News AirPlus 13 Mar 2018

Go Big (data) or go home

| 13 Mar 2018

The 3rd Asia Conference on Big Data and Analytics for Insurance heard industry experts exchange insights aplenty from strategy and case studies, to recruitment and even a take glimpse at the next data frontier. Here are our top 10 takeaways.

  1. Data cannot drive strategy

The question of which would come first in the triad of “vision, data and strategy” was decisively answered by panel experts which include Mr Vincent Shi, Managing Director of SCOR Global Distribution Solutions Asia, Mr Bill Lee, Managing Director of Azendian Solutions, and Mr Neil Gardner, Chief Customer Officer – Asia, Generali.


"Data should never come first, that's a given," Mr Shi said, explaining that otherwise, businesses run the risk of letting a specific few data points determine their strategy, while ignoring or missing out on other relevant data points. “Always ask the business question first. What is my plan, or what is preventing me from achieving my plan? The answer to that is probably where your data project comes in,” he said.


Mr Lee meanwhile, described the issue as the difference between strategy and tactics. “At the strategic level, we can never let data tell us what to do. But on a tactical level, data can point us in the right direction. Data can tell us how to focus and understand our customer better.”


  1. Three key tech implications on insurance industry

There are three key implications of technological advancements for the insurance industry and these are 1) a change in risk pools, where the availability and capacity to process even more data will impact the nature and intensity of traditional risks; 2) automation, which will allow not only for increased capacities but also enhanced quality in risk analyses, performed to more consistent standards for example; and 3) disruption of the insurance industry structure.


These impacts could potentially catalysed or inhibited by factors such as regulations, consumer behaviour and competitors, among others, said Dr Ladina Caviezel, Analytics Consultant, Group Digital & Information Services Swiss Re.


  1. Data analytics objectives and maturity – we’re not there yet

It does not matter what the names or terms are, the three fundamental objectives of data analytics is to achieve revenue and service enhancement, resource and operations optimisation, and risk management in any organisation, said Mr Bill Lee, Managing Director of Azendian Solutions.


He outlined the three main levels of analytics maturity and what they entailed. The basic level is descriptive analytics, where insights gained are more reactive and relate to past experience, and is also the stage at which most organisations are at. A level two maturity reflects analytics that can provide insights of a predictive and forecast nature, allowing organisations to “uncover what you do not know”. Finally, the third level of analytics maturity – prescriptive analytics – is where companies are able to use data to optimise operations and resources that ultimately translate to risk management and cost savings.


  1. If you want to invest in analytics, infrastructure is crucial

Mr Lee also shared practical advice on how to get started on adopting analytics in an organisation.


The analytics value chain consists of five pillars: 1) data warehouse design and implementation; 2) the ETL (extract, transform, load) process; 3) prediction forecasting optimisation; 4) dashboard reporting and visualisation; and 5) organisational change management.


Of these, he highlighted that the first, second and fourth pillars are the most important, and that “for every dollar spent, 50 to 60 cents” should be allocated to infrastructure.


  1. Organisational culture and work environment determines talent attraction

Hiring the right data science talent for the company in itself is almost a science. Mr Raymond Au, Head of Data Science at Allianz Asia Lab, said the question insurers must ask is not what “can fit”, but what “can be happy” in their organisation. Context as to what the role is, and why and where it is needed, all play a part.


A key determinant in being able to attract suitable talent lies in the company’s data science readiness, not only in terms of infrastructure sophistication, but also in terms of the support of the respective senior management and business functions. “Your environment determines its people. In other words, you get who you deserve.”


  1. Will actuaries eventually be replaced by data scientists?

The short answer is “No”, although in order to adapt to advancements in tech and analytics, as well as the changing business landscape, actuaries will eventually need to enhance their existing skill set with some data science knowledge, agreed panellists Ms Mandy Luo, Chief Actuary & Head of Data Analytics, ReMark International; Mr Lee Sarkin, Head of Data Analytics, Life & Health, Munich Re Asia Pacific, Middle East & Africa; Mr Tan Suee Chieh, Former Group CEO, NTUC Enterprise Co-operative Ltd; and Mr Simon Lee, Chief Data Scientist, AXA Hong Kong.


Fundamentally, whether an actuary or data scientist, Mr Tan highlighted that it is all about maintaining the right mindset and continuous learning, or else “whatever skill set you have will become outdated eventually”.


  1. Scaling AI adoption: It (almost) takes a village

Adopting AI throughout one’s operations, be it in mining for customer opinion and satisfaction, or attrition prediction among others, will depend on the company’s strategy and choice to optimise its use where the organisation deems to generate the most ROI.


What’s interesting is that the most daunting challenge in embarking on the AI journey, is not the acquisition of technology, but in getting the dream team to execute the vision, according to Mr Albert Antoine, Senior Vice President of Asia Pacific for Dataiku.


He and fellow panellists, Mr Ofir Shalev, Chief Technology Officer at CXA Group, and Mr Vikram Mengi, Co-founder & CEO of Latize, noted that aside from data having to be integral to the company, senior management and Boards must also understand the AI strategy in order to drive it. The Chief Analytics Officer (or Chief Technology Officer) must be the able strategist and politician in managing stakeholder relations and buy-ins, while insurers must also hire the right quality and quantity of data scientists for the project.


  1. The increasing need for blockchain

The need for and use of blockchain will increase with time and change the way business is done in a data-driven world. And because of InsurTech and exponential technologies, together with the customers’ desire to change, things will move very fast, said Mr David Piesse, Advisory Board Member for Guardtime and Asia Pacific Ambassador of the International Insurance Society (IIS).


Among the several blockchain trends he outlined, emphasis continues to remain on ensuring data provenance and integrity. Even though focus on data security is increasing, Mr Piesse noted that this is not sufficient as encrypted data does not necessarily equate to data integrity.


And as governments continue to step up efforts on consumer privacy and protection, such as the implementation of the GDPR in Europe, he added that organisations may be more inclined to adopting blockchain.


  1. Data quality remains crucial: Garbage in is still garbage out

Before investing in complex systems and infrastructure, Mr Nick Mair, CEO of Atticus Associates, reminded the audience not to neglect their “small data”, or internal data.


The Harvard Business Review found bad data can cost insurers 10-15% of annual revenue, and that a 1% error rate can double the cost of a transaction. Thus, ensuring the quality of one’s internal data should not be overlooked. Combining good external data with bad internal core data, he said, is a recipe for disaster. Much like supercar engines, one cannot possibly expect a Ferrari to run on diesel.


Mr Mair offered a couple of actionable takeaways for companies. First, to conduct simple audits, looking at data flows, pinning down pain points, and focusing on material data (using the 80/20 rule). Second, he recommended to start with creating a simple data control framework that can be built up over time.


  1. From IoT to IoU

The Internet of Things (IoT) may be an excellent data source to harness customer insights, but it is limited to the network of physical objects. It appears that the next level in harnessing data and to “truly make it smart” is to look at the “Internet of You” (IoU). And the key to unlocking IoU is in sensor data, said Mr Tom Vandendooren, Chief Business Development Officer for Sentiance.


Ambient intelligence is the art of sensing, understanding and predicting behaviour from data sources, such as the smartphone, which is carried by virtually everyone everywhere. Sensor data collected promises real-time, continuous, dynamic and objective insights – for example, shifting from providing protection to enabling prevention – to the organisation that is able to make sense of the information.


This being said, Mr Vandendooren noted that the industry is “not there yet”. A key challenge, he said, lies in understanding the context in which data is collected.



The conference, which drew more than 100 delegates, was organised by Asia Insurance Review and sponsored by ReMark.

| Print | Share

Enter the code shown above in the box below.

Note that your comment may be edited or removed in the future, and that your comment may appear alongside the original article on websites other than this one.


Recent Comments

There are no comments submitted yet. Do you have an interesting opinion? Then be the first to post a comment.

Other News

Follow Asia Insurance Review