This is the first part of a two-part series. You can read the second part here.
Digitization is affecting all industries, but perhaps none quite so much as automotive. From autonomous cars, over-the-air feature enhancements, new conversational interfaces, right up to the fully transparent and connected car on your mobile. All of this is transforming traditional car manufacturing. It’s a breathtaking opportunity to shape these changes.
At the vanguard of this change are companies such as Tesla, Google, Apple, and Uber, tech companies trying to disrupt, to extend or to change the market. Meanwhile, the incumbents like Ford, Toyota, VW, Mercedes, and BMW are striving to become digital, agile and fast moving.
It’s hard to predict who will win the race — or even if there will be just one winner in the end. It’s not easier to guess who will gain market share or who will be a dominant player in a disrupted and fast-changing market. I like to believe, that there is value and asset in the size and the experience which the incumbents have. Perhaps I’m biased because I like their cars and their traditions. I favor European cars — to be honest — especially the German ones.
In these series of articles, we’ll explore the most common challenges we see in our daily work with automotive incumbents and look at the lessons we’ve learned. This first article lays the foundation for the series by analyzing the context of the automotive industry with our TOPS framework and by highlighting the corresponding challenges.
In the second article, we present ideas around how to improve the decoupling of hardware and software development. In the last two, we plan to address additional potential improvements, such as 'Power and Priority of Procurement' and the 'Culture of Agility'.
TOPS as-is analysis for connected car features on mobile devices
One of the most pressing challenges for automotive incumbents is in mastering the adoption of agile practices and culture that is common in the software and innovation industry. At Thoughtworks, we support these firms by helping them build digital products that communicate directly with the cars and also by building IT-solutions that support the ecosystems around them.
To describe our context thoroughly, we use TOPS — which stands for the four dimensions we look at when analyzing situations and suggesting improvements to our clients: Technology, Org-Design, Product-Design, and Strategy.
All elements and concepts of the software and innovation industry fall into one of those dimensions. All the elements we encounter can’t stand on its own, but they must be aligned to the others to generate the full value potential.Underpinning all the formal concepts there is the corporate culture, which highly impacts how people behave in that context.
Ultimately, it’s like playing Tetris. It is most likely that there is a legacy that you need to take into consideration. There is room to shape your business. And there are new technologies, methods, org-designs, lean product approaches and new business strategies. They must fit together — and you need to align them. Or you will lose the room to maneuver successfully.
Let’s put it to work.
Within the domain of connected cars and the digital products communicating with them, we usually face a high-level IT architecture that looks like this:
Customers and their devices are at the center and the focus of automotive incumbents these days (well, they’re on the left in the picture). Technically, those devices send out requests that travel via public telecoms infrastructure until they reach components that are controlled by the automotive enterprise. It may be public cloud infrastructure (e.g., for content delivery networks) or private cloud infrastructure (e.g., dedicated applications servers) and at some point in time they will touch the private core IT landscape (e.g. for car- and customer-data and perhaps for billing).
The core IT landscape usually features a connected-car-API or -backend, encapsulating security and communications with the car itself. The requests are passed on to that API traveling on a similar route away from the core systems, via public telecoms infrastructure until reaching the cars. Whatever needs to be done in the car is done, and an acknowledgment travels back the same route to the customer's mobile device.
Latency and distribution are big challenges in this setup. Customers can be anywhere in the world. Cloud infrastructure can be relatively close to them — but core IT landscape, up to now, is most likely centralized in dedicated data centers close to the enterprise’s home base. That has implications for companies operating on a global basis.Take Chinese customers, who represent a large and interesting market: the data transfer latency from China, to, for instance, a German central data center, to a car in China and back again is unacceptable. It can add up to four times the latency you typically have between China and Germany.
Another challenge are release cycles. Cars models are updated on multi-year cycles. And they stay the same for a very long time. Software gets often released up to hundreds of times a day and it changes very quickly. Synching those two worlds, especially during the development phase, is a major challenge. None the less, the benefits are substantial: in time-to-market and relevance of the new features. Downwards compatibility is also taken to an extreme here, because new versions need to support old cars — for a very long time (decades).
From an organizational perspective, we typically see siloed departments that separately take care of marketing and sales, corporate IT, group IT, engineering, testing, governance, and procurement. Departments span multiple locations and multiple legal entities.
The process frameworks that incumbents have in place are closely aligned to, and supportive of, the car engineering and production. Usually, that spans over multiple years and is heavily optimized for the production, sales and maintenance services of the physical cars, as well as the best practices and learnings around that.
Budgeting is siloed too, reflecting the strong departmental structure, and is as inflexible as the process framework for building cars. Budgets are attached to projects, which usually last a year or longer and are tied to the manufacturing of a car. It’s hard to do something that isn’t supporting a certain model and a certain car.
All of these characteristics are challenges for agile software development teams, where we prefer autonomous and cross-functional teams, that have end-to-end responsibilities for business results and which can work in a continuous and iterative way.
While incumbents want to adopt digital and agile ways of working, we see they have core product design processes which are tailored to physical products, which are hard to prototype and experience before they are available. With more and more digital products becoming relevant and driving revenues, the existing product design processes post some challenges.
For example, teams face pre-defined requirements that they’re asked to implement. These requirements often cannot demonstrate the value of the market and are based on inadequate assessments of user needs. Instead, Agile teams deploy hypothesis-driven and data-driven development of features and work autonomously on achieving business results.
Traditional processes lack direct and early feedback from end-users. In the digital world, teams can release features often and quickly and get immediate input from the market through low- / high-fidelity prototypes or A/B tests. Missing this feedback and their input increases the risk of introducing features that customers don’t want, and gambling with their loyalty.
Every company we talk to has a digitization strategy that challenges the status quo. Increasingly, business leaders understand that their business model and revenue streams will change; that digital brand experiences will add value to customers’ lives.
However, most senior managers face the challenge that there is a very large and inflexible organization serving them and their customers. That organization and people are formed around processes and techniques designed to engineering complex physical products en mass. The new digital brand experience, however, requires agility, flexibility; agile software delivery demands a close connection between the physical product and the software systems around it.
We’ve seen some common patterns emerge, as automotive firms try to adapt and change. An early and easy response to the challenge was to just 'call it agile' — and hope that the organization and the people will change following the name. But while dissolving operations departments and concentrating them in one unit with software development may seem like a valid approach to introduce DevOps, it’s insufficient. It’s the underlying culture that drives people’s behaviors. Hence the term 'Innovation Theater'.
Also founding a digital lab alone (Germans do that in Berlin) will not yield a business result by itself. We believe in creating an MVP within roughly four months, even in a complex enterprise setup. Once the MVP is there, we can test its value. This requires the experience, the culture and the means to have impact and deploy software to production quickly. Many of the required conditions are missing in those digital labs.
To some degree we also see top management focus on tough decisions, without taking the business value into consideration. Rather than thinking about a central or regional solution, we would ask: what will yield the greatest value?
Today there are powerful feature toggle frameworks available, that enable businesses to route users, depending on location and profile, to different features and version of the software. Rather than engaging in discussions around power and influence, it should be much more about business outcomes.
Summary TOPS as-is analysis & Outlook
We applied the TOPS framework to identify challenges and potential improvements along the formal areas of Technology, Org Design, Product Design and Strategy. The details are in the following table:
Decoupling hardware & software releases
CI / CD for systems
Distribution of IT components
Separate legal entities
Distribution across locations
Budgeting along physical products
Business value as objective for teams
Missing direct & early feedback from customers
Low- / high-fi prototypes
Digital labs without business outcomes
Central vs. local discussions instead of business value focus
Change of culture
Delegation of control
Alignment on vision and strategy
Value-based portfolio management breaking down investments and outcome
In future articles, we’ll look at three areas for automotive incumbents to address, where they can improve their current practices and move towards the best practices of the software and innovation industry. The three areas we will tackle are:
Decoupling of hardware and software — related to technology
(through digital twins, API as products and end-to-end test automation)
Power of procurement — related to Org Design and Strategy
(and how to manage agile IT service providers and contracts at scale)
Agile management culture — related to Strategy
Enterprises who succeed in addressing these problem areas, will benefit from the following:
Short time-to-market for new features and changes in their customer ecosystem
Relevant features for their clients — generating loyalty and revenue streams
A digital engineering brand to attract skilled employees and educated customers
Potential to scale out the transformation to more parts of the enterprise
Disclaimer: The statements and opinions expressed in this article are those of the author(s) and do not necessarily reflect the positions of Thoughtworks.