At PwC, I worked long-term on a project for NewsCorp, a mass media publishing company.
Problem
Content is NewsCorp’s core product with research demonstrating it is more than half the reason why customers subscribe to their mastheads
NewsCorp is a data-rich environment but previously had siloed data sources. This inhibited real-time data sharing and dynamic newsroom decision making on insights drawn from a complete data set
Goal
NCA’s strategic objective was to drive subscription revenue growth by leveraging their understanding of the return on content through data insights. This would be achieved by:
Audience (Pay): increasing the number of people consuming and subscribing to content;
Engagement and Retention (Stay): improving the value to subscribers and consequently their desire to continue subscribing; and
Effectiveness: improving the return on content creation.
A data analytics platform that delivers data and insights to journalists, editors, and management to ultimately drive and support content growth was needed. The goal was to develop a data product for journalists and management on top of this data platform called “Verity”
Output
I led the data modelling in BigQuery for a new data platform on GCP involving high volumes of news data. Modelling the data in BigQuery increased performance of the dashboards and decreased the cost through optimised query patterns.
Worked with senior stakeholders to define key data requirements. Led the functional design of newsroom dashboards surfaced in an internal app presenting near real-time data to 1000+ journalists, editors, and management for informed decision-making on relevant content.
I led the functional development and build of the newsroom dashboards and key data visualisation designs. For this I used Looker Studio and Python packages (plotly) for some use cases.
Quantified Impact:
Verity now provides News Corp Australia with an invaluable competitive advantage that will strengthen their newsrooms for future growth.
Verity has accelerated the use of data in newsrooms and shifted the culture of the business to insight led decision making
In 2018 new subscriptions driven from premium content breach increased by +133% year-on-year
Content driven acquisition as a channel accounted for 35% of new subscriptions in 2018 up 19 percentage points versus 2017
True Crime Australia launched in July 2018 packaging crime content within its own branded vertical and improving crime content subscription sales by +272% year-on-year
Awards:
News Corp Australia and PwC Australia were recently awarded “Best Use of Data Analytics” by the International News Media Association (INMA) for Verity
PwC Australia were finalists for two ARN Innovation Awards, ‘Customer Value’ and ‘Cloud Innovation’.
A Showcase of the type of work I performed are shown below
Verity combines multiple big data sources to give real time visibility of performance metrics across content and customer behaviour.
I dealt with content management systems, audience metrics, and web analytics, which resulted in millions of rows of data daily.
The data product was available in various formats - web, phone, and on big screens in the news rooms to ensure that data driven decision making was front of mind.
The article view dashboard included insights to identify promotions (i.e., have they been included in newsletter or social posts) and the predicted benefit/gain of actioning a particular promotion for this article type.
Journalists were able to A/B test article changes and the effects of promotions and get feedback in real time.
Types of insights displayed include:
Live insight matrixes by content category and trending topic allow the business to rapidly identify where topics are underrepresented and refocus on high value content
Author performance and category performance dashboards with information such as page views, # breaches to the subscription paywall, subscriptions, subscription conversion rates, engagement time on article, # social referrals.
Developed logic for daily email blasts to send to teams so that they could be used for morning team meetings and daily planning. This compared progress against targets and recapped the prior days performance.
One version of this report was built and the security and permissions were built into the single report at a row level so that information was kept private amongst teams.
Used Natural Language Processing (NLP) to analyse article content such as keywords, and people and organisations mentioned. This allowed teams to identify topics that were performing well and leverage this.
Monitored data quality and built reports to score teams on metrics such as completeness, conformity, and uniqueness. This incentivised teams to be responsible for their own data quality where possible and care about any manual data entry points.