Calculating the Correct Used Car Price using Machine Learning - CARS24

There is no denying the fact that India, on the whole, is a very price-sensitive market. While for a new consumer product, deciding the right price for which it should be sold it isn’t very complicated and is done quite accurately by the OEMs. However, when we consider a used product, pricing becomes a very complicated attribute to evaluate correctly. Unlike the 90’s, the social stigma associated with selling a car or purchasing a used car doesn’t exist anymore. If anything, the trends show that both the people as well as the market are upgrading their commodities at a rapid rate. Nowadays people are also as open to investing in used premium cars as they are on equally priced new cars. This shift in trends can certainly be attributed to the flourishing middle class with increasing income rates and greater mobility. When you consider a commodity like a used car, the sheer number of variables and factors that are required to be considered in evaluating its price makes it a very complicated task. Hence because of this, people tend to use very arbitrary or inefficient methods to calculate the used car price. This is done mostly for the sake of convenience or even out of lack of awareness. Regardless of why it is done, one thing is for sure, these methods can never get you the correct used car price.

CARS24's Intelligent Pricing Tool

Keeping all these factors in view, we have designed a very sophisticated pricing tool which takes into consideration several factors and computations before arriving at the correct used car price. Unlike the baseless and arbitrary price calculation methods which are common in practice, the CARS24 pricing tool is based on real-time data.

We at Cars24 are trying to solve for the inefficiencies in the used car ecosystem by leveraging a real-time digital auctioning & transaction platform; customer focused retail fronts; state of art inspection approach; and advanced ML techniques in pricing & buyer-seller matching algorithms. Having transacted in excess of 1.35 lakh cars and inspected & auctioned around 5.50 lakh cars since inception about 3.5 years, we have access to a huge training data to establish a relationship of transacted price point. For every new car that we inspect or sell, we further expand our database with the data of that particular car as well. This allows us to increase our precision literally for every customer we get. Click here to read the full article, written by our very talented Data Science Head, Naresh Mehta.