How SCMP uses data & automation to drive digital transformation
South China Morning Post's director of data analytics, Korey Lee, shares exclusive insights
After you were acquired by Alibaba Group, you implemented big data and expanded with a team of data scientists. What are the learning’s so far?
The impact was particularly apparent in our data-driven transformation – democratizing data via various tools within the organization, empowering our teams to make smarter decisions. Measuring the impact of those choices gave us the agility to adapt quickly when necessary.
With the advent of these initiatives around data storage, centralization and automation, we've seen massive gains in :
- Growth in traffic & revenue,
- Rising excitement, morale and engagement at the SCMP.
“I’m a firm believer that data is not useful unless it drives an action. In other words, a recommendation should at minimum, motivate the consideration of a shift in thought, strategy or execution.”
What is the SCMP strategy to make audience data actionable?
The framework for driving actionable data is three-pronged:
1. Data cleaning
It’s necessary to not only aggregate data from across different sources and platforms but also to ensure the integrity of the data. Make sure it’s all housed on one platform for ease and efficiency to access.
2. Data mining
Once the data is clean and usable, we are able to start generating reports, adhoc analyses on customer behavior, retention, user journeys, top stories, best performing topics and so on. These are the “ah-ha” moments and nuggets of gold that we mine from the mountain of data gathered in our warehouse.
3. Data actions
On a daily basis, my team provides recommendations cross functionally to different business units to drive incremental traffic, review what we did well to repeat wins and understand where we underperformed so we can mitigate against them in the future.
“Creating a culture around testing over gut feel is healthy and helps to drive data informed decision making across the company.”
How does this strategy enable the SCMP to build new products and refine existing ones?
1. Content recommendation:
Implementing content recommendation solutions have enabled us to increase recirculation and serve more tailored content to our users on a regular basis. While we’re in the early stages of this, we’ve already seen promising results and more engaged users.
Our aspiration is to get to a place where every user who comes to SCMP.com sees a customized view based on their preferences and what they or someone like them enjoys reading.
2. A/B testing
Building the tools, educating and creating a culture around editorial headline testing has been challenging but the numbers have been fruitful. Chartbeat enables us to test headlines on the fly across various pages and react to reader engagement in real time.
If we think a story should be trending but it’s not picking up traffic, we’ll start testing different headlines to see if we get better pickup via SEO and better engagement across our internal linkages. A/B tests on headlines have yielded varying increases in overall traffic as we continue to iterate and improves processes.
Creating a culture around testing over gut feel is healthy and helps to drive data informed decision making across the company. While we can’t reveal what new products are launching or the nature of them, we can say that our deep dive research into our user profiles, personas and behaviors has helped to inform and drive the creation of these new products.
A/B site and A/B headline testing will a large part of these refining the shape of these new products as we continue to leverage a minimum viable product (MVP) philosophy to launch, iterate and continually improve as we get feedback both from testing and from our users.