Big Data: Boosting Insurance Development and Innovation

Aptly known as the “prophet of the big data era”, Viktor Mayer-Schönberger observed in his work entitled Big Data: The Revolution That Will Transform How We Live, Work and Think that “Big data marks the beginning of a major transformation … Just as the telescope enabled us to comprehend the universe and the microscope allowed us to understand germs, new techniques for collecting and analyzing huge bodies of data will help us make sense of our world in ways we are just starting to appreciate”. Big data has become the driving force behind new inventions and new services, and more changes are on the way.

As Mayer-Schönberger pointed out, big data is now present in all aspects of our society from technology, government and healthcare services to education and economic activities, and it is reshaping the way we live, work and think at an alarming speed. Data are the building blocks of the insurance industry, where data resources abound. Many insurance firms have started tentative applications of big data technology, hoping to notch up further development and innovations through big data analysis. While applying big data to drive business growth, insurance companies should also adapt their strategies and make the most of big data in marketing, product development, pricing, service, and management operations, thereby transforming themselves into "next-generation" insurance businesses.

First, we need to know what big data actually means.

I. The notion of big data

During every single minute in 2015, there were more than 4.16 million “likes” for Facebook posts on the internet, nearly 700 trips transacted on Uber and 51,000 Apple apps downloaded. The same level of activity is seen on other social media platforms – every minute, 300 hours of new video clips are uploaded to YouTube, 347,000 posts are tweeted on Twitter, and 110,000 voice calls are made via Skype, not to mention innumerous ecommerce and blog posts, photo uploads, and more.

The internet has gone through a period of explosive growth. During the last two to three years, in particular, the use of social media has increased exponentially, generating a tremendous amount of data. As we reach a point when certain collections of data can no longer be perceived, obtained, managed or processed using traditional IT technologies and hardware/software within a specified time limit, we need new data analysis techniques, hence the introduction of big data technology.

Gartner, a technology research institute, defines big data as “high-volume, high-velocity and/or high-variety information assets that demand cost-effective, innovative forms of information processing that enable enhanced insight, decision making, and process automation”. The definition given by Wikipedia is as follows: “Big data is a term for data sets that are so large or complex that traditional data processing applications are inadequate”. Big data are characterised by high volume, high velocity, high variety and high value. The strategic value of big data technology lies in the professional processing of meaningful data, rather than the acquisition of a huge amount of data itself.

II. Three shifts in the big data era

With the advent of the digital age, data processing has become easier and faster, enabling us to process massive datasets within a split second. In essence, big data technology requires three “shifts” in our way of thinking, which will change our perception of society and the way it is constructed:

The first shift concerns the “increase in data volume”. In the big data era, we have the capability of processing more data – sometimes even the entire body of data available concerning a specific phenomenon – instead of relying on random sampling.

The second shift dictates that we must accept an “increased variety of data”. Presented with massive amounts of data, we are no longer obsessed with data accuracy. Instead, the emphasis shifts to data variety. Data quality, i.e. minimum inaccuracy, is the most fundamental and important requirement for “small data”, where we must ensure the accuracy of data recorded due to the limited amount of information available. However, as we now have access to a vast amount of data, accuracy is no longer the prime requirement; we can gain insights into development trends without all the data being perfectly accurate. As far as big data are concerned, perfect accuracy is not only unnecessary but also impossible to achieve.

The third shift is the product of the first two shifts – that is, we will be more interested in establishing “relative relationships” than causal relationships. If millions of electronic medical records suggest that cancer can be cured by combining orange juice with aspirin, discovering the cure itself is more important than finding out the actual pharmacological mechanism. Similarly, once we define the best time to book flight tickets, the actual cause of wild fluctuations in ticket prices is irrelevant. Big data tell us “what” and “how” but not “why”.

III. Implications of big data technology for the insurance industry

As far as big data application is concerned, the insurance industry has unique advantages. First, it abounds with data resources; second, insurance businesses are operated based on the law of large numbers, and estimation – which is part and parcel of big data technology – is the most powerful technique in assessing probabilities. The influence of big data is present throughout routine insurance operations and the insurance value chain, especially during marketing, product development, pricing, service and management activities.

1. Customer segmentation and precision marketing

Insurance companies need to collect data concerning insurance customers as well as information beyond the insurance system, e.g. data derived from partners, and social and behavioural data retrievable on the internet. Such data are all relevant to insurance market segmentation. Through in-depth analyses of customer information and behaviour, insurance companies acquire an understanding of customers’ needs, allowing insurers to identify potential customers and recommend suitable products to prospects, and ultimately implement precision marketing that sets them apart from competitors.

In the big data era, instead of trying to attract different groups of consumers using the same advertisement or applying the same marketing technique, marketers should carry out promotional advertising customized to refined customer segments. Categorization of behavioural and intent data enables insurance companies to understand how much a specific user group is interested in certain products, thereby informing insurers in selecting the most suitable marketing techniques. For example, health and mobile related accident insurance should be marketed to consumers who spend more than five hours a day on mobile internet surfing; critical illness insurance should be recommended to meat eaters who also drink a lot; insurance for broken screens should be recommended to users of big-screen phones; and travel related products should be recommended to Ctrip and Tuniu users.

Insurance companies can identify target customers through accurate data analysis, conduct well-targeted marketing and promote relevant insurance products according to customers’ individual needs, so as to avoid offending potential customers by sending mass advertisements.

2. Personalized product development based on a clear understanding of customers’ needs

While developing new products, insurance companies typically innovate, modify or bundle existing products according to market demand, and enhance their own competitiveness. Once we have a clear understanding of customers’ needs, we can estimate potential losses based on relevant big data and existing data about risks associated with the insurance coverage. This allows insurance companies to develop personalized products tailored to the actual needs of different customer groups.

On 29 September 2013, several famous Chinese companies including Alibaba, Tencent, Ping An and Ctrip established Zhong An Online Property Insurance Co., Ltd., the first internet-based insurance company in China. It launched innovative insurance products covering shipping costs incurred for returned goods, account security risks, credit card fraud and debt guarantee, among others. The insurance industry has started to apply the “internet thinking” in product development.

3. Pricing optimization system: “big data” vs. “law of large numbers”

Some people argue that big data debunked the law of large numbers. In reality, although both aim to discover patterns and analyze and predict risks based on a large amount of data, big data and the law of large numbers fulfill different purposes within the pricing mechanism for insurance products.

Insurance is the science of risk management. The law of large numbers, with its roots in statistics, is at the core of the actuarial theory. It lays the foundation for ensuring adequacy and impartiality in the determination of the industry benchmark pure risk loss ratio. In other words, the law of large numbers governs mathematical logic for insurance operations and management, and is unshakable as the theoretical and pricing underpinning of the insurance industry.

On the other hand, big data plays a supportive role in insurance product pricing. It involves collecting online data about customers’ transactions, searches and other behaviours to identify behavioural patterns through relevance analysis, enabling insurance companies to refine customer segmentation, optimize actuarial pricing models, and work out effective floating insurance rates and differentiated pricing mechanisms.

Therefore, instead of debunking the law of large numbers, big data serve to optimize and enhance the market-based insurance pricing mechanism, applying the latest technology to refine risk management operations.

4. Settlement of insurance claims with improved accuracy and timeliness

Claims management is critical to an insurance company’s profitability credibility, and big data allows settlement of insurance claims with improved accuracy and timeliness.

In the big data era, insurance companies have access to real-time information of customers involved in accidents covered by insurance. Therefore, they can contact the customers proactively and offer claim settlement services. Various big data techniques are suitable for different accident reporting methods. For example, the insurance company can locate the customer if an accident is reported by telephone, or obtain the exact location of the accident through GPS. Pictures of the scene can be remotely submitted or uploaded to identify the cause of the accident and determine the responsibilities of parties involved through image analysis and mining technology. This enables insurers to settle claims more efficiently and therefore enhance the overall customer experience.

Big data can also be obtained from third-party partners. For example, if a customer has an accident on the motorway and reports it to the police, the insurance company can access relevant information in a timely manner and process the relevant compensation payment without the customer ever filing a claim; insurance companies can obtain vehicle repair and maintenance data through cooperation with franchised car dealerships, enabling surveyors to make appropriate and reasonable claim settlement decisions based on readily available data about the vehicles; victims of accidental injuries do not need to visit insurance companies to file claims, and hospitals can transmit information directly to insurance companies. Therefore, advance payments can be directly transferred to customers’ accounts to meet their urgent needs.

5. Easing information asymmetry about the insured item and fraud prevention

Big data can also improve insurance companies’ management capabilities and prevent insurance fraud. Information asymmetry is at the root of insurance fraud; big data can partially reduce information asymmetry and therefore lower compensation costs on the part of insurance companies, safeguarding their legitimate rights and interests. In terms of fraud prevention, insurance companies generally rely on a set of fixed assessment criteria and the personal experience of their claim processing staff to filter out fraudulent claims. Due to the absence of a collaboration mechanism or information sharing platform in the insurance industry, the quality of claim investigations largely depends on claim settlement officers’ empirical skills and sense of responsibility, and cooperation with the police authorities. The strategic value of big data technology is not limited to the availability of high-volume data. More importantly, the data are professionally analyzed and processed to discover the otherwise hidden useful information that can translate into economic and other benefits. Some foreign insurance companies have already started applying big data to insurance fraud prevention and accumulated extensive experience in this area, setting an example that should be followed by insurance companies in China.

Applying big data in the insurance industry is a promising direction going forward, but we should also be aware of the issues and challenges involved in the transformation. Firstly, the growing amount of personal confidential information involved in big data analysis raises the need for improved data storage security  and privacy protection. Secondly, given their currently limited big data processing capabilities, insurance companies are confronted with a serious challenge as to how their own data resources can be put to effective use. Furthermore, an ever-increasing number of internet companies and technology companies capable of big data processing have entered the insurance industry, making it a more competitive marketplace.

All in all, the arrival of big data technology brings with it endless possibilities for transformative development. Insurance companies should adapt themselves to “internet thinking”, and keep building on their big data processing capabilities, incorporating big data technology into their marketing, product development, pricing, service and management operations. This way, we will be able to take China’s insurance industry to a new level and promote insurance innovation.


  • Viktor Mayer-Schönberger, Kenneth Cukier. Big Data: A Revolution That Will Transform How We Live, Work, and Think.
  • Chris Smith. What happens on the Internet in one minute?
  • Big data, from Wikipedia, the free encyclopedia.
  • Big data, from Gartner.
  • The Boston Consulting Group. The Age of Internet Plus: Improving Big Data and Reforming China’s Insurance Industry. August 2015.
  • Wang Wei. An Exploration of Big Data Applications in the Insurance Industry. Financial Computer of China. April 2015.


Published July 2016