Neil Postman, the American author, critic and educator, died exactly twenty years ago from the writing of this article. “Technological change is not additive,” he scowled. “It is ecological. A new technology does not merely add something; it changes everything.” A few years after his death, iPhones hit the market and transformed everything from mobile data traffic to working styles to even (shockingly!) bubble gum sales.
And then almost a decade after Postman’s passing, the Model S hit the market and another transformative awakening hit the automotive market: software would eat vehicles as well. And not just embedded lines of code (LOCs) growing from 2.5 million in his 2003 sedan to an estimated billion LOCs 25 years later, but automotive manufacturers as software factories supporting “car-as-a-device” business needs.
Amongst thousands of industry professionals amassed at the MOVE America ’23 conference this past week in Austin, Texas, several experts including some from Mercedes and Porsche discussed how automotive manufacturers are having to become software companies.
“Right now, the customers have a choice, and the phone is better. Plain and simple,” states bluntly Frank McCleary, Partner and the North America Lead for Automotive and Financial Services within Porsche Consulting. “If [manufacturers] want to drive adoption in-vehicle systems instead of CarPlay or Android Auto, they have to improve significantly. And until they get the knack of making intuitive software, it’s going to be a challenge for them to truly accelerate. There are 4-5 OEMs that are committed to $20B in annual revenue on software-as-a-service on the vehicle. It’s a big business. They have to change ….”
“I think the biggest change is a mindset change within the OEMs,” states Jana Breitkopf, Managing Director & CEO of Mercedes pay USA LLC. “Coming from the analog world to actually embracing the digital world. And not just copying what’s out there or copying a pdf to an app. That’s not real transformation, just like adding a battery to a car doesn’t make it an electric vehicle.”
This requires a few massive steps for companies already struggling with overhauling their powertrains and business models: converting from physical architectures to software networks, building software factories and understanding the full data science landscape.
Converting From Physical Architectures
As Jim Farley, CEO of Ford, explained earlier this year, “We farmed out all of the  modules that control the vehicles to our suppliers because we could bid them against each other.” And that worked for decades. There certainly was some technological differentiation, but most manufacturers attempted to win via styling. But along came the need to control and wirelessly update code – for business models, cybersecurity and unfortunately a shifting safety culture – and suddenly that distributed control became problematic. “The problem is the software is written by 150 different companies. And they don’t talk to each other. And so even though it says ‘Ford’ on the front, I have to go to Bosch to ask permission to change their seat control software.”
Therein, creating and owning a cloud-to-cross-vehicle software network is a massive overhaul and enabler to a manufacturer wanting to affect quick change.
“It [has to be] a mini-data-compute center that’s running on wheels,” explained Siddarth Shah, Vice President of Engineering at Canoo.
Building Software Factories
To support that enormous software, these companies must grow organizations that have never existed previously and populate them with the cultural mindset alluded to by Breitkopf.
That’s a difficult road.
“If the leadership team has no understanding of what’s needed in the automotive industry, you’ll never be able to execute …,” Christoph Schroeder, Vice President of Software Development at Luminar Technologies, explained in a 2022 Forbes interview. “But then you need people who aren’t too used to ways in the automotive industry, so they don’t understand where everyone else is coming from. So there’s a balance between experience in the automotive industry and still enough flexibility mentally.”
The Full Data Science Landscape
An entire novel could be written about the ever-shifting landscape of data science. Obviously the recent evolution of Artificial Intelligence (AI) and how its associated data management will inform algorithms that will enhance vehicle functionality, will create winners and losers. Insurance programs and rates, while maintaining or adversely affecting privacy might drive customers one direction or another.
But interestingly, the discussion last week centered around the growing need to share data to learn quicker. “Currently there is no way to share data cleanly,” pronounced Shah. “You can curate it, create data lakes, and make it available, but there’s no standard that I know of where you can do it easily. There’s tremendous opportunity in that space, but it requires tremendous cooperation, which has not been a strong suit for the industry.”
“There’s so much that we can do now via telematics [for predictive maintenance],” explained Ken Johnston, CEO of Autonomic, a subsidiary of Ford. “The challenge that I have is a Big Data problem. [My commercial customer’s] fleet uses the same tires as everyone else, and I could [theoretically] predict the need for maintenance and have the correct parts onsite to protect uptime. How do we do that better? Even with 16 million vehicles on the road, without data from other places, it’s Small Data.”
Wise data management and sharing will assuredly affect the software-as-a-device businesses.
If you happen to work for an automotive manufacturer and you’re thinking “We got this,” may I suggest reexamining your hubris. Just last week, Stellantis recalled 273,000 Ram trucks for software incompatibility involving the backup camera and radio that “… could lead to a crash.” The week before that, spokespersons for Kia and Hyundai reported that “…more than 760,000 cars have been updated so far [and] distributed more than 280,000 steering wheel locks” to address an ongoing cybersecurity issue. And, oh by the way, right now I’m driving a truck that simultaneously instructs me to drive opposite ways around the same roundabout when using Apple CarPlay.
Meanwhile, on the exact same Ram recall day, Tesla pushed an over-the-air update enabling a new safety feature: auto-hazard-lights after a crash. Although this might save few lives, it’s another indication of the entrant idea-to-cloud-to-vehicle methodology versus the incumbents’ slow cleanups.
And so my message to you executives: don’t skimp. You don’t know what you don’t know, and catching-up to Tesla et al really isn’t about the electric vehicle. It’s remaking the foundational vehicle and development itself while trying to run at the speed of Silicon Valley.
Get help … now.