Mumbai: Kotak Institutional Equities, a unit of Kotak Securities Ltd, in collaboration with Google, launched Consumer Querimetrix—a tool that simplifies and predicts near-term Indian consumer behaviour by analysing Google Trends data.
The first edition of Consumer Querimetrix provides consumer insights into India’s evolving car buying behaviour. Using machine learning techniques and merging big data from Google Trends, the tool enables ‘nowcasting’ (near-term predictions) on consumer activity, capturing inflection points earlier than traditional forecasting tools to give a complete picture — on car launches, last mile hiccups and competition. Here is a snapshot of a few:
Assessing new car launches
Each edition of the Querimetrix series will focus on consumer behaviour in a different industry. The first issue establishes that Google search volumes can be used as a gauge for assessing new car launches. The report underscores the linear relationship between search interest and advance bookings during a car launch. Interestingly, higher traditional media spending does not always guarantee higher search interest.
Classifieds impeding new car sales?
With over 223% year-on-year growth in used car queries in India, the report also outlines that horizontal classifieds (OLX, Quikr) may also be impeding new car sales as first-time car buyers may be substituting entry-level models with used car options. In contrast, vertical classifieds seem to facilitate new car sales.
Car sales to grow 16% in January
In its first projection, Kotak Institutional Equities predicts 195,000 passenger cars to be sold in January 2016. As Google queries on car loan inquiries begin picking up, it is anticipated that total passenger car sales will grow at 16% over last January.
There is very strong indication of growing competition in the market, as industry leaders have seen a substantial drop in search-share since 2010 even as advertising spends rose sharply.
Altering the ground rules
The Internet is altering the ground rules for vendors of cars and allied products and services. With growing access to easy information online, the Indian car buyer’s journey from a whim to final purchase has changed dramatically. More than 75% of car buyers are researching online for reviews, comparative specifications, and financial products and used car markets before making a purchase. The first edition of the tool also explains how the traditional ‘funnel’ model is giving way to a more complex purchasing path where ‘initial consideration’ may not always guarantee sales. Although higher auto-related searches correspond to higher demand for cars, this does not hold true on a brand-wise basis.
“The digital wave is challenging conventional business practices across industries," said C. Jayaram, joint managing director, Kotak Mahindra Bank at the launch of the report adding that ground rules are evolving rapidly along with the consumer. Therefore those in the business of business intelligence need new tools to keep up with the changing land-scape.
The sheer volume of consumer-centric search data available today presents a tremendous opportunity to analyze and throw up actionable insights, he added. These takeaways would be useful to both companies and investors.
With over 300 million Internet users online and growing, India today has a sizeable population which accesses the Internet on a daily basis, making search queries as the most dynamic data input to arrive at consumer insights through machine learning as illustrated by Kotak’s research. “If we look at the search trends related to car finance and car purchases, we’re seeing a 40% year-on-year growth in car purchase queries on Google in India, said Vikas Agnihotri, industry director, Google India.
Over half of the people who evaluate car purchases change their consideration set during their research phase-adding two to three new car models in their consideration, the only non-negotiables are price and colour of the car. This alone proves the growing complexity for car makers in the country. “With this report we’re attempting to demystify this changing consumer behaviour into actionable insights for the industry," said Agnihotri.
Saifullah Rais, quantitative analyst at Kotak Institutional Equities and the architect of Consumer Querimetrix, said, “In the absence of conventional rules, traditional decision-support systems are not very effective. They fall short on scalability and adaptability. Machine learning algorithms learn from data and do not rely on explicit rules, making them the most effective method of dealing with data explosion," he added.