Recent reporting from Warranty Week noted that after more than a decade of gradual decline the automotive industry in the US has for the past several years seen a steady increase in overall warranty expenses. Digging into the numbers,CX3 Marketing has been tracking the rise of software-related recalls in the US. According to Andy Gryc, co-founder of CX3 Marketing, “Software-related issues affected around 8% of recalled vehicles in 2017, with four times as many vehicles affected compared to five years prior.”

The stats are not comforting. As an industry we are getting worse, not better.

Figure 1 Courtesy CX3 Marketing/Strategy Analytics

Strategy Analytics’ Roger Lanctot told us “The annual cost of non-OTA updates have risen to as much as $6B in the US alone”

Additional insight from Stout Research confirms what we intuitively thought, increased software in the vehicle is adding additional recalls to every system in the vehicle. For example:

• In 2016 Nissan announced a recall of 3.3M vehicles in the US due to an error in the occupant detection system that could incorrectly classify an adult as a child and not deploy the airbag in a crash.
• In 2016 GM recalled 3.6M vehicles to due to a software error in an airbag sensing and diagnostic module caused the module to incorrectly enter a diagnostic mode that could stop deployment of the airbags in a crash.
• In 2017 Chrysler recalled 1.3M pickup trucks due to an error where a severe impact to the truck underbelly would cause the system to misdiagnose a sensor failure and would not deploy the airbags.

As I look at the three examples above, I can’t help but be very worried for the industry. Compared to the intelligence and required software complexity of even the most basic Level 2 Autonomy systems the crash detection and airbag deployment systems are relatively simple and should be very mature. If it were one automaker you might say it was a fluke. But with three you have to admit that while the automakers rush headlong into full autonomy they still need to work harder to get the software of the most basic safety systems right.

An additional challenge is getting the recalled vehicles actually fixed. Stout has found that historically recalled vehicles that are more than 5 years old have a fix rate of approximately only 40% within 9 months of the recall. That rate only increases to 70% if the vehicle is less than three years old. NHTSA has set a goal to get to 100% fix rate, but that seems unlikely without a dramatic change in the industry. For the ignition switch recall, GM has resorted to sending service crews door-to-door across the US. There simply is no other way to locate the vehicles.

Think about it – if an app is regularly crashing on my phone I get an update pushed to me right away. But if there is a software bug that could cause me to die in a crash the best I can hope for is that the automaker finds me by snail mail and I remember to schedule a warranty repair. Seems the technology deployment is backwards to the seriousness of the issue.

Adding full-vehicle OTA and connected diagnostics can save costs in many ways:

• For many updates, the update can be deployed directly with no additional service center costs
• Today service centers oftentimes have to remove and re-install electronic units just to update the software. The use of OTA at the service center can reduce this shop-time.
• Diagnostics and reporting can be used to locate and notify drivers of safety recalls, reducing what can be a very arduous process to track owners of older vehicles.

For OEMs to continue to add sensors and advanced features to vehicles and not get swamped by warranty expenses they need to add full-vehicle OTA like eSync now. Getting a POI update to track the latest Starbucks openings is nice, but how about you make sure that I don’t die first.