Carvana vs. Vroom: The Two Largest Online Used Car Retailers

Avatar
Posted on Nov 5, 2020

Link to the Code

Background

The used car market is massive. It amounts to a TAM (total addressable market) of almost $1 trillion with 40 million+ cars sold each year. Despite such a size, the industry is notorious for its lack of transparency and poor customer service. Competition is fierce, as the largest player (Carmax) has only ~2% market share. Yet the industry had not focused on innovation until recently when the e-commerce boom started to influence everything customers buy including cars. 

Carvana (CVNA), Vroom (VRM), Shift (SFT), and Carlotz (ACAM) entered the industry in the last few years to transform the car buying experience into one that can be entirely done online. Carmax (KMX) started to follow as its stock languished behind that of its new tech peers. While the coronavirus pandemic only accelerated this trend, e-commerce for used cars is still less than 1% of total units sold. Online car-buying will only accelerate from here and the winner(s) will be handsomely rewarded for taking advantage of this massive shift in consumer behavior. 

Goal

Inventory is one of the most important elements that drives sales growth for a used car retailer. The goal of this project is to compare inventories of CVNA and VRM (the two largest online used car retailers) to assess which company has a more desirable inventory, which improves its odds of driving sales growth and stock price appreciation. 

Analysis

Both CVNA and VRM websites were scraped on 10/27/20 for used cars in the price range of $22,145-29,527. This range was derived from a downpayment of $2,500 and a monthly payment of less than $500, which is what a typical American pays for a used car.  CVNA was scraped using Scrapy and VRM was scraped using Selenium. Inventories of the two companies were then compared for differences in their age, brand, models, miles, and price profiles. 

The average age of a used car sold in the US is 3-5 years old. Both CVNA and VRM tend to focus on younger cars as shown below but CVNA offers a wider range of selection and has higher percentage of inventories in the 3-5 years old range. There also seems to be a statistically significant difference in the mean age of the two companies' inventories with CVNA cars ~1 year older than VRM cars on average since the p-value from a two-sample t-test was almost 0. Overall, the age profile of CVNA inventory looks better than VRM's.

Brand and model distributions don't seem to yield much insight. Both companies have ~50% of their inventories in the top 10 brands and ~7% of their inventories in the top 10 models. The list of top 10 brands and models are from Autobytel and Autotrader, respectively. While CVNA does offer a wider selection, the difference is marginal.

Intuitively, the distribution of miles should resemble the distribution of age since each vehicle in the US is driven ~12,000 miles on average per year. The chart below illustrates the mileage distribution for each company is indeed similar to its age distribution. Interestingly, however, miles per age is ~20% lower for a CVNA car than a VRM car. This difference in mean miles per age is statistically significant (p-value of almost 0) based on a two-sample t-test and speaks to the higher quality of inventories at CVNA. 

Lastly,  CVNA also seems to have an edge on the most important variable in a car buying process, i.e., price. On average, CVNA cars are $312 cheaper with p-value of 0.024 from a one-sample t-test. Adjusting for age yields a slightly less compelling result, though the price differential is still meaningful enough to warrant attention. 

Conclusion

As the chart below shows that while CVNA and VRM both entered the used car business in 2013, CVNA has been growing much faster than VRM. The analysis of their inventories indicate that this trend is likely to continue for the foreseeable future unless VRM changes its inventory sourcing strategy. VRM needs to change how it prices its inventories as well as improve the quality of its inventories in order to compete with CVNA. Therefore, despite the higher valuation, CVNA is still a better investment than VRM at this juncture. 

About Author

Avatar

Hong Chan Kim

Hong is a data science fellow at New York City Data Science Academy (NYCDSA) with expected graduation date of December 2020. His domain expertise lies in the US equity market, where he spent 7 years in the hedge...
View all posts by Hong Chan Kim >

Leave a Comment

No comments found.

View Posts by Categories


Our Recent Popular Posts


View Posts by Tags

#python #trainwithnycdsa 2019 airbnb Alex Baransky alumni Alumni Interview Alumni Reviews Alumni Spotlight alumni story Alumnus API Application artist aws beautiful soup Best Bootcamp Best Data Science 2019 Best Data Science Bootcamp Best Data Science Bootcamp 2020 Best Ranked Big Data Book Launch Book-Signing bootcamp Bootcamp Alumni Bootcamp Prep Bundles California Cancer Research capstone Career Career Day citibike clustering Coding Course Demo Course Report D3.js data Data Analyst data science Data Science Academy Data Science Bootcamp Data science jobs Data Science Reviews Data Scientist Data Scientist Jobs data visualization Deep Learning Demo Day Discount dplyr employer networking feature engineering Finance Financial Data Science Flask gbm Get Hired ggplot2 googleVis Hadoop higgs boson Hiring hiring partner events Hiring Partners Industry Experts Instructor Blog Instructor Interview Job Job Placement Jobs Jon Krohn JP Morgan Chase Kaggle Kickstarter lasso regression Lead Data Scienctist Lead Data Scientist leaflet linear regression Logistic Regression machine learning Maps matplotlib Medical Research Meet the team meetup Networking neural network Neural networks New Courses nlp NYC NYC Data Science nyc data science academy NYC Open Data NYCDSA NYCDSA Alumni Online Online Bootcamp Online Training Open Data painter pandas Part-time Portfolio Development prediction Prework Programming PwC python python machine learning python scrapy python web scraping python webscraping Python Workshop R R language R Programming R Shiny r studio R Visualization R Workshop R-bloggers random forest Ranking recommendation recommendation system regression Remote remote data science bootcamp Scrapy scrapy visualization seaborn Selenium sentiment analysis Shiny Shiny Dashboard Spark Special Special Summer Sports statistics streaming Student Interview Student Showcase SVM Switchup Tableau team TensorFlow Testimonial tf-idf Top Data Science Bootcamp twitter visualization web scraping Weekend Course What to expect word cloud word2vec XGBoost yelp