Andreas Weigend | Social Data Revolution | Fall 2015
School of Information | University of California at Berkeley | INFO 290A-03

external image bmw-logo-AD930473AC-seeklogo.com.gif

Homework 5 - Intuitive Telemetry for BMW

by Abhi Sharma

The Problem (and why it matters, at least for BMW)

The automobile began as a status symbol--something people aspired to own. Later, especially in the United States, it became essential to navigate increasingly sprawling cities and affordably travel long distances. Especially among youth, though, it was a symbol of freedom and fun.
external image freedom.jpg
Generations of youth grew up as car enthusiasts alongside relatively affordable and fun cars like the BMW 2002 (less affordable but more fun Turbo variant pictured) and E30 M3. Among the premium German brands, BMW specifically has always been known for emphasizing the driver's experience; many of those young enthusiasts likely became repeat customers as they got older and had more disposable income.
external image 1974_BMW_2002_Turbo_For_Sale_Front_resize.jpgexternal image 640px-BMW_M3_E30_front_20090514.jpg
Today, affordable and fun cars are still around--look at the new M235i, heir apparent to the E46 M3. Much of their buyers are actually nostalgic older car enthusiasts, though. Young car purchasers, let alone enthusiasts, are becoming an increasingly rare breed.
external image number.jpg
external image dontwant-car.png&w=1484
This presents an obvious problem for BMW. Decades from now, how many people will be loyal repeat BMW purchasers? If the majority of today's youth rely on public transit or ridesharing services like Uber, only a minority will purchase automobiles of their own. Of that minority, an even tinier sub-minority would care enough about the driving experience to purchase a BMW over a more economical rival brand or even a fully autonomous car. How can BMW capture more of the future automobile market while continuing to offer a differentiated (driver-focused) product in an increasingly crowded (to the point of commodification) range of options?

The Hypothesis (the solution's main idea)

There are many ways to tackle this challenge, each of which incorporate at least one of two key strategies: an appeal to environmentalism and/or a conspicuous demonstration of onboard technology. Why these two specifically?

Although young people these days may not tend to care about cars as much, they are quite likely to care about protecting the environment. They are also increasingly fond of technology, to the point of not being able to imagine life without it.
external image Motif-Millennials-Impact-Investing-Risk-Reward-graph.pngexternal image sense-of-smell-millennials.jpg
This data merely confirms common sense; those who grew up with smartphones in a world marked by climate change would naturally hold both technology and the environment dear to their hearts. The key for BMW's driver-focused premium automotive offerings to stay relevant to these new consumers is to make both technology (in a very consumer-facing way) and environmentalism core brand values. This is already being done with the i3 and i8 (pictured) models. Both offer a stimulating experience for the driver while showing off an electronics-focused cabin and an energy-efficient hybrid system.
external image bmw-i8-inside-beautiful-2015-bmw-i8-interior-at-bmw-5.jpg external image 7433.jpg
Not every BMW can feasibly resemble the bleeding-edge i-models for now, however. The i3 is slightly larger than a 1-series but costs nearly as much as a 528i. The i8 is more expensive than any 7-series! To capture the interest of youth, all of BMW's models need to somehow exhibit consumer-facing technology and an eye towards efficiency--even the M range, at least to some degree. Existing onboard technology and equipment have to be used in a novel way.

Luckily, considering how computerized modern cars are, there is plenty of equipment to work with. GPS sensors, accelerometers, pedal force pressure sensors, rain sensors, wheel speed sensors, 360° cameras, and many more already offer pinpoint insight into what exactly is happening in, to, and around the vehicle. Surely if there is enough information available to enable completely unassisted automatic driving, the same information could be used to make manual driving itself more enjoyable and intuitive. Instead of romantically abandoning present-day technologies like electrical power steering, automatic transmissions, and traction control to provide a more stimulating driving experience, what if those technologies could be used to work with the driver instead of overriding his/her input? In addition to the data available locally to each individual car, what insights could be gleaned by pooling together the data of all the BMWs on the road nearby?

The Action (the implementation of the hypothesis)

All of those data sources can pooled together to effectively implement something comparable to but more powerful than Waze in every new BMW's vehicle dynamics and infotainment system. Waze, as it exists today as a smartphone application, relies in large part on manually input location-only data from users actively opening the app and reporting accidents/hazards to accurately determine traffic patterns and route users accordingly.
external image waze-ios-7-1024x846.png
If a similar data-sharing model was implemented passively in each new BMW, the built-in navigation system could find the fastest route and calculate a pinpoint-precise ETA (perhaps even based on the driver's driving style and the vehicle's capabilities in relation to current road conditions and legal limits) without any input from any user. Thanks to how much data each car collects today, though, that represents just the beginning of this method's potential. Nearly every aspect of the driving experience can be perfectly tailored to the driver based on his/her prior actions (indicated by the car's local data) and the current local environment (indicated by the data of surrounding cars). Think of each car's computer system as a supervised neural network, operating in parallel with its counterparts! Data can be shared between very nearby cars using peer-to-peer short-range radio or Bluetooth communication (comparable to racecar telemetry systems), while data can be shared with more distant cars or BMW corporate using wireless Internet.
external image bmw-active-cruise-control-3.jpg
Rolls-Royce (a subsidiary of BMW) has already brought to market a GPS-enabled automatic transmission that tailors shift patterns to the topology of the current road for an optimally smooth ride (e.g., downshifting in anticipation of an upcoming hill instead of downshifting upon encountering the hill and thus jerking the car slightly). If that GPS data was crowdsourced and combined with data collected by the local car about its owner's driving style, it could be used to develop an automatic transmission that would be able to predict and make the exact shifts you would manually--except without ever making a mistake or compromising on efficiency. The car and driver together would together be able to develop a novel yet efficient driving style even on unfamiliar roads. The same ideas can be applied to any of the car's electronically controlled manual inputs--variable-ratio steering, pedal force/travel, etc. Even non-manually-controllable but dynamic features of the car, such as air suspension, can dynamically adjust themselves in response to this continuous feedback loop. Overall, the automobile would "interpret its driver's driving style" in the most optimal (fast and efficient) manner appropriate for the current environment.
external image 201303-P90115731-zoom-orig-300x199.jpg
Though these ideas may at first seem like a continuation of the modern trend of taking agency away from the driver and giving it to electronic nannies, they can be implemented in such a way that they make the driver more connected to the machine; technology can make the driving experience more intuitive yet more efficient. For instance, the impact of modern rear-wheel steering systems can only be felt intuitively in the steering wheel with variable-ratio electronic power steering. In other words, BMW's core brand value of the "ultimate driving experience" can be maintained while promoting technology and efficiency to appeal to younger buyers present and future using this data-driven method. The beauty of this is that nearly all of it can be done using existing technology; it is just being used differently and exposed more to the end-user this way.

The Metric (how the success of the action will be quantitatively measured)

Two simple metrics would be sufficient for roughly/broadly measuring the impact of this data-driven system overhaul: the percentage of users (perhaps conditioned based on various qualitative variables, such as age or type of car owned) who disable or otherwise complain about the data-sharing and data-based driving enhancement features described above, as well as the percentage of users (conditioned once again) whose driving experience is materially improved by the driving enhancement features (i.e., their average speed has increased and/or their average fuel efficiency has increased since the features were enabled). These specific metrics would be measured using the cars' existing sensors and computers and sent to BMW corporate using the online data uplink capability mentioned above.

The Evaluation (how the solution will be judged overall)

Aside from the obvious interpretation of the metrics (the more users who keep the new features on and benefit from them, the better), the qualitative impact of the new data-driven drive-by-wire system can be measured by receiving honest and informal feedback from young BMW purchasers as well as longtime BMW enthusiast communities. Think "community managers" and "ears on the ground" over "focus groups." The end goal is to capture mind and market share for BMW among young people who may not have an initial interest in the driving experience while continuing to please the devoted minority of car enthusiasts.