Editorial analytics, that is tailored in cultures supportive of data-informed decision making, is the gold standard of audience development, said Won Kok Sung, data scientist of Mediacorp, Singapore.
There are bespoke tools to address gaps in existing off-the-shelf solutions and specific newsroom needs, said Won, who
believes the industry is moving towards data-centricity. Won was one of the speakers at the sixth annual Big Data & AI for Media conference in Hong Kong from 7-8 December.
From data infrastructure and data and measurement to audience insights and audience activation, data plays an important role in monetisation and advertising targeting and optimization, explained Won.
Popularity indicated by lifetime pageviews, participation shown as social pageviews or interactions and users interest as average dwell time are the three indicators of content performance, Won added.
To understand consumption context, they adopted a 360° view assessing the audience, the content they consume and the way they consume the content. They then look into audience development effort, which gives them the answer to “where did they find the content?” and “how engaged were they?”.
RIPPLE score based on predicted page view volume is built with data feeds, taxonomy classification, interest imputation, demographic inference and social engagement.
Won also explained that because more than 90 percent of lifetime page views are realised in the first 72 hours, their RIPPLE model only takes data from the first three days after an article published for analysis.
To help editors in the newsroom to act based on data, Mediacorp also developed apps for alerts, provide analysis (insights) and information about trending content.