McKinsey | When Things Get Complex: Complex Systems, Challenges and Where to Focus
2 likes1,279 views
Georg Doll, a senior expert at McKinsey Digital, emphasizes the increasing complexity and challenges in the automotive software landscape, predicting significant growth in the software market essential for modern vehicles. He highlights the importance of a holistic approach to manage the interdependencies and complexities arising from trends like automation, connectivity, and electrification in automotive development. Successful organizations must focus on development tools, program management, and talent management to address these challenges effectively.
McKinsey | When Things Get Complex: Complex Systems, Challenges and Where to Focus
- 1. CONFIDENTIAL AND PROPRIETARY Any use of this material without specific permission of McKinsey & Company is strictly prohibited Georg Doll, Senior Expert, McKinsey Digital Munich Complex Systems, challenges and where to focus When things get complex
- 2. McKinsey & Company 2 Georg Doll Senior Expert at McKinsey Digital Munich Georg Doll, Senior Expert at McKinsey Digital in Munich As Member and Co-Lead of the EMEA Software Service line at McKinsey Digital, he supports clients along the product development LiveCycle. With his background of over 20 years’ experience in Automotive and embedded Software delivery and management of international teams. He supports clients in Market introduction planning, introduction of systems engineering, improving in project execution excellence and agile software development and talent management. He has served Tier1s, semiconductor vendors and vehicle manufacturers around the world in Japan, Asia Pacific, EMEA and US. In 2009 he was instrumental in setting up the GENIVI Alliance, and served as member of the Strategy Council and the Board for several years.
- 3. McKinsey & Company 3 Abstract Software is on the rise. Software is the no. 1 topic in the development of new functions. The software market is expected to grow from today to 2030 with an average CAGR of ~10%. So what could go wrong? A closer look at the four major automotive trends shows that they depend on the success of software. OEMs and Tier1s have recognized the situation and are investing heavily in software. To the extent that some speak of a "software defined car" or a "computer on wheels". But where there is light, there is also shadow. Highly automated driving, connectivity, powertrain electrification and new mobility services introduce additional dependencies between functions. Dependencies that – as we know continue to increase the system complexity. The development of software is a constant challenge for projects. What seems trivial at first glance turns out to be a much greater challenge than many people expect. Falling productivity, increasing communication, declining quality, constantly rising development costs and project delays are omnipresent. These challenges can only be tackled with a holistic approach. Successful organization have mastered the most important dimensions. Development tools, program management and talent management are just some of this dimensions.
- 4. McKinsey & Company 4 Software is on the rise.. and it is complex.
- 5. McKinsey & Company 5 Linus Torwald Principal developer Linux Kernel The value of Software is not in the code, its in the head of the people who developed the code Christoph Grote SVP Electronics at BMW Today 95% of innovation in automotive is software based Olla Källenius Chairman of the Board Daimler AG and Mercedes-Benz AG To stay relevant we have to control the Software in our vehicles
- 6. McKinsey & Company 6 63 76 85 50 81 30 44 63 92 129 15625 34 37 50 20 20 13 2020 238 2025 2030 362 469 Components 2,755 3,027 3,800Automotive sales USD billions 3% 7.0% Estimated CAGR 2020 - 30 Electronics ECUs/DCUs6% Integration, verification and validation10% Sensors7% Software (functions, OS, middleware)9% 3% Other electronic components (harnesses, controls, switches, displays) Power electronics (excl. battery cells)18%
- 7. McKinsey & Company 7 67.8 1.9 25 103.3 89.5 88.4 2020 25.7 2030 DCU ECU 91.5 156.2 129.0 ECU/DCU market USD billions Estimated CAGR 2020 - 25 21% 68% 3% 0% 6% 2020 - 30 92 156 92 362 469 238 2025 20302020
- 8. McKinsey & Company 8 4 Autonomous Connectivity Shared mobility Electrification
- 9. McKinsey & Company 9Idle stop/start Airbag deployment Adaptive front lighting Adaptive cruise control Automatic braking Electric power steering Electronic throttle control Electronic valve timing Head-up display Night vision Engine control Windshield wiper control Cylinder deactivation Blind spot detection Lane departure warning Electronic stability control Active vibration control Remote keyless entry Trans- mission control Seat position control Parking system Anti-lock braking Tire pressure monitoring Regenerative braking Hill-hold control Active suspension Active exhaust noise suppression Security system Navigation system Digital turn signals Electronic toll collection Lane correction Battery management Entertainment system DSRCCabin environment controls Active cabin noise suppression Voice/data communica- tion Interior lighting Auto-dimming mirror Event data recorder Driver alertness monitoring Accident recorder Instrument cluster Parental controls OBDII Active yaw control
- 10. McKinsey & Company 10 Eco functions bring the need for new sensors Highly automated driving increases functional dependencies Connectivity brings new features to the fleet ADAS functions increase safety levels of vehicle functions Connectivity introduces security threads Road safety requirements drive the need for new sensors Emission regulations drive need for new power train solutions Electrification brings new technologies into the car Eco functions require tighter function integration
- 11. McKinsey & Company 11 Productivity Complexity 2,25 3,00 0,75 2,75 2,00 1,25 1,00 2,50 1,50 1,75 Relative Change Infotainment Mission critical SW Industry Infotainment Mission critical SW Industry SOURCE: McKinsey Numetrics analytics on SW complexity and productivity in automotive • A widening gap between complexity and productivity levels is driving to performance and schedule challenges for automotive players • In order to overcome those challenges, SW development organizations need to both: Manage the increase in complexity Increase efficiency to enable management of complexity Productivity and Complexity relative growth over 10 years 2,4x 2x
- 12. McKinsey & Company 12 5 Unmanaged interdependencies between systems are developed as point-to-point interfaces leading to a high complexity and variety of interfaces within and beyond the system Point to point Interfaces 1 2 6 Portfolio Product Lifecycle 74 Multiple applications covering same functionality redundantly in the portfolio Overlapping functionality between components in the same system Multiple versions of the same application/ system are “live” at the same time Code size, quality, and documentation as further sources of complexity throughout the lifecycle … Description Operating system, HW complexity, and testing environment with strong influence on system complexity Components within a system are developed as monoliths impeding accessibility of single elements for updates/maintenanc e and integration within new development Functional redundancy Versions variety Code and documenta- tion quality Complexity dimension Multiple HW platforms Closed systems Multiple applications/ sub- systems within a SW platform are competing for similar resources (compute, storage, power) 3 Interfering sub-systems Multiple applications/ sub- systems within a SW platform are competing for similar resources (compute, storage, power) 3 Interfering sub-systems
- 13. McKinsey & Company 13 Interfering subsystems Isolated subsystems Main steps Sub-systems Test effort, required tests # Sub-systems Test effort, required tests # Large amount interfering systems test effort is rising exponentiallyA Small isolated systems test effort is rising linearB Test complexity per function: The smaller is the isolated system, the smaller is the exponential part of the test complexity Interfering sub-systems 3
- 14. McKinsey & Company 14 Vehicle Middleware layer/OS E/E hardware UI/UX/HMI Connectivity (back-haul) Artificial intelligence/Advanced analytics ActuatorsSensors Power components Cloud platform Applications
- 15. McKinsey & Company 15 Centralized Computing Collaboration of ECUs Isolated Functions Computer on wheels Cross-functional connection Domain- centralized Vehicle centralized Distributed 1st 2nd 3rd 4th 5th Domain controller Central gateway Body/comfort Chassis Powertrain Infotainment Central gateway Sensor Actuator Cockpit / Displays Central Compute COM Zone Zone Zone Zone Today
- 16. McKinsey & Company 16 How to standardize Software across my different product lines and product generations? How to transform 10.000 hardware-oriented developers to an agile-minded, software-driven organization? My software org is a black box to me. How do I assess and boost the embedded software dev. productivity of my 5.000 distributed developers and my suppliers? How can I ensure that my 1 billion USD software investment is delivered on time and on budget? How do I get access to the best software talent? How to transform our management systems to drive world class embedded software performance? How to organize software developers across my divisions SOURCE: McKinsey
- 17. McKinsey & Company 17 Top organizations show the potential increase in Software development performance for average and bottom quartile organizations SOURCE: Numetrics embedded SW project database Bottom quartile Average1 Top quartile 65 100 175 2,7x 65 100 224 3,4x 155 100 27 -83% Productivity Complexity units per man week1 1 Average indexed to 100 Complexity units per week1 Development throughput Residual design defects1 Quality
- 18. McKinsey & Company 18 Developer Velocity Index (DVI) Technology Deep structured interviews 100+ industry experts Working practices Comprehensive survey 440 large organizations across 12 industries and 9 countries Organizational enablement Statistical correlation analysis Business performance (financial performance, innovation, customer experience, brand, talent) against the various dimensions of DVI
- 19. McKinsey & Company 19 DVI is calculated as a weighted average of the scores for the 43 drivers across 3 dimensions Team characteristics Cross-functional teams Autonomous scope Co-location Limited context switching Product management Product manager/owner capabilities Product telemetry Rapid prototyping Clear product vision and requirement Linkage between strategy and team metrics Organizational agility Agile funding mechanism Portfolio management Dependency management Culture Collaboration and knowledge sharing Continuous improvement Culture of customer obsession Psychological safety/fail fast and learn Servant leadership Talent management Recruiting Employee value proposition Capability building Career path Performance management Team health management CI/CD practices Repeatable builds, continuous integration, delivery, and deployment Engineering practices Code reviews Coding guidelines Technical debt management Agile team practices Ceremonies Definition of done WIP management Open source, inner source Open source usage and contribution Inner source adoption Architecture Software architecture Data architecture Infrastructure and platform Public cloud adoption (IaaS, PaaS) Infrastructure as code Testing Test automation Test driven development Security and compliance Security practices Compliance practices Tooling Tools (planning, development, DevOps, collaboration) Al assistance Low code/no code Technology Working practices Organizational enablement
- 20. McKinsey & Company 20 1. Calculated using Johnson's Relative Weights: % importance is relative importance of driver on business outcomes. Total sums to 100%. Higher % indicates stronger impact on business performance 2. Average score for Innovation, Customer Satisfaction, Brand, and Talent Source: McK Developer Velocity Survey, Expert Interview Foundational drivers R2 = 0.6 N = 440 1.0% 4.0% 7.0% 2.0% 0% 6.0% 5.0% 3.0% 7.5% 0.5% 6.5% 5.5% 1.5% 2.5% 3.5% 4.5% Publiccloudadoption(IaaSandPaaS) Infrastructureascode Psychologicalsafety/"failfastandlearn" Testautomation Testdrivendevelopment Servantleader Security Compliance Collaborationtools Techdebtmanagement Developmenttools DevOpstools Collaborationandknowledgesharing Dependencymgmt Codereviews CI/CD Lowcode/nocode Producttelemetry Employeevalueproposition AIassistance Codingguidelines PMCapabilites Opensource Recruting WIPmanagement Definitionofdone Ceremonies Teamhealthmanagement Autonomousscope Limitedcontextswitching Fundingmodel Productvisionandrequirements Linkagestrategyandteammetrics Rapidprototyping Portfoliomgmt Continuousimprovement Innersource Customerobsession Crossfunctionalteam Capabilitybuilding Co-location Careerpath Planningtools Softwarearchitecture Dataarchitecture Performancemanagement Drivers relative importance1 on overall business performance indicators2 % Importance1 Infra & platform TestingArchitec- ture ToolingSecurity & Com- pliance CI/ CD Engineering practices Open source/ inner source Agile team practices Team characteristics Organizational agility Culture Talent managementProduct management Organizational enablementTechnology Working practices Softwarearchitecture Compliance
- 21. McKinsey & Company 21 A Developer tools B Product management (PM) capabilities C Culture D Talent management 1. Calculated using Johnson's Relative Weights: % importance is relative importance of driver on business outcomes. Total sums to 100%. Higher % indicates stronger impact on business performance 2. Average score for Innovation, Customer Satisfaction, Brand, and Talent Source: McK Developer Velocity Survey, Expert Interview Foundational drivers R2 = 0.6 N = 440 1.0% 4.0% 7.0% 2.0% 0% 6.0% 5.0% 3.0% 7.5% 0.5% 6.5% 5.5% 1.5% 2.5% 3.5% 4.5% Publiccloudadoption(IaaSandPaaS) Infrastructureascode Psychologicalsafety/"failfastandlearn" Testautomation Testdrivendevelopment Servantleader Security Compliance Collaborationtools Techdebtmanagement Developmenttools DevOpstools Collaborationandknowledgesharing Dependencymgmt Codereviews CI/CD Lowcode/nocode Producttelemetry Employeevalueproposition AIassistance Codingguidelines PMCapabilites Opensource Recruting WIPmanagement Definitionofdone Ceremonies Teamhealthmanagement Autonomousscope Limitedcontextswitching Fundingmodel Productvisionandrequirements Linkagestrategyandteammetrics Rapidprototyping Portfoliomgmt Continuousimprovement Innersource Customerobsession Crossfunctionalteam Capabilitybuilding Co-location Careerpath Planningtools Softwarearchitecture Dataarchitecture Performancemanagement Drivers relative importance1 on overall business performance indicators2 % Importance1 Infra & platform TestingArchitec- ture ToolingSecurity & Com- pliance CI/ CD Engineering practices Open source/ inner source Agile team practices Team characteristics Organizational agility Culture Talent managementProduct management Organizational enablementTechnology Working practices Softwarearchitecture Compliance
- 22. McKinsey & Company 22 Linus Torwald Principal developer Linux Kernel The value of Software is not in the code, its in the head of the people who developed the code Software is a people business The key success factors - Culture - Talent Management - Development Tools - Product Management
- 23. McKinsey & Company 23 1. Pycholocial safety (fail fast and learn) 2. Collaboration and knowledge sharing 3. Continues improvement 4. Servant leader Development Tools Product Management Culture Talent Management 1. Planning tools 2. Collaboration tools 3. Development tools 4. DevOps tools 1. PM Capabilities 2. Product telemetries 1. Performance management 2. Team health management 3. Capability building 4. Recruting
- 24. Thank You