Data Science and Analytics use in the Media Industry
When Media Prima embarked upon its digital transformation (DX) journey two years ago, data science was to play a huge role. Head of Data Science and Analytics of Media Prima’s digital unit, Peter Kua, knew they had to reduce the gap between IT and its data science division.
In doing so, he shed more light upon what a serious and pervasive data science initiative actually entails. “The most important imperative was to improve all digital touchpoints across the consumer lifecycle, to optimise content, sales and marketing, enhance brand value and achieve operational efficiency.
“In order to do that, we had to create high-value, data-driven use cases.”
“But, what about the part where you need IT’s help? They are most of time, the custodians of all the systems you need to get data from,” he pointed out.
“We need them to help push data on a daily basis, to data engineers. (IT) are the ones that export raw data to our data engineers who in turn extract, transform and load data to the data warehouse. In addition, IT helps prepare the digital infrastructure to operationalise analytics and ensures the infra is secure and reliable.”
Dog eat dog industry
In a fiercely competitive landscape where unique customer experiences are the key differentiators between brands in an industry, a customer’s interaction with a single brand could be via various different touchpoints. One touchpoint, for example a website, may have customer data going into siloed systems.
The more touchpoints a customer goes to, the more systems that are involved.
“So, how do you make sure that the customer’s identity in these different systems are in sync and that you have a single holistic view of them?” Kua asked, adding that for Media Prima Digital, their systems are linked and they are able to extract and combine data from various systems to build a customer profile.
“There are certain industries that are a lot more advanced (in the use of data science and analytics), but the media industry has started recognising they were in trouble when channels like Facebook and Google started taking in 80-percent of digital ad spending in Malaysia.”
Warning signs
Aware that traditional advertisement revenue has been dropping in recent years, Media Prima, a leading integrated media company is ramping up their digital ad revenue and using data analytics to provide personalised and targeted ads.
“Leveraging data for better ad targeting to our audience is one of the most important use case for our data science division,” said Kua.
The data science team also enhances Media Prima’s digital properties via personalisation and content recommendation.
But it isn’t all about retargeting advertisements only – digital properties that belong to the media group, is also leveraged to promote brand recall.
The two-year old data science team is starting to see results, with a Customer 360 view developed and deployed, allowing the media company to obtain a complete view of customers by aggregating data and matching a single person across the various touch points/digital properties that they land on.
Also of note, is a Natural Language Processing (NLP) Engine, which was developed and deployed for advanced content recommendation, tagging, brand safety, categorisation and search based on English and Malay languages.
Progress for data science
He said, “Anything significant in terms of results, needs an investment of 2 to 3 years. Improvement won’t be that fast. When you start a data science team, there has to be a series of progressions that you make.
“In the first year, it was about setting up structures and standard operating procedures (SOPs). This includes getting other divisions to take on a data driven mindset, and can start with simple stuff like business intelligence.”
The second year saw more advanced data science methods like predicting the kinds of ads a user would like to see, to encourage higher engagement with the ad.
“There is steady progress and we are confident of achieving our KPIs in the next couple of years,” Kua said adding that there is an ongoing strategy to roll out more high-value data-driven use cases across the media group.
Challenges for the industry
Kua also wears the hat of a big data and analytics advocate, and actively conducts classes on big data strategy and applications.
He observed that government-linked corporations in Malaysia understand the importance of being data-driven, and noted progress in maturity of adoption since 2013.
“Now, there would be some form of structure and budget in place in most big organisations like our incumbent utilities providers,” he shared.
“You can be ready (for big data and analytics) when you want to be. Big data is about your strategy – do you have gaps in your business unit that can be solved by data science?
“So, you have to internalise the big data (mindset) and brainstorm about what your current gaps are and from there, launch your data science programme with a list of high impact use cases that will impact your bottomline.”