February 17, 2023
Is there anybody who still doubts the value of data? Highly unlikely. However, as with any asset, data needs to be properly made use of to realize the full potential. This means to find an answer to the questions of why a particular set of data is valuable, how this value can be extracted and what tasks the extraction involves. The question of extraction is not trivial, for it is not really data which is valuable, it’s the insights derived from it—similar to the difference between raw materials and a product.
In the automotive space in particular, there is more data available than ever before, for better or for worse. By 2030, nearly 95% of new vehicles sold globally will be connected, sharing data with devices inside and outside of the vehicle. The total volume of data transmitted will soon reach 100 petabytes per month.
This constant flow and high volume will lead to opportunities and challenges. One challenge is that data will have to be properly collected and filtered because most data on the vehicle is highly repetitive and for this reason not of much use. There needs to be a way to gain access to the information that is meaningful and actionable. The other aspect of this issue is that the software will have to be developed and maintained. The second challenge is finding the best use for any given corpus of data. To begin with, use can be divided into external and internal.
If an automotive OEM or a fleet management company generates data from fleets of vehicles, external use means to sell the data to third parties, that in turn apply the data to their own purposes. Examples would be aftermarket sales or deriving insights about traffic patterns. However, the external use of data is neither high on the priority list for many companies nor as easy as it may sound. Among the reasons for this low priority are concerns about data privacy (it is difficult to get customer consent to third-party use) and the risk of giving away valuable intellectual property.
The internal use of data is simpler, with questions of data privacy and explicit customer consent not posing problems. Internal use has many aspects, depending on sector and organization. On the one hand, data can serve to improve parameters like customer spend or the acquisition of new customers. Customer retention is a classic category for the application of customer insights, with targeted advertising, pricing, and products. It is important because it is much cheaper to retain an existing customer than to acquire a new one; thus such efforts need to be as calibrated as possible.
On the other hand, there is an internal use of data which is not customer directed and instead has purposes such as reducing cost and adding value to products and services. At times it is even possible to develop new products and services based on insights derived from real-world use cases. This aspect will become more important and better targeted with further improvements on the information and analytics front.
Examples for these uses are plentiful. For cost reduction, there is reductive design, with the goal to change or simplify features that aren’t used or to eliminate them altogether. Today, data allow this the first time, revealing how people actually drive and use their vehicles. It may turn out that a feature designated as very desirable in market research sees little use in daily driving and thus should be improved upon or sometimes plainly eliminated. How and how often, for example, are features like four wheel drive low or the modes for teen drivers or valets really used?
Further use cases that will reduce cost for manufacturers (and often for drivers as well) involve the early detection of issues. A certain sensor parameter that is outside its defined limit can be an indicator of a problem or impending failure. To observe this allows timely maintenance and sometimes even a correction via over-the-air updates before a problem worsens. Thus, the necessity and cost of returning the car to the shop might be eliminated. Instead a virtual mechanic diagnoses software-related issues and failures (and sometimes even rectifies them), while also performing or suggesting preventive maintenance, which may, in turn, reduce downtime.
While the options of added value for existing products are numerous, some of them profit from data-driven insights in particular. New features and upgrades may be introduced based on information from real-world driving, similar to reductive design. The beneficiaries of such features may be the driver directly, the vehicle (benefitting owner/driver in a proximate way), or the OEM. Whatever the category and whoever the beneficiary, the added value over the vehicle life cycle can be considerable. The table shows only a small fraction of the benefits that will become available.
These value-adds demonstrate that there is a shift in the products and services—away from traditional sales and aftermarket, towards services and subscriptions. Examples here are mobile apps that will see continued development, with possible interfacing to third-party services that may make features, such as in-car deliveries, possible by providing temporary vehicle accesses to the delivery service if the driver is not present. With machine learning (ML) and artificial intelligence (AI), a car may now know the driver’s preferences and provide for these, even in yet unknown environments. For example, the navigation system might suggest a similar coffee shop in a different town, or AI can predict where the driver is going and inform them of traffic patterns. Memory-based parking and self-parking will be attractive options; in general, drivers will perceive optimized and more convenient processes valuable. The whole model of vehicle ownership may change because data can reveal how a certain driver treated a vehicle, what range is available, or how the vehicle was left. A driver who wants to share their vehicle would consider the option to specify what kind of drivers to share with certainly valuable.
The corollary question is what is necessary to get to insights and customer improvements and subsequently to reduced cost or added value. To realize these steps, three areas of specialization have to be covered: product, data, and analytics. However, these specializations are often not conjoined in one organization. Because of this, collaboration and partnerships will lead to optimum results.
What has to be provided in the categories of data and analytics? OEMs and tier 1 suppliers will need a place with tools to collect data, update software, and send remote commands. Ideally, this would be a platform that is comprehensive but user friendly as well as safe and secure. Sibros’ Deep Connected Platform (DCP) enables all this over the air, with the functionalities available via a Web Portal and APIs.
The DCP works with major cloud providers, such as GCP, AWS, and Azure; it can be applied to purposes as different as to develop new applications or to reduce end-of-line testing costs. Analytics tools such as Google BigQuery can be used to drill down into data with great granularity.
The combination of OEM expertise with software/data analytics capabilities yields manifold advantages: deeper insights, less cost, better product offerings, improved ROI and profits, reduced loss and waste, shorter time to market, better driver experience, and more sustainability, among others. To put it simply, vehicle manufacturers want to focus on what the data offers to their core business, instead of scrambling to build underlying technology that they could easily get through a partnership with a sophisticated supplier.
The resulting feedback loop of better knowledge through data leads to even more product expertise, which in turn opens up more options and opportunities as well as further refined analytics parameters and categories. To learn more about adding value with data, contact Sibros today.