As the South China Morning Post moves from 'local paper' to a global one with more than 94 per cent of online traffic coming from overseas, it's building a stronger data culture with more data-informed decisions.

Vice president of data Korey Lee says the 117-years-old publisher has been taking steps recently not to just use data, but to "thrive on it" as a means for providing readers with high-quality journalism.

He says relevant data can be used to improve the way a newsroom works and thinks, with the most critical way to build a good data culture being "training, educating and evangelising" why data matters.

"The story-telling process requires the cross-pollination of data tools and culture to shift decision making from intuition to data-driven.

"Weekly online editorial meetings, Slack and email alerts for under-performing and over-performing content, daily insight emails summarising wins/losses, and dynamic TV dashboards all contribute to driving this change."

Lee says these help form rhythms encouraging data reliance on real-time reporting. "KPIs enable teams to better track progress and have increased accountability for targets," he says.

Training workshops - dozens of them - were held to ramp up staff on tools, and features, and to tie data relevance into each individual's day-to-day. "As the organisation continues to transform, we're encouraged - albeit kept quite busy - by the tenfold increase of data requests over the past three months."

The focus on creating a transformative culture has started with investment in people. "The impact was particularly apparent in our data-driven transformation, democratising data via various tools within the organisation, empowering our teams to make smarter decisions," he says.

"Measuring the impact of those choices gives us the agility to adapt quickly when necessary. With the advent of these initiatives around data storage, centralisation, and automation, we've seen massive gains in efficiency, growth in our core metrics (such as traffic and revenue) and a rising tide of excitement, morale and engagement."

He says that while data is "not usually" the sole variable in driving a decision, it should play a role in most decisions. "I prefer to call this data informed decisions.

"Throughout the course of the past two years, we've seen an increasing trend in business leaders asking for and considering the data piece before making decisions which is a big step from the 'gut-feeling' or intuition driven decisions that were more the norm previously.

An example is A/B testing headlines in the newsroom. When editors have differing opinions on which headline works and which doesn't, putting it in front of users to see which one they click most, identifies what works, "regardless of which one we might like".

It has also helped identify language usage issues such as wether to use 'row' or 'war' in coverage of the US-China trade war. Lee says several headlines were written in the British style with 'trade row', but since most audience is now coming from the US, 'trade war' worked better.

"Americans generally read 'row' as 'row, row, row your boat' not as 'fight'," he says.

He says the best data culture is an environment where everyone has the knowledge and ability to access the data they need. "Each person across the company should understand the value of data to their team as well as their individual work and be able to have self-serve access to actionable data to execute on their day to day work.

"Data tools should be largely automated and the data team's focus should be focussed primarily on generating insights, enhancing recommendations, and building tools that employ AI, ML and algorithms that help drive the business onwards and upwards."

Lee says that while there is still plenty of work to be done in order to become a data-first organisation, "we see a clear path towards building this culture into our future as a step in the right direction".

Based on a post for INMA's Big Data For News Publishers blog, with permission