メイン コンテンツにスキップ
ダッシュボードに移動
どこから手をつければよいかわからない方は、簡単なクイズに答えて、パーソナライズされたおすすめを参照してください。
South China Morning News

South China Morning Post's data-driven approach boosts user engagement

Creation of a dynamic recommendation engine delivers personalized user experience
South China Morning Post newsroom

The challenge

With over a century of experience as a leading global news company reporting on China and Asia, SCMP has gone from strength to strength in its journey from a traditional newspaper to a digital-first media organization. As SCMP's global audience continues to grow, the company needs to better understand its readers to ensure that users on its platform are consistently offered content relevant to their interests, drawn from the various publications in its portfolio.

In 2021, SCMP collaborated with the Google News Initiative (GNI) to build a recommendation engine that dynamically creates User Types based on various data sets. This allows it to serve up relevant content and calls-to-action, using algorithms to determine preferences and demographics. The goal of the project was to increase both user engagement across the board and conversions to paid subscriptions.

Headshot
This project exemplifies our commitment to elevating our users' experience by serving readers with content based on unique demographic, psychographic, and behavioral data. By better personalising user journeys, we can meaningfully engage with our readers with relevant content while achieving stronger retention and conversion.
Vhanya Barba Mackechnie
Vice President, Product, SCMP
6x Increase in CTR via “People also Read” widgets
25% Increase in average time on page

The results

Aiming to improve content recommendations based on User Types, SCMP split testing into two rounds, each with a different focus. The first round tested variations of the article page structure; the second tested different algorithms based on the winning page structure identified in the first round.

By setting up a Standardized User Type retargeting framework, users that hit the paywall were targeted first and consistently converted. Identifying detailed User Types also allowed SCMP to serve more personalized content. With further iterations in content recommendation, it was able to improve the “People also Read” suggestions; clicks on the recirculation widget increased by 6x, demonstrating higher reader engagement. Removing Perpetual Scroll incentivized readers to browse the widgets and select which articles to read next, driving a 25% increase in average time on page versus the previous page setup.

移動して進行状況が失われてもよろしいですか?
このページから移動すると、現在のレッスンの進行状況がすべて失われます。続行して進行状況が失われてもよろしいですか?