Everyone in the logistics industry is aware that supply chains are becoming more complex. On top of this increasing complexity, technology is playing a bigger role, as cars become more and more like drivable computers or smartphones.
The rise of the so-called ‘connected car’ – those equipped with internet access – doesn’t look likely to slow down anytime soon, with developments coming from all OEMs, and the involvement of tech giants, as seen in examples including the Google Car and Apple’s mysterious ‘Project Titan’.
All of this development means more software going into vehicles. A modern car now has more than 100m lines of code, with that figure expected to grow to 200m or even 300m over the coming years (compare that to the measly 145,000 lines of code that got Apollo 11 to the moon).
There are strict checks in all areas of production and logistics, but as OEMs and manufacturers add more software to vehicles, and install more software to help run the plants and processes that make vehicles, there is an increasing risk that it can go unchecked.
“The common feedback we get is that software should just work, whether a company is making it themselves or getting someone else to install it,” said Dr Martyn Jeffries, head of automotive solutions at SQS. “The same logic in manufacturing is that you wouldn’t just build a car and never check the parts. You can’t guarantee that it will work, and we’re trying to instil that in manufacturers.”
SQS tests software for other companies across a range of industries, including automotive, retail and banking. For the automotive industry, SQS is involved at different stages. “I usually divide it into three domains,” said Jeffries. “We work with the business, where an OEM is designing and engineering a car; operations, when they’re making and moving it; and product software, which is what goes into the car that makes it work – the connected car. The business aspect encompasses large enterprise systems, whereas the operational aspects are transformational, as in implementing a new system to offer a new capability or logistics system, such as warehousing.
Still a lot to learn
Tesla has already significantly changed the automotive industry, according to Jeffries, and there is a danger that OEMs could miss the next big thing. A software-centric company could come in, and outsource the manufacturing side of building a vehicle. “The old players have to be careful that someone doesn’t disrupt the system – especially with driverless cars and city transport,” Jeffries said.
Moving code, rather than cars
While logistics are generally quite efficient, when software is brought into the equation, it becomes more difficult. “Maybe three times before the car gets to the customer, there will be software updates. How do you manage all that?” Jeffries asked. He explained that the physical parts of a car will eventually become the least important parts. “The challenge will be when people are buying features that are not part of the original supply chain. How do you integrate connected car technology while keeping it up to date? If there are 100,000 cars around the world, how will you do that? The oily bits generally don’t go as wrong as often as they used to. For the warranty period, it’s as much about software as it is a door falling off,” Jeffries said.
Going forward, it seems that the challenges for the supply chain will become virtual, rather than the physical challenge of getting a vehicle or part onto a lorry and driving it somewhere. “The supply chain will become more about moving code. People now expect a fully customised car. You can choose everything, and while that’s adding complexity to the supply chain, there will be fewer physical parts, and much more software,” Jeffries said. Compared to a few years ago, many software companies are now bigger than OEMs. “OEMs are used to working from a position where they are the dominant player. They could say, ‘If you want to work with us you have to do this,’ but you can’t do that with Google. The dynamic will change,” Jeffries added.
Staying safe?
The complexity for testing software on this scale is inevitably more complex, but Jeffries explained that the difficulty is the connectivity of relatively simple parts, to different systems. “That’s where you get unexpected results. The danger is that people just test what they know will work. The big story with quality assurance is what if people are doing something they shouldn’t? Would a system still behave appropriately? Those unknowns, along with the negative and false tests, are what you have to prove to be secure and safe.” OEMs need to be careful that they aren’t pushing development too quickly in order to meet demand, which could lead to safety issues, recalls, and further logistics complications.
While it’s “no longer about the metal, and now about the code and the chips,” according to Jeffries, traditional cars will be made for a long time, and this is especially true in emerging markets, where traditional cars are being built in a different way, with less of a focus on software and connectivity.
That said, Jeffries was keen to reiterate that manufacturers need to keep one eye on the future. He forecast that the value of the vehicle’s physical parts would become lower and more commoditised. “It’s not how you do it, but how you do it profitably,” Jeffries said.
Suppliers start to make money when they’ve made a significant volume of cars, but with software, the costs come later. Maintaining the software ten years later would be difficult, and this is where tech companies have the advantage. “Metal bashing companies that are big today could find themselves being subservient to software companies, which could dramatically alter the supply chain,” Jeffries said.
Naturally, it’s harder to fix a software problem the further down the development process it is, so SQS likes to be involved early on. “We manage risk. Companies tend to avoid problems which could cost them more money later on,” Jeffries said. Testing can be 20% of the cost of developing software, so if a company hasn’t budgeted for that, it can be a bit of a shock. But in the end, a customer is not likely to be tolerant if their car doesn’t work properly.
Whether it is in or out of the car, functioning software throughout the supply chain is vital, and will only continue to grow in importance as cars become more connected. Whether it is managing inventories, benefiting warehouse management, improving visibility on the line, or connecting a car to the internet, like the nuts and bolts scanned out of the bins, software needs to be of a similarly high quality.
