“Data science” and “Big Data” were among the trending phrases last year and will continue to be for some time, but they’re not all that new. The science of collecting and learning from vast quantities of data has been around for a long time.
Within consumer industries, data science has been more associated with sales and marketing than with manufacture and design. What is new is that this branch of science is being taken very seriously by vehicle manufacturers, who are bringing it in-house to a greater extent than ever before.
The finance and banking industry was an early adopter of Big Data. In Central London, San Francisco and New York, banking and insurance company policies were based on derivatives driven down from data analytics and complex formulae.
Many automotive businesses have had their own data analytics departments for some time – but initially these were concerned with analysing marketing and sales data to drive business, understand consumer behaviour and interpret website traffic.
The rise of data use in the automotive sector
Many of us will be aware of the increasing role data now plays in vehicle servicing. As cars become more complex and able to collect more data, a regular car service entails downloading information from a vehicle whose sensors monitor wear and tear and report on mileage, fuel efficiency and routes.
The big OEMs have teams that work with that data and use it to redesign products or create new ones, correct faults or improve supply chain quality; this data impacts how those businesses follow up on sales and marketing activity and create a better customer experience. As vehicles become more complex, this creates a huge amount of data, but it’s still only the tip of the iceberg for the automotive sector.
As cars become more integrated technologically, the real explosion in data science is being driven, if you’ll pardon the pun, by connected and autonomous vehicles.
Autonomous vehicles, data and data analytics go hand in hand. Indeed, this new branch of design and manufacture is only made possible through the creation and deployment of advanced analytics. The key elements of data science here are the Internet of Things and Machine Learning. For a machine to learn it needs vast quantities of data, and the reason why autonomous vehicles are only being realised commercially now is because it is only now that data science has become sophisticated enough and accessible enough for the automotive sector to use it.
In order to use data to learn, we must be able to collect it, store it and interpret it.
The advent of cloud storage and Wi-Fi has enabled increasingly large amounts of data to be analysed, which enables the development of autonomous systems such as sensors including cameras, radar, LIDAR, IMS, ultrasound, GNSS and more besides.
The geniuses arrive: the lure of the industry
The automotive sector is now beginning to benefit from people who have made money as silicon entrepreneurs. People who have built empires by developing microprocessors used in computers and storage systems now see the automotive sector as a landscape rich in opportunity. Tesla founder Elon Musk is perhaps the most high-profile example of someone who has been drawn to the sector from outside, at the same time making automotive technology sexy.
Five AI is another business that demonstrates how leading talent from within software and communications tech is being drawn to the automotive sector. Founder Stan Boland sold two microprocessor businesses and in Five AI is building a company that, rather than developing autonomous vehicles, is creating the software and systems for them. These will be sold to automotive giants who will develop the technology within their own businesses.
This is a burgeoning and potentially gigantic field of endeavour. From a talent perspective, it links automotive skills and technology skills more than ever before and represents a seismic shift within two industries, each creating new opportunities for the other.
A lot of the data experts the automotive industry needs to entice will have previously worked for marketing companies and start-up finance firms. The automotive firms hiring these individuals would previously have applied data analytics more to their own marketing than to their vehicles. Concept design and data analysis are beginning to work together, because vehicle design can learn a lot from data.
Luxury brands like Bentley have always been able to speak directly to a closed, accessible consumer market to gain insightful feedback about product design. Through VIP events and various high-touch feedback mechanisms, they can learn what customers like and dislike, what evolution they’d like to see and what changes are ultimately blind alleys for the next generation of vehicles.
That’s all well and good for boutique brands with a consumer base that can be treated more like a circle of friends. But for high volume manufacturers selling tens or hundreds of thousands of units per month, a database of millions of customers can be a fruitful way of gaining feedback – but only providing you have the sophistication to assess, identify and interpret trends. Analytics and design can therefore work hand-in-hand, and it is data scientists who can unlock the potential of the one to positively impact the other.
Where will the talent come from?
Good data candidates are out there, but a big challenge lies ahead.
I recently watched Stan Boland advising a parliamentary committee that helps to shape future government policy for AI and autonomous vehicles. The transcript reveals a discussion about the idea of “the car mechanic of the future (being) a software engineer as much as he is a mechanic”, with Mr Boland opining, “we definitely need more software engineers as a nation…so we are probably not ready for any of this in terms of the total number of skills that we need to go alongside companies the size of Silicon Valley Companies…”.
The increasingly common marriage between mechanics and software means embedded software engineers will become even harder to find - and more important - than they are now. But demand for computer scientists, mathematicians and strong programmers will also rise. This won’t apply only to embedded electronics professionals with mat lab skills but also to Master’s students with strong Python programming and analytical skills. Data analysis is the back end: new data scientists will play a crucial role in creating complex algorithms, and maths and Python will feature on the “desirable” skills list.
In the recruitment sector we often refer to the “war for talent”, but I suspect this won’t be too much of a hyperbole when applied to the battle for the best data scientists over the next five years and beyond. The big challenge for the automotive sector stems from two areas: its ability to compete with other sectors to attract top talent, and a skills gap stemming from educational trends over the last few years.
The automotive industry has to compete against the banking and security sectors who are well entrenched in data science and pay very well. A car manufacturer may pay £500 a day for a data scientist, but the banking sector could double that figure.
Then there’s the skills gap. Over the last two years, trends in pedagogy have resulted in more students coming out of colleges and sixth forms with good science and maths ‘A’ levels. But this is only now beginning to balk a trend that extends for the previous decade, when we did not have enough people studying maths and science to the level required to go on and study it at undergraduate level.
Those generations of graduates are now working and that gap in the sciences is being felt in UK industry. This coincides with a period in which our ability to recruit and attract talent from outside the UK remains to be properly understood.
Over the next five or ten years, our immigration policy could be one of the most significant parts of its executive portfolio for the British government to manage, and a number of industries will be watching very closely whilst, no doubt, making their voices heard. This is a huge challenge for the UK and its businesses.