Developing Business Strategies for Vehicle Data

Connected and autonomous vehicles (CAVs) will create and use ever increasing volumes of data.

There will need to be close consideration of the development of strategies around the data influx, software generation, data harvesting and analytics because they will be crucial. This will necessitate the need to define data ownership and to embrace cloud and connectivity orchestration as they are at the forefront of the automotive industry agenda.

Maite Bezerra, principal analyst in software-defined vehicles at Wards Intelligence, says the key data trend has to be about using the data from the car and users to make better informed decisions. This requires an artificial intelligence capability. However, this still leaves the question of how the data would be used. She says it can be used to understand consumer to “provide better customer experience and customization”.

Fleets can use the data to health check their vehicles remotely to reduce the number of recalls or warranty claims too. Safety is another aspect of using vehicle data. For example, with vehicle-to-vehicle (V2V) communication information it’s possible to collate data and relay information between connected vehicles about an accident that may be nearby. This might also be about warning a vehicle or a human driver that they are approaching a pedestrian crossing; or enabling information about other hazards to be signposted in real-time. “The last part is to improve ADAS and autonomous driving algorithms to enhance these systems in the future,” she says.

Level 3 automated driving

Varun Krishna Murthy, industry analyst for connected and autonomous driving, Frost and Sullivan adds: “There are many trends such as data security (cyber-security for production vehicles), collaboration for compute and storage between automakers and cloud players, and the evolution of electrical architecture to accommodate newer features and related data.”

He believes the data influx will occur with the deployment of Level 3 automated driving and on a large scale. The commencement of the timeline for this isn’t just around the corner in his view. “You are looking at a timeline of 2025, being the starting point” but it will take years find a tipping point.

Data ownership

Like Bezerra he says the automakers will own all vehicle data. To make sense of the data they will work with third-party partners, to enhance functionalities to offer new services with the aim of creasing monetizable opportunities from which they may profit or use to attract customers. He explains: “Automakers will use this data and a continuous learning process to enhance Level 3 piloted driving on newer road stretches, cities or geographic locations.”

Bezerra says the vehicle data will only be shared with third parties in an “aggregated, anonymized form so that users are not identified”. So, for example, she explains that third parties will be able to learn how and where people are driving but not “how and where a person is driving”.

In contrast there is also a need to consider customer data, which is treated differently. It is only shared with customer consent and she says the data ownership belongs to the provider the user is sharing the data with. This includes sharing your location data with Google Maps via your vehicle system to receive real-time traffic updates, allowing Google Maps to have access to your data.

She adds: “There are worries about this but what consumers believe to be acceptable to share in exchange for a better user experience will change. Look at our phones, the more data you share about yourself, the better experience you have – more personalized services and offers. Eventually, we will all be sharing more data with the car but, of course, with user consent.”

Software perspective

In terms of the software perspective, she comments that automakers generate vehicle data. She explains: “They use the input from these data to deliver value to the end consumer trough their brand experience. Whereas the role of the software company is to be the enabler; they handle and analyze vehicle data though their sophisticated AI models providing input for carmakers to improve existing and develop new use cases.

“Some of these companies also happen to have consumer facing brands that compete with automakers offerings or are even investing on their own vehicles themselves. However, working with them is inevitable. Automakers can try to develop their own algorithms and AI capabilities but it is unlikely that they will reach the same level as software companies because this is their expertise. Therefore, they must find a common ground of collaboration.”

Nascent understanding

Collaboration is key to the automakers’ success. Krishna Murthy says this is because their currently level of understanding is still nascent. “Software-defined vehicles are complex and automakers have not attained full expertise, which is why they still have to rely upon partners including Tier 1 suppliers, semiconductor players, software developers, rather across the value chain,” he clarifies before predicting that collaborations will eventually lead to market consolidation. An example of this is Bosch’s takeover of FIVE, an AI software company. It was purchased because it has expertise in developing the AI for autonomous driving software.

The key to understanding the data is about how to monetize it. Bezerra says that only 0.1% of all vehicle data is presently sent to the carmaker’s cloud and much less, at 0.001%, is used for monetization. Opportunity will therefore knock for the automakers who develop the right strategies, partnerships, collaborations, expertise and knowledge. However, she points out that they can’t transfer all vehicle data to the cloud because it would be too costly and inefficient. CAVs tend to have several sensors and they can subsequently generate much data.

She elaborates: “OEMs are still figuring out how much of vehicles’ data is useful for delivering value, what portion should be processed in the car, what portion should be transferred to the cloud and ultimately how to translate the data in the cloud into monetizable use cases. Moving forward there will also be high emphasis on data standardization. Today OEMs only use data from their own vehicle fleet but that limits what they can do with AI and machine learning because these algorithms require large and diverse datasets to be optimized. However, data sharing relies on data standardization, and today the industry is very far from it.”

Who’s paying for services?

There is also a need to work out who’s paying for services such as 5G to enable a seamless data flow. At present she says no-one knows.  Bezerra believes it very much depends on who is benefiting from the deployed services. So, for example, if consumers are the beneficiaries, they are likely to be the ones paying for it. Yet, if it’s of value to the automakers, they are likely to pay for it while recovering their investment through the price of a vehicle. There may also be circumstances where both automakers and consumers can both benefit. In this situation, the case of who pays will depend on who’s gaining the most value.

Krishna Murthy nevertheless comments: “Initially automakers might offer it as a part of the connected vehicle subscription. However, the cost will be eventually passed on to the customer as part of a monthly/annual subscription or as part of the overall price of the vehicle.” Bezerra adds: “Overall, there will not be a single monetization strategy for all services. Some will be included in the vehicle price or subscriptions, others will be paid by OEMs, and some will be paid by third-party companies.”

Enhance customer experience

She concludes that the customer experience is enhanced when automakers extract the most value from the data generated in the car or vehicle more generally. What of the future? Well, it will involve a combination of vehicle, user, and cloud data to allow a car, for example, to understand its environment and to adapt to changing circumstance accordingly. In this sense vehicles will become more sensitive to the needs of their driver or passengers.

Sensing vehicles

So, if the vehicle senses that the driver is stressed, the vehicle would reduce the amount of information on the car’s screens or play your favorite song while re-routing itself to a road that is less congested. With regard to autonomous driving, an example would be the vehicle recording data on where and how you park your car every evening. Using AI and machine learning it would learn how to it, and then have the capability of performing the task autonomously. The technology will enable the analysis of vehicle data, allowing them to learn and to improve to offer a personalized experience. Automakers will therefore wish to tap into the data influx, and to develop strategies to offer a range of experiences.