME variable, however, the coefficients for thesearchitectural
technologies declined in magnitude and statistical significance from the first to the
second periods in the split-sample analysis. This seemsto be the result, according to
industry experts, of cross-architecture learning about mechanical and electronic de-
sign. Designers of each successively smaller architecture reduced the part count sig-
nificantly by incorporating more functions that had previously been handled me-
chanically into the electronics of the drive. Designers of larger-architecture drives
were then able to incorporate these design insights into subsequent generations of
their 14-, 8-, and 5.25-inch designs.
The bottom section of the table showsthe coefficients for the head-disk interaction
term. This wasincluded to test the possibility that simultaneous adoption of modular
head and disk technologies could contribute synergistically to performance improve-
ment. Note that although the total-period interaction was negative and of marginal
statistical significance, the interaction in the first period was significantly negative,
18. LIMITS OF THE TECHNOLOGY S-CURVE: I 351
while it was significantly positive in the latter period. A possible interpretation of
this, which is consistent with information I obtained in interviews with company
engineers,is that in the early yearsof a component technology’s commercial existence,
the system-wide impact of incorporating it into a design is inadequately understood.
Designers who used thin film heads and disks together in the early years therefore
suffered a performance penalty-negative synergy-because the system-wide impact
of the components was unknown. As designers learned about these technologies,
they were able to optimize the design to capture the full benefits, so that use of thin-
film heads and disks created positive synergy in the second period. There was not a
collinearity problem in measuring the individual contribution of thin-film heads
and disks to increases in recording density. Thin-film disks penetrated much more
rapidly than disks, so that there were many models using ferrite heads with thin-
film disks. Conversely, there were also a number of firms which used thin film
heads with oxide disks.
The R* value of 0.91indicates that the variables included in the equation accounted
for 91% of the variation in the area1density of the 1,033 models in the total-period
data base.
To explore whether firms’ technology strategies might differ systematically along
the spectrum of relying upon incremental versus modular component technology
development asthe primary engine of systemperformance improvement, I estimated
the equation specified in Table 1separately for several of the industry’s leading firms.
Table 2 presentsthe results for two technology/performance leaders,IBM and Hewlett
Packard. The average area1density of the product lines of both firms increased at
approximately the same rate as the industry average over the period studied, 35%.
The coefficients for the TIME variable shown in Table 2 indicate, however, that the
source.sof that improvement were different for the two firms. For IBM, the coefficient
of the TIME variable, Br , was0.14, indicating (when the coefficient is exponentiated)
that IBM realized about 15%density improvement each year from incremental in-
novations. Since IBM’S total annual improvement in density was 35%, this means
that the other 20%annual improvement came from switching to the new component
technologies specified in the equation. In contrast, the Br coefficient for the TIME
variable in the Hewlett Packard equation was0.24. This means, when the coefficient
is exponentiated, that Hewlett Packard achieved a 27% annual rate of improvement
in density through incremental technological change alone, without relying on the
modular adoption of new component technologies. Adoption of new, modular com-
ponent technologies accounted only for 8% annual improvement (27 + 8 = 35%
over-all annual rate of improvement). Indeed, there were a number of component
technologies that IBM had adopted which Hewlett Packard never employed during
this period.
In other words, these two firms employed very different means to achieve nearly
identical’annual rates of density improvement: IBM derived most of its improvement
from switching to new component technologies in the “strategic leaps” tradition
noted by Hayes ( 1985), while Hewlett Packard’s performance improvement came
largely from extending the efficacy of technological approaches it already employed.
This statistical finding is consistent with the views of IBM and Hewlett Packard’s
technological strengths which industry engineering managers expressedin my inter-
views with them. IBM managers and their competitors uniformly viewed IBM’S tech-
nological strength as component technology development, while they saw Hewlett
20. LIMITS OF THE TECHNOLOGY S-CURVE: I 353
Packard’s technical team as masters in system design-as being able to wring more
performance from a given set of components than other firms in the industry.
Summary
Although technology S-curves seem to provide useful insights at an aggregate,
industry level about the potential for continued improvement of fundamentally dif-
ferent technologies, the application of this framework at amanagerial level to planning
component technology development seemsto be very ambiguous. In the disk drive
industry, it appears that the perceived flattening of a component’s performance tra-
jectory is for practical purposes a firm-specific phenomenon. In fact, it may be that
a slowdown in improvement is the result of forecasts that improvement potential
hasbeenexhausted and the resource allocation decisions that follow from that forecast.
Limits to performance improvement, while often clear in retrospect, are changing,
dynamic concepts in the world of the operating manager. Since there are many
different component and system technology levers to pull in the pursuit of perfor-
mance improvement (there is more than one way to skin the cat), even limits imposed
by widely understood natural laws have been circumvented. These options seemto
have created substantial leeway in the technology strategieschosen by different com-
petitors. Some have gotten most of their performance improvement by extending
the performance trajectories of existing component technology, whereas others have
followed a technology strategy of switching technology S-curves rather aggressively.
Although S-curve patterns in component technology progress clearly exist, there
was no clear evidence of any first mover benefits or “attackers’ advantage” (Foster
1986). Firms that switched late to new technology S-curves successfully matched
the product performance of the early adopters. In the industry over-all, it was the
leading, incumbent disk drive manufacturers that consistently led the industry in
switching to new component technologies. Would-be attackers, which entered the
industry employing new component technologies asa source of product performance
advantage were rarely successful: attackers seem to have been at a decided disad-
vantage in exploiting new component technologies. The second paper in this series
(Christensen 1992b) shows that the opposite caseis true at points of architectural
technology change. In that study, I show that architectural technologies also follow
an S-curve pattern of performance improvement and that first-movers and attacking
firms enjoyed a decided advantage over late-adopters and incumbent firms in ar-
chitectural technology innovation.’
i I thank Professors Rim B. Clark, Robert H. Hayes, and Steven C. Wheelwright of the Harvard
BusinessSchool; ProfessorsRebeccaHenderson and JamesUtterback of the Sloan School of Management,
MassachusettsInstitute of Technology; and the anonymous refereesfor invaluable guidance and suggestions
for improvement to earlier drafts of this paper. Any remaining shortcomings are my sole responsibility.
Appendix A. Data Sources and Research Methodology
I have taken the data about disk drives reported in this paper from a larger study of that industry
(Christensen 1992a). The products upon which I focused that study were rigid disk drives, a product
category including drives commonly labeled as Winchester disk drives, which have one or more nonre-
movable rigid (hard) disks hermetically sealed in the drive housing, as well as drives that employ packs
of removable rigid disks. I did not include floppy disk drives in the study. Data in this study essentially
drew information from three sources. The first was Disk/ Trend Report, an industry survey published
annually. The editors of Disk/ Trend collect from each firm participating in the world disk drive industry
their revenues and product shipments by “form factor” (disk diameter) and capacity and use that data
21. 354 C. M. CHRISTENSEN
to calculate the size of each product-market segment, as well as average pricing levels in each segment.
They also report the disk drive revenues of each firm and market sharesof the principal competitors in
each product-market segment. In addition, Disk/ Trend publishes detailed product performance speci-
fications and a listing of component technologies used in each mode1currently offered for sale by each
of the manufacturers. This listing includes the year and month of first shipment for each disk drive mode1
aswell as the list price for a majority of the models. In addition, the editors of Disk/ Trend allowed me
to draw additional data not published in the Report from manufacturers’ product specification sheetson
file in the Disk/ Trend archives. I usedthis data to identify the specific models in which eachnew component
and architectural technology was first used in the industry and to trace the patterns of diffusion for each
of these new technological approaches. By charting each firm’s revenues, by size and capacity of drive
over time, I could reconstruct the commercial fortunes of each firm in considerable detail. I gratefully
acknowledge the generous assistance of the editors and staff of Disk/ Trend Report during this project.
The second source of data for the study were trade publications, particularly Electronic Business
Magazine. I searched each monthly (and more recently, twice-monthly) issue of Electronic Business
since it was first published in 1976 for notes and articles about disk drive technology, the disk drive
industry, and firms participating in it. My purpose was to understand more completely the corporate
histories, organizational structures, and competitive strategies pursued by the competitors in the industry,
as well as to identify additional disk drive manufacturers that might not have been captured in Disk/
Trend Report (I found only one such firm). I combined this information with the data from Disk/ Trend
on the sources and patterns of diffusion for each new technology to analyze which types of firms tended
to pioneer the development and adoption of each new technology. This enabled me to determine the
patterns of commercial successand failure among different groups of firms. Comparisons of entrants
versus established firms (building on the work of Henderson and Clark 1990) proved particularly fruitful,
as did comparisons of firms whose corporate forms were different: venture capital-backed start-ups,
vertically integrated computer manufacturers, integrated firms that produced other magnetic recording
products, and horizontally diversified firms that produced other computer peripheral products, such as
printers and tape drives.
The third category of data used in the study was information from over 60 persona1interviews with
founders and key engineering and marketing executives associated with eight of the major disk drive
manufacturers: IBM, CW, Digital Equipment, Micropolis, Quantum, Seagate Technology, Miniscribe,
and Conner Peripherals. In addition, I interviewed executives at the three largest independent component
manufacturing firms: Komag, Read-Rite, and Applied Magnetics. I also interviewed other industry experts
and consultants. My purpose in these interviews wasto understand and reconstruct ascarefully aspossible
the managerial decision processesthat led to these firms’ decisions whether or not to develop and deploy
particular new technologies, whose importance to the industry was highlighted through work with the
data described above.
I gratefully acknowledge the support of the Harvard Business School Division of Research, whose
financial assistance made a study of this scope possible.
Appendix B. Glossary of Technical Terms
Actuator
The mechanism that positions the head over the proper track on the drive, The class of actuators that
hasbecome most commonly usedbecauseof its superior positioning ability is called a “voice coil” motor.
This operates on a principle similar to that used in telephones: an arm is moved in and out via electro-
magnetic forces. Voice coil motors have been made in linear and rotary designs, but the rotary design,
which works like the arm on a phonograph, has become the dominant design because it requires less
space.A much lessexpensive actuator mechanism is a stepper motor, in which a shaft rotates in discrete
stepsto new positions in responseto changesin the surrounding magnetic field. Stepper motors are much
lessexpensive than voice coil motets and were used primarily on low-capacity drives targeted to price-
sensitive markets. Torque motors and DC motors were also used on a limited number of models in the
low-moderate performance range.
Areal Density
The amount of information that can be stored in a square inch of disk surface, measured in megabits
per square inch (mbpsi). This is determined by multiplying the number of bits of information storable
along a linear inch of track (bit density) by the number of tracks per inch of disk radius (track density).
Disk
The round, rigid platter on which data is magnetically recorded. It is comprised of a substrate, typically
made of aluminum polished perfectly flat, coated with particles of magnetic metal oxide or thin metal
films. These magnetic coatings are, in turn, coated with lubricating and protecting materials.
22. LIMITS OF THE TECHNOLOGY S-CURVE: I 355
Drive
The computer industry’s term for the equipment that contains rotating magnetic media-reels of tape,
flexible (floppy) disks, or rigid disks-and that controls the flow of electronic information to and from
that media.
Embedded Servo System
Mechanical shocks, differential thermal expansion, and a host of other factors can affect the accuracy
with which an actuator ca’n position a head over a particular track on a disk. Low-performance drives
using stepper motor actuators got around this problem by spacing the tracks far enough apart that such
subtle changes and misadjustments rarely caused the head to be mispositioned over the wrong track.
High-performance drives, however, require a closed-loop feedback system to the actuator, so that the
head can continuously be repositioned precisely over the proper track on the disk. This enables much
greater track density. One way of keeping precise head-disk alignment was to dedicate one complete
surface of one disk on the spindle to tracking information only. The head reading information off that
track and feeding it back to the actuator motor provided such a closed-loop, continuous-adjustment
mechanism. In an embedded servo system, track identification markers are written (embedded) on each
individual track of each recording surface. This frees up for user information the entire surface that
otherwise would have been reserved for tracking information only.
Ferrite
A magnetic compound comprised of iron and oxygen. In disk drives, the primary use of ferrite
has been as the core material around which fine copper wires were coiled to form an electromagnet
in the head.
Head
A device that contains a tiny electromagnet, positioned on an arm extending over the rotating disk.
When the direction of current through the head changes, its polarity switches. Becauseopposite magnetic
poles attract, changesin the polarity of the head causean opposite change in the polarity of the magnetic
material on the disk as it spins immediately beneath the head. The head writes information in binary
code in this fashion. Heads read data in the opposite manner-changes in the magnetic flux field over
the disk’s surface as it spins beneath the head induce changes in the direction of current in the head,
reversing the information flow. In rigid disk drives, heads are aerodynamically designed to fly a few
millionths of an inch above the surface of the disk, they generally rest on its surface when the drive is at
rest, take off asthe disk begins spinning, and land when the disk stops again. Heads in floppy disk drives
generally do not fly but glide on the disk’s surface.
Interface
This refersto the electronic circuitry through which the drive and computer communicate. A description
of the differences among interfaces is beyond the scope of this paper. Originally, interfaces were custom-
written by each drivemaker for each customer. Although some standard interfaces such asSMDemerged
as &inch drives were used with minicomputers, the trend toward standardization was accelerated by
SeagateTechnology’s ST4 12 interface, which required that the rate at which the drive took data off the
disk was equal to the rate at which the drive could transfer data to the computer. While low-cost and
efficient, this effectively put a ceiling on the bit density of the drive. Subsequent interfaces such asSCSI
(used primarily with Apple computers); AT (used with IBM-compatible computers), and ESDI (used
primarily with engineering workstations) decoupled these activities. With these interfaces, the drive could
take data off the disk as rapidly as its designers wanted, cache it, and then transfer it to the computer as
rapidly asthe computer could accept it. This enabled much greater bit densities than had been possible
under the ST412 interface. Other interfaces used only on a limited number of models were IPZ-I, ZPZ-
2, and ANSI.
An acronym for modified frequency modulation, an early coding technique used in writing data on
disks, wherein a magnetic marker was placed on the disk to denote the beginning and ending of each
individual piece of information.
MIG Heads
An acronym for metal-in-gap, a version of ferrite head wherein a strip of metal was deposited in the
gap between the leading and trailing portions of the head. This strengthened the magnetic flux fields that
23. 356 C. M. CHRISTENSEN
could be created and sensedby the head, enabling data to be written and read on smaller domains on
the disk surface.
Oxide
The term used in the industry for particles made from a compound of oxygen and a magnetic metal,
such asiron, cobalt, and chromium. Oxide particles were used to coat mylar substrates to create magnetic
tape and floppy disks, and to coated aluminum disks used in rigid or “hard” disk drives. The oxide
particles are the media in which, through changes in the particles’ magnetic polarity, data is stored
magnetically. The particles are generally of an elongated, needle-like shape.
Photolithography
The manufacturing process through which a desired pattern of one material is applied onto another
substrate material. Typically, the substrate is first coated (by plating or sputtering) with the material
from which the final pattern is to be made. This is in turn coated with a light-sensitive monomeric
material, called a photoresist. A mask of the desired pattern is then held over the photoresist, and the
unmasked material is exposed to light, causing the exposed material to cure. The unexposed photoresist
is then washed away. Through a subsequent seriesof etching and washing steps,only the desired material,
in the desired pattern, is left on the substrate. Integrated circuits are built on silicon wafers, and thin film
heads are built, through photolithographic processes.
PRML
An acronym for partial response, maximum likelihood, a coding technique that has followed RLL and
MFM recording codes.
Recording Density
Seearea1density.
RLL
An acronym for run-length limited recording codes, which enable data to be written more densely
that was possible with MFM codes. Two versions of RLL codes have been used: 2,7 and 1,7.
Spin Motor
The electric motor that drives the rotation of the spindle upon which the disks are mounted. In l4-
and 8-inch drives the spin motor often was situated in the corner of the drive and drove the stack of
disks via a pulley. In the 5.25 and subsequent drive architectures, a flat, direct-drive “pancake” motor
was positioned beneath the spindle.
SpindIe
The shaft upon which one or more disks was mounted.
Stepper Motors
SeeActuators.
Thin Film
A continuous, very thin film (often only a few angstroms thick) of a material (often a metal) on
another substrate material. This is generally applied through a process called sputtering, in which a
substrate is placed at the bottom of a vacuum chamber. A target of the film material is then bombarded
with electrons, which dislodge ions of the target material. These ions float like a vapor in the vacuum
chamber and then gradually settle in a thin, continuous film on the surfaceof the substrate.This deposition
techniquk is one of the early production steps in the manufacture of integrated circuits and thin-film
heads. It is also the technique used to coat disks with very thin films of magnetic material.
Torque Motors
SeeActuators.
References
BECKER,R. H. and L. M. SPELTZ( 1983), “Putting the S-Curve Concept to Work,” ResearchManagement,
26, September-October, 31-33.
24. LIMITS OF THE TECHNOLOGY S-CURVE: I 357
CHRISTENSEN,C. M. ( 1992a), The Innovator’s Challenge: Understanding the Injluence ofMarket En-
vironment on Processes of Technology Development in the Rigid Disk Drive Industry, D.B.A.
dissertation, Graduate School of Business Administration, Harvard University, Cambridge, MA.
CHRISTENSEN,C. M. ( 1992b), “Exploring the Limits of the Technology S-Curve. Part II: Architectural
Technologies,” Production and Operations Management, 1, 4, 358-366.
CONSTANT,EDWARDW. ( 1980), The Origins ofthe Turbojet Revolution, The Johns Hopkins University
Press,Baltimore.
COOPER,A. AND D. SCHENDEL( 1976), “Strategic Responses to Technological Threats,” Business Ho-
rizons, 19, February, 6l-69.
DREXEL, BURNHAM & LAMBERT ( 1985), The Disk Drive Industry, Drexel, Burnham, and Lambert,
New York.
FOSTER,R. ( 1986), Innovation: The Attacker’s Advantage, Summit Books, New York.
HAYES, ROBERT( 1985), “Strategic Planning-Forward in Reverse?Harvard Business Review, 63, No-
vember-December, 190- 197.
HENDERSON,R. ( 1988), The Failure of Established Firms in the Face of Technological Change, Ph.D.
dissertation, Harvard University, Cambridge, MA.
HENDERSON,R. M. ( 1993), Of Life Cycles Real and Imaginary: The Unexpected Old Age of Optical
Lithography, mimeo paper, Sloan School of Management, Massachusetts Institute of Technology,
Cambridge, MA.
HENDERSON,R. AND K. B. CLARK ( 1990), “Architectural Innovation: The Reconfiguration of Existing
Systems and the Failure of Established Firms,” Administrative Science Quarterly, 35, (March),
9-32.
O’BRIEN, M. P. ( 1962), “Technological Planning & Misplanning” in Technological Planning at the
Corporate Level, J. R. Bright (ed.), Harvard University Press,Cambridge, MA, 73-97.
ROUSSEL,P. A. ( 1983), “Cutting Down the Guesswork in R&D,” Harvard Business Review, 61, Seg
tember-October, 154- 160.
ROUSSEL,P. A. ( 1984), “Technological Maturity Proves a Valid and Important Concept,” Research
Management, 27, January-February, 29-34.
SAHAL, D. ( 1981), Patterns of Technological Innovation, Addison-Wesley, London.
STEELE,L. ( 1983), “Managers’ Misconceptions About Technology,” Harvard Business Review, 61,
November-December, 133-140.
TCHIJOV, I. AND E. NOROV ( 1989), “Forecasting Methods for CIM Technologies,” Engineering Costs
and Production Economics, 15, August, 323-389.
THOMAS, L. J. ( 1984), “Technology and Business Strategy-The R&D Link,” Research Management,
27, May-June, 15- 19.
TUSHMAN, M. L. AND P. ANDERSON( 1986), “Technological Discontinuities and Organizational En-
vironments,” Administrative Science Quarterly, 31, 3, 439-465.
TWISS,B. ( 1979), Management of Technological Innovation, Longman, London.
VAN WYK, R. J., G. HAOUR, AND S. JAPP( 1991), “Permanent Magnets: A Technological Analysis,”
R&D Management, 34, October, 30I-308.
EDITOR’SNOTE: An earlier combined version of this paper and the following paper published in this
issuewon the 1991 William Abernathy Award for the best paper in management of technology.