Questions and perspectives from Stockholm publisher's analysis

Mar 30, 2021 at 05:45 am by Staff

Patterns from time-related metrics are helping a publisher learn about its readers' preferences and habits.

In an INMA Big Data post, Dagens Nyheter editorial data scientist and analytics team lead Lovisa Bergström shows what can be learned from changes in reader behaviour, including those driven by external factors such as pandemic restrictions.

Behavioural analysis is also a simple way to shift to a more outside-in perspective on your publication schedule and trigger internal discussions, she says.

"The goal is to publish our articles when they have the best possible chance of being read - by many readers and to the end of the article - converting new subscribers, and being shared."

Examples include the times when readers have enough time to finish long articles, when they usually come from organic search, and what kinds of articles they tend to look for.

She urges users to agree on the question to be answered, and choose metrics and a timeframe accordingly. "Behaviours change, and you don't want to draw conclusions from obsolete patterns and outdated data," she says.

To avoid specific news events that affect your numbers, exclude outliers for each slot, perhaps by limiting outliers to three standard deviations.

The "fun part" is the visualisation and conclusions, and Bergström advises plotting each slot in chronological order, typically with heat maps in which weekdays are on the x-axis and hours on the y-axis, plotting both average and the percentage change.

Experiments with different selections can follow, with heatmaps of different metrics side by side. She recommends Jupyter Notebook and Google Colab as good notebook tools, which can be shared within an organisation and are very time efficient to run and configure while it's in place.

Then there are questions about correlation and causality: Do readers spend time because of long articles, or do we publish long articles at specific times because readers spend so much time on our page then?

With data fuelling the discussions, the analysis delivers new perspectives and a basis for internal discussion. See the full post here.


or Register to post a comment