The journey of Star from data to
decisions with Gramener’s
Insights Platform
Challenge
Identifying insights flowing in from over 60 channels was tedious and time-consuming
Approach
Gramener’s insights platform is enabling quick and efficient viewership data analysis
Outcome
80% reduction in data extraction efforts, analytics and communication efforts halved
ABOUT THE CUSTOMER
Star India Private Limited is an Indian media company and a wholly acquired subsidiary of The Walt Disney Company – India. Star has a network of 60 channels telecast in eight languages. The network reaches approximately 790 million viewers in a month across 100+ countries including India. Altogether, Star India generates more than 30,000 hours of content every year.
THE BUSINESS CHALLENGE
Globally, media companies have started experiencing new business opportunities and possibilities with data, analytics and technology. The potential of viewership data and television ratings to deliver actionable insights and unlock business value is now no secret.
Till a few years ago, Star was unaware of the opportunities that the sea of TV viewership data and media ratings held across their organization. A holistic approach to bring together, analyze, and mine it for audience measurement and meaningful decision-making was yet to be understood.
Being a media company, there was no single source of data but many hiding viewership data and television ratings. Collectively, it led to challenges in terms of data volume, data access, timely availability and usability of data for decision-making. The sooner Star democratized data with the help of technology, sooner they could enable quicker and better decision-making.
THE NEED FOR AN INSIGHTS PLATFORM
Star is a pioneer in leveraging data to drive decisions in the Indian media industry. They have continued to find ways of acquiring new sources of data. Each data source is available at a different level of aggregation, sampling and follows different data dictionaries and distinct timeliness. It was important to extract the right data from the right channel before doing viewership data analysis.
This created a challenge in terms of consolidating them in the entire Extract, Transform and Load (ETL) flow. They had to be creative in treating each flow differently, incorporate dynamic data validation and ensure that transformations of data across sources are uniform and synchronised.
Viewership data at fingertips with insights platform
The insights platform team was placed uniquely in the organization with an insight into business and expertise to build products. A TV audience analytics product was built on the foundation of subject matter expertise, technological knowhow, proven data science use cases, deeper understanding of data sources and data to analyse consumer behaviour and drive business decisions.
Extraction of data was the most effort-intensive task and we moved it to the cloud and ensured it happens in a distributed manner. This cut down our ETL time from 24 hours to 1 hour for key data sources
We had to transform and host this data for applications that warranted aggressive response times. We used a host of technology solutions to suit each source and application As we gathered new data sources with raw events instead of aggregated events, we updated the latest databases and also upgraded to one which could cater to requirements in the foreseeable future
The insights platform APIs also helped other teams do viewership data analysis by allowing access to the prepared data layer and use it as a base for other products
Visualization to ease the discovery of patterns
Data visualization was still a large missing piece and Star, being in a traditional set up with respect to technology, had limitations in terms of skill sets of users. The partnership with Gramener changed the way a story could be narrated through data.
This led to adopting data as a culture within the organization, incorporating it into decision making, from daily tactical calls to ideating long-term strategies.
A WORD OF APPRECIATION FROM THE CLIENT
Modern features embedded into the insights platform made patterns visually obvious.
- Customized notifications to BUs promptly after official data release
- A storyboard to enable users to create weekly custom reports that automatically refresh and provide a platform where they can be shared and consumed as required
- Compare data across sources at the click of a button for a holistic understanding of a viewer
- Connect multiple analyses through vertical and horizontal navigations over time, markets and consumer segments, while retaining the analytical context
- The Star team was able to classify the viewership based profiling to do audience measurement for a specific show/channel.
The audience demographics revealed geospatial information of viewers to find relevant markets to target and sharpen the marketing effectiveness was also available on the fly.
THE BUSINESS IMPACT
Kaushik Das, the Executive VP of Big Data and Data Science at Star says that the insights platform has simplified the data analysis process for the team.
- The insights platform has become a one stop shop for consumer and TV viewership data for users across hierarchies and Business Units and accepted widely as the single consolidated source to turn to
- With centralization of ETL, there was 80% reduction in data extraction efforts and ~50% reduction in analytical and communication effort
- Reduced decision-making time drastically. The turnaround times to get data in an analysed form reduced from days/weeks to minutes
- Every user now has complete visibility of the data, thanks to democratization of data. Viewership metrics are standardized across teams
- It was transformed and visualized in a manner that reduced efforts in contextualizing data for quicker diagnosis and hypotheses validation
FUTURE SCOPE OF THE INSIGHTS PLATFORM
Star and Gramener teams are together in the process of building an AI chatbot that processes the natural language of the user and redirects him/her to relevant pages according to their needs. Star is using data to bring in intelligent recommendation engines which use advanced analytics techniques such as sentiment analysis to decode TV viewership data and ratings.
