Of Burning Teslas and Big Data

October 5, 2013
Tesla Model S

Tesla Model S “Tartare”

This week, Tesla Motors suffered what many called a PR meltdown as video appeared on the Internet of a Model S outside Seattle, WA on fire, streaming liquid from the undercarriage.   Message boards on enthusiast and investment sites, Twitter were quickly engulfed in their own flames as keyboard quarterbacks rushed to analyze what happened.  Certainly this isn’t the first vehicle to catch fire, in fact nationwide driving statistics show that roughly 150,000 vehicles a year catch fire.  The difference is this was the first Tesla electric vehicle.  And what happened in the following 48 hours, how Tesla Motors reacted was as different as their vehicle.

Yesterday at 3:44pm, an email caught my eye.  Entitled, “About the Model S Fire”, it was from Elon Musk, Chairman and CEO of Tesla Motors.  In it, the company explain in excoriating detail exactly what happened:

  • A metal object struck the Model S’ undercarriage while driving at highway speed
  • Acting as a lever, it impaled the vehicle’s 1/4in armor plate with a peak force of 25 Tons
  • The first of 16 battery modules was damaged and started emergency shutdown procedures
  • The vehicle’s onboard alert system notified the driver who was directed to safely pull over and exit the vehicle

The email and subsequent blog post goes into additional detail that explains factually the procedures that emergency personnel followed.  The last paragraph in the blog post is what really caught my eye:

The nationwide driving statistics make this very clear: there are 150,000 car fires per year according to the National Fire Protection Association, and Americans drive about 3 trillion miles per year according to the Department of Transportation. That equates to 1 vehicle fire for every 20 million miles driven, compared to 1 fire in over 100 million miles for Tesla. This means you are 5 times more likely to experience a fire in a conventional gasoline car than a Tesla!

Tesla Model S Telemetry captured via REST APIs

Tesla Model S Telemetry captured via REST APIs

How can Tesla possibly know how many miles have been driven by thousands of Tesla vehicles?  Big data.  The Tesla Model S captures an insane amount of data – acceleration and deceleration, voltage loads, steering, pressure,  in what equates to a flight data recorder which is connected to the company via a 3G wireless data connection.  This data is used by Tesla for everything from monitoring vehicle performance, wear and tear, and making adjustments via regular software updates delivered to each vehicle that actually improve the driving experience over time.

Somewhere in Silicon Valley I imagine Jerome Gullen, VP of WW Sales and Service at Tesla Motors and his boss, the enigmatic Elon Musk were sifting through data alongside their local investigation team in Seattle.  Together with their PR team, they crafted a response with brutal honesty and transparency, even posting (with permission) Jerome’s email exchange with the owner (who received a loaner vehicle by the way).

While Tesla Motors doesn’t report exactly what telemetry is sent to the company, like other auto manufacturers they capture deep logfile data via the CanBus, delivered locally to diagnostic equipment via the OBDII port or via a USB stick.   To give a sense of the type of data being captured on-board, see this post on the Tesla Motors Club site. For other developers out there, this just scrapes the surface of what’s captured.

This is how you handle a PR disaster – brutal honesty and transparency.  This is also why I’m a long-term investor in Tesla, and will be taking delivery of my Model S in November.