Deal or No Deal?

Original article was published by Sanjay B Bhakta on Artificial Intelligence on Medium

The automotive industry is highly influential within economic development, supplier industries, political regimes, and consumes about half the world’s oil. This industry contributes about 3% to the U.S. GDP, bringing jobs, enabling revenue within the local and state governments, and consumes goods and services from other sectors, such as computers, financial, healthcare, machinery, raw materials, semiconductors, and others¹.

The U.S. is the second² (compared to China) largest automobile market in the world, with approximately 273.6 million² vehicles registered in 2018, with many autos transmitting a beaucoup volume of data. Regarding this volume of data, the monetization of car data³ is estimated to potentially generate revenue between USD 450–750 billion by 2030.

Data monetization³ from vehicles reveals new business models such as subscription based predictive maintenance (optimizing total cost ownership), auto insurance (driving behavior incentives premium), risk mitigation of highway safety, road conditions for preventive maintenance, new products, cost optimizations, logistics optimizations for OEMs, and host of others.

Today’s new normal⁴ of purchasing autos is more heavily dependent upon e-commerce with contactless delivery. The first half of 2020⁴ experienced a significant decline in car purchases, both new and used, while the second quarter and remaining year are expected to have larger gains in the used car market. The profit margins⁵ are larger for used cars and even more in demand in 2020 due to inventory challenges of new cars.

In the U.S. 69% of car buyers prefer used versus new, while still expressing concerns on the reliability and repair needs of used cars⁶. Interestingly, 2020⁷ seems to be opportunistic for the buyer of used cars as prices have dropped an average of approximately 6%, depending upon geography and model. Luxury⁷ cars seem to be the loser, Fiat 500e price drop >15% and BMW i3 price drop est. 10%.

Oh yes, let’s not forget the psychology of car buying, where the model and color signals your personality, such as black symbolizing power and elegance⁸. So now’s a great time to develop a data science model for predicting the price of used cars, and optimize your next purchase, as I’ve done with a preliminary ML model within the Juypter notebook, from my management consulting with dealerships, fleet management companies, and others⁹. FYI, in the works is debiasing the model (race and gender), as well as further conditioning the ML model with gigs of pricing data, demographics, social, total cost of ownership¹⁰, and of course psychology for that “personalization”.

The insights gleaned from the automotive industry are astounding, the glands are salivating for the prime pickings of new business models. It’s the most influential industry in the world, and tightly integrated within our economies. Monetizing the data from autos gives the synthetic in “data oil”. The year 2020 shall be unforgettable, but how about trying to capitalize on the circumstance in purchasing that next vehicle. Let data science lend a hand, so you may negotiate as “Deal or No Deal”!