Interview With Karthik Chandrashekar, Director of Data Science At Netflix

One of Chandrashekar’s most significant achievements at Netflix has been building a unique low-priced mobile tier in India.
Listen to this story

“My job here is to make more money for Netflix, it is the short way to describe my mark,” Karthik Sriram Chandrashekar said in a conversation with Analytics India Magazine. Director of Data Science and Engineering at Netflix, Karthik and his team work on acquiring and retaining customers at the Netflix scale.

Karthik is also behind introducing the low-cost mobile Netflix plan in India. Analytics India Magazine caught up with him to understand his current responsibilities, past work and other projects in the pipeline.

Improving RoI on investment

Karthik has a degree in computer science engineering from BITS Pilani. Since then, he has worked for ten years in software development and machine learning. One of his earliest assignments was to work on ads ranking. “I also dabbled in healthcare tech, working to provide the right data infrastructure to execute electronic medical records, recommend the right medicine/product dosage, and coordinate hospital operations.” 

Karthik said that this role was not machine learning oriented, so he started working with ads ranking again. “It was like a location-based ads company. So, we were trying to essentially show ads that are relevant to the current location of a person based on their mobile data,” he said. 

His next stint was at Netflix, where he joined as Software Engineer (Machine Learning). He is currently the Director of Data Science & Engineering at the streaming company. He works with a team of 60 people, which is a mix of data scientists, engineers, and statisticians who are focused solely on causal inference. 

One of Karthik’s earliest projects was to customise the non-member experience. “We use machine learning to make decisions on gauging the highest return on investment to get customers,” he said. Personalised feed for a paid subscriber is understood, but how difficult it is to do the same for a ‘non-member’. When we asked this question to Karthik, he responded, “Personalisation for a non-registered member involves multiple steps. We do get a lot of data points, even from the non-members, for example, what mode they are using to access Netflix—mobile, web or TV app. Further, more than personalisation, I think contextualising is a more appropriate term.”

Building a unique plan for India

One of Karthik’s most significant achievements at Netflix has been building a unique low-priced mobile tier in India. This mobile plan costs INR 149 per month and is designed to attract Indian users and get a competitive edge over rivals like Disney, Amazon and others.

“Here, I would like to talk about how data played a role since it is the single most important factor that aids in making strategic decisions like pricing plans. So, the thought first came when we were trying to figure out how to unlock the next phase of growth in India. Not just India, we were looking at adopting a mobile-first strategy in the APAC region. This coincided with the larger changes in the industry overall with Jio’s launch,” told Karthik.

The first step here was to develop a market understanding using customer segmentation models. The team did a deep dive analysis to estimate the kinds of customers which could be retained better if they did introduce a low-cost plan like this. “That said, whether a business idea is good is largely a strategic question. We would not know much without actually launching and then checking what happens. This resulted in us developing a sort of quasi-experimentation strategy where we launched the plan across several markets in the APAC region. We then used observational methods to estimate how much impact it would have in other markets if we continue to expand to other regions,” said Karthik.

Once the plan was launched, there was another challenge of measuring the RoI in terms of revenue. What if customers using more expensive plans switch to cheaper mobile plans—was one of the questions. “One thing with rolling out a new service or product is that it can result in many kinds of outcomes. Performing many tests across markets and learning the aggregate impact of those things through causal inference techniques was a core part of what we did to launch the mobile plan,” he said.

Current projects

One of the projects that Karthik has been working on is monetising account sharing. About this project, he says that it is “fundamentally an extremely data-informed growth strategy”.

“It starts from understanding accounts that are being shared. Based on these pieces of information, we want to develop a product strategy that can help in monetising that behaviour in a very customer-friendly way. Evaluating those strategies through complex experiments is at the core of the project I am working on,” he said. Netflix already uses cyclical/concurrent streams which entails a limitation to the number of people who can watch at the same time. He added that the strategy they are currently working on will be novel in nature.

Other major projects that Karthik and his team are working on are improving machine learning infrastructure and bringing more automated experiments, including adaptive experimentation.

Download our Mobile App

Subscribe to our newsletter

Join our editors every weekday evening as they steer you through the most significant news of the day.
Your newsletter subscriptions are subject to AIM Privacy Policy and Terms and Conditions.

Our Recent Stories

Our Upcoming Events

3 Ways to Join our Community

Telegram group

Discover special offers, top stories, upcoming events, and more.

Discord Server

Stay Connected with a larger ecosystem of data science and ML Professionals

Subscribe to our Daily newsletter

Get our daily awesome stories & videos in your inbox