Astro’s Analytics Journey
This is the fifth piece in a series of blogs by Phuah Aik Chong, Astro Chief Technology Officer
In one of my earlier articles, I briefly touched on the importance of Data Analytics across all aspects of our business.
As part of our 3-year digital transformation journey, we recognise that it is key for us to build up our internal Analytics capabilities to enable Astro to leverage on data to serve our customers better, and to stay competitive.
As we progress into this journey, we increase the number of data points available to the business, thus opening a new world of opportunities with Analytics. Each business unit has several Analytics initiatives underway and these initiatives have grown organically to further refine our understanding of our customers in both households and down to an individual level, ultimately enabling highly targeted personalisation across all our products and services.
We have enabled personalised content recommendation across our On Demand service on connected set top boxes, and video streaming apps Astro GO and NJOI Now. Similarly, personalised product recommendation is available on our Go Shop ecommerce web portal. There are more in the pipeline to come.
Challenges from legacy practices
As we accelerate the pace of enabling data driven features in our products and services, we have created an unprecedented "data wealth" which requires a systematic approach to data and industrialisation of our Analytics practice. Without such an approach, we take the risk of being "Data Rich" but "Insights Poor".
We do have a few challenges to overcome during the first few months of the journey in order for us to be able to really leverage on the rich data set we have, mainly due to how Analytics have evolved over the last few years in the organisation.
Our data is stored across a few different locations, on premise and in the cloud. Whilst we have a centralised Enterprise Data Warehouse (EDW) for the structured data from our PayTV business, the unstructured data from other digital products was stored across multiple systems in the cloud. It was complicated to understand the overall data landscape and time consuming to pull data across multiple sources for analysis.
There were existing Analytics and data skills in many parts of the organisation, with different maturity levels that sometimes operated in silos which meant we needed to strengthen the operating model and collaboration between the teams. The silo-ed engagement model limited the possibility to drive enterprise wide standards on data quality and best practices, and at the same time didn’t provide good visibility of the overall integrated Analytics roadmap across all business units.
The first few months were spent making sure we started off with the best foundation, putting the right organisation and processes in place and empowering our people with Data Driven decision tools powered by the latest technology.
A modern data driven organisation
To facilitate data collection from all end points, we are quickly evolving from a traditional Enterprise Data Warehouse reliant organisation to one that is driven by a Cloud based Data Lake.
When centralising our data into a Data Lake architecture, we realised that we needed a business focused centralised organisation to leverage on the richness of the data across business units to ensuring the pollination of information throughout the Astro group.
With a centralised organisation, we aim to:
- Provide an interface to the business leaders of each unit, helping them understand how to develop and refine their Analytics based Use Cases
- Capture all business initiatives and break down the business KPIs into Analytics capabilities before they become an implementation project led by the project office
- Add an Analytics component to all existing projects to ensure they are ready for future data analysis
- Enhance data collection and processing from existing systems
This Analytics Practice or Center of Excellence (COE) comprised:
- Data stewards in charge of the governance of data across all products and services
- Analytics practitioners that are currently spread across the group
- Data scientists who are already serving specific business units, plus new talents recruited Analytics mentors and coaches
- Oversight by Analytics experts from Amazon Web Services
These mentors and coaches enabled our workforce to "think data", back-up assumptions, and validate hypothesis with real data. They will increase our Exploratory Data Analysis capability among our employees and equip them with storytelling capabilities through data, giving greater insight to our business decision makers.
Data enabled Projects
Measuring success of an Analytics project may not be that straight forward and as such we make it a must to pre-define business KPIs for each Analytics project to ensure data points are made available for constant monitoring and re-adjustment if required.
With this approach, we aim to identify the opportunities for improvement, becoming a leaner, ROI-driven organisation, and only invest in projects and technology if they solve business problems at a realistic cost.
Our flexible, Cloud Based Data Lake architecture enables us to perform experiments faster by scaling up and down as we need, without requiring upfront investment through lengthy procurement processes.
This Data Lake collects, cleanses, and makes data available to Data Analysts, Data Scientists and ultimately the business decision makers, so that our business leaders can have better and faster insights into their areas of responsibility.
We are already beginning to see positive results from the Analytics investments we have made to date. We are very excited by the road ahead with the positive contributions these initiatives will have on our customers and business.
Click here to read previous blog entries.