No Results Found

NDTV, a leading Indian broadcaster, wanted to boost digital content consumption by growing new audiences and cultivating loyalty. They recognized that while trusted methodologies (like editorial judgment, readability formulas, and basic web analytics) were useful in measuring engagement, there was scope for improvement by leveraging real-time insights to refine content.

NDTV decided to create a framework for increasing engagement based on what content resonated most with readers and what they found valuable. To do this, NDTV partnered with Google and adopted a data-driven, iterative approach to enhance the quality of published content. This approach, The Newsroom Quality Index, involved a two-stage tool that provided editorial teams with real-time, data-driven insights.

Stage 1: Pre-Publishing Quality Score

In the first stage, NDTV focused on strengthening the foundation of article writing. To ensure that the articles are robust, original, and engaging, NDTV developed a comprehensive checklist of essential inputs.

Essential Elements for Quality Articles

Base Inputs (5Ws and H): Ensuring that the story answers all the basics – Who, What, Where, When, Why, How.

Multimedia: Using original photos, videos, infographics, and social media embeds for visual engagement.

Supporting Elements: Quotes and data for credibility and depth.

Core Hygiene: Meticulous checks for spelling, grammar, plagiarism, bylines, and headlines.

Once these elements are in place, the story is pushed into NDTV's Content Management System (CMS). The CMS generates a pre-publish quality score before the story goes live. Editorial teams must ensure the highest possible pre-publish score for every story.

Stage 2: Post-Publishing Predicted Engagement Score

In stage 2, NDTV combined several data sources such as event-based and user data via Google Analytics, real-time data from Datastream,raw data and article data from NDTV’s native CMS, by leveraging APIs and Google Cloud buckets to enable seamless extraction of data into Google BigQuery. The final unified data set allowed NDTV to effectively analyze and apply a quality score to each article.

NDTV leveraged Machine Learning to build a robust and accurate model which analyzed CMS details and real-time user interactions (consumption, time on site, etc.) to predict an article's engagement trajectory.

Refer to the detailed explanation here.

With access to predictive quality scores, editorial teams were empowered with a data-backed indicator of how any content would likely perform, and could make real-time updates to articles and evaluate the impact of the changes made. The continuously learning ML model ensured accurate predictions, updated every 30 minutes.

The tool was launched for NDTV English news articles in July 2023. To measure the impact, NDTV compared page views, average session duration and bounce rates before and after implementation. Improvements reflected across the board. NDTV also observed a parallel success story with respect to its unique visitor count on these pages.

**The comparison period had no major outlier news events (such as wars (Oct’23 and Nov’23) or elections (Dec’23)) to influence these results.

By implementing a machine learning-based process to analyze user engagement data, NDTV was able to identify articles performing well and those that needed improvement. This empowered editors to optimize content strategy by focusing on topics and formats with the highest viewership. NDTV registered a spike in website traffic and viewer engagement by adopting this strategy.