AXA Insurance Pte Ltd | Singapore

Introduction

We are in the midst of a digital age, surrounded by an ever-increasing volume of data which is estimated to grow to 181 zettabytes by 20251 (1 zettabyte = 1 trillion gigabytes). Accompanying this explosion of data is the need (real or perceived) to come up with use cases to utilise such data.Under the weight of these use cases, centralised teams of data analysts and data scientists face challenges in deriving and serving insights in a timely manner. Instead, a key differentiator for companies nowis the ability of their business units to embed data-driven decision making into their daily work.

Even though the value of a data-driven culture is widely recognised, it is not an easy task to implement in traditional organisations. Employees face a multitude of issues such as a lack of training, uncertainty over how to handle data and following their intuition2. Hence, it is important to orchestrate a strong foundation to prepare teams for the transition to a data-driven culture.

Organisational Change

Adopting a data-driven approach commonly entails organisational and process changes. Hence, one of the key elements for it to succeed is strong support from top management and senior leaders.They would need to lead by example by trusting the data more than their intuition gained over many years of experience.A robust framework would also need to be designed to determine how and which data points should be selected for use in the decision making.

 Adopting a datadriven approach commonly entails organisational and process changes 

Senior leaders also need to walk the talk by allocating resources such as funding and manpower to support the change. Data management and governance would be essential to prevent poor data quality from skewing the decision making, and these initiatives require executive sponsorship. Technology and appropriate tooling will help to smoothen the process of retrieving and analysing data as well.

Infrastructure Requirements

In order to have a sustainable data-driven approach, the underlying data infrastructure must be able to support the myriad requirements of the business. Established firms would find themselves struggling with legacy data warehouses or data lakes which require time and effort to extract the granular data points that are needed to drive decisions. Depending on the use cases and system complexity, it may be worthwhile to explore newer technology stacks such as data lakehouse, data virtualisation or even data fabric which can overlay the existing legacy technology to meet the immediate needs of the business.

It is also vital to conduct a thorough data discovery exercise to obtain a clear understanding of the data siloes within the organisation. This exercise has the added benefit of exposing the interdependencies between various data sources which will be used to identify the sources of truth during the integration stage.

Building Capabilities

One of the most challenging dilemmas companies face in adopting a data-driven approach is striking abalance between “buy or build”. While there is no one-size-fits-all solution, it is advisable for companies with minimal experience to engage professional services to set up the data infrastructure. This allows for knowledge transfer with the correct architecture and prevents technical debt. It is common for small enterprises to fully build their own infrastructure, only to later realise that it is not scalable and significant rework is needed.

For the same reason, companies should focus first on building their foundational capabilities like data engineering and data governance with a small core of data analysts to train business units on the data-driven mindset. Without these foundational capabilities to ensure high quality and timely data, the business will quickly get frustrated and the data-driven approach will fail.