Data Deep Dive
How can brokers leverage big-data analytics? You have more options available to you than you might think
Amount: The amount of data being produced today is staggering and is only expected to grow exponentially. This data provides opportunities for both carriers and brokers to better understand their risk profile. Integrating new social data along with traditional data sources, such as policy and claim information, is facilitating a deeper understanding of your particular risk exposures. From a brokerage perspective, this can support advanced risk selection and placement activities. The sophistication of these analytic efforts is only expected to increase as the consumption of big data increases.
Speed: Data today is available for analysis in real time from a variety of sources, including traditional quote, policy or claims information and more sophisticated contemporary social media and other unstructured data. The speed at which information is available will continue to increase in lockstep with the adoption of mass mobile technology and the speed at which people can communicate. The way the data is being delivered is also changing. In the past systems would update overnight for using batch processes. Today data is streaming live and is updated continuously, implying the need for continuous data analysis. For brokerages this means more efficiently managing both the incoming flow of data, which is increasing, and also monitoring the data affecting our clients.
Type: There are two major data types: structured and unstructured. Structured data is standardized and easily entered and managed information. Examples include numbers or pre-designated codes for financial information, lines-of-business product classification, etc. Typically, it has a defined range of values and a fixed length. Conversely, unstructured data is free-form, non-standard and not bound by type or medium. Examples include free-form emails, imaged documents, photos, video recordings, etc. Effectively categorizing and maximizing the value you as a broker add to your clients is challenging in an increasingly unstructured data environment. Utilizing new tools and approaches may be necessary to properly capture that information’s business value.
In addition to these three attributes, companies with established management information practices are discussing adding additional attributes such as validity and visibility to these criteria. Regardless, the speed and availability of data has been challenging insurers and brokers in assessing what to do. The implication for the insurance industry and brokers is that big data requires building new, or enhancing existing, analytical applications on new data types to better serve customers and drive better competitive advantage. So what are those types of data?
Types of Big Data
The move to enhanced analytics is predicated on more sophisticated data classification and categorization within your organization (sometimes referred to as an “information domain”). From a business decision making and analytics perspective the following are typical classifications for ‘big data’ types.
Internal Data. Every entity, regardless of its size, has some internal data. Examples might be payroll information, client information or product information. This data provides an opportunity to further develop or enhance your client profile based on various types of strata, identifying strengths and challenges of your sales force, product mix and selection.
Social media is data generated through social sharing sites allowing people to connect with each other and from a marketing perspective, can expose a client’s brand affiliation and other positive/negative brand sentiment. Real time analysis of customer sentiment through social-media monitoring is increasingly common and is one way to establish a new communication channel with both current and prospective customers. We see brokers and carriers including social data as part of product development and sales channels for extensive marketing.
Machine data is systematic, technology-related data. It is also one of the largest areas of data generation. One example is mobile apps leveraging address geo-location based on the app user’s unique Internet or wireless location. This could allow the website owner to identify customer “hotspots,” (e.g. inquiries conducted by potential clients, the type of the products inquired about, etc). Similarly, the explosion in use of mobile devices and mobile browsing has created a unique opportunity for data analysis.
This type of data plays a major role for the insurance industry. From paying premiums to managing claims and checking other available insurance products, mobile apps make it easier for clients to remain connected to insurance providers. Understanding what information, product or other content is drawing customers to your sites will enable you to better position those products and map out customer trends.
Environment/Geophysical data. The industry has moved from classic weather forecasting to more sophisticated models that include predictive indicators for environmental concerns. This has been broadly used with mapping software and is now being supplemented with usage-based-insurance (UBI) products that leverage telematics data and other geophysical data points.
While UBI is relatively nascent in Canada, we see this as reflective of the foundation of better data analytics practices being pursued within the industry. Applying environmental and other geophysical data points to risk modeling and underwriting practices can be a point of differentiation for the sector. Through better profiling of your existing customer data, understating how to better position UBI may become possible.
Boots on the ground. Last but not least is the age-old and proven data-gathering technique of having people resident in a local market. There is no substitute to direct interaction with existing and potential clients. This conventional source of information coupled with modern data types and classification facilitates informative decision making processes.
Nothing compares to the knowledge a broker gets when meeting a client, prospect or other brokers. However, the single biggest challenge associated with this data type is leveraging it across your organization. The insurance sector is undergoing a change to broader knowledge management concepts, which we see as a differentiator going forward.
In addition, there continues to be a large amount of other government and industry data available. Leveraging statistical information from government census, Statistics Canada or industry groups continues. What is changing with big data is the ability to apply it to different data sources and apply different decision criteria. For example, identifying urban growth areas, propensity for rainfall as well as social media mentions of wet basements is an interesting intersection of current data points.
After briefly discussing the aspects and components of big data, the question that arises is, What does it all means from insurance brokerage perspective? Let’s try to answer this in terms of key opportunities and challenges.
“Slicing and dicing” a brokerage’s client data provides a wealth of information. For example, developing various client profiles; based on industry segments, geographic location, annual revenues clients produce and types of insurance products purchased. Client profiles should also include social media data sources as they provide a unique opportunity to increase the share of the wallet by offering different products or services.
Client profiling can also be done based on their business lifecycle. Clients in various stages of their lifecycle have different wants, needs and risk appetites. This type of profiling allows brokerages to develop products and services that will be unique to a client’s stage of business lifecycle and design new market segments by mixing various profiles. Brokers can also identify cross-business insurance needs based on a client’s sector. For example, transportation, processing and distribution can form compelling product bundling opportunities.
Operational efficiency can also be achieved through internal data analytics. Data analysis can assist in minimizing operational cost, uncovering fraud, and driving efficiencies. Current data allows for real-time monitoring of where risk exposures are occurring or likely to occur. Brokers can leverage predictive models based on past performance to identify improvement trends.
An insufficient or unclear data strategy can lead to a lack of analytics direction. Establishing your information and data strategy, along with its governance, is the first step to better analyzing your data.
Another challenge is establishing a talent management model or capacity for data analytics and business intelligence efforts. Not all companies or brokerages need a dedicated team, but supporting the effort to develop data analytics as a competency is a trend in the marketplace.
Additionally, many companies face data integrity issues at the tactical level and lack of integration between systems at the strategic level. Developing a roadmap for the resolution of these IT issues will help enable a more agile data environment and support advanced analytic efforts.
Ultimately big data holds the promise to enable better business insights and the ability to optimally serve your clients. The clear market winners in today’s data rich environment are the ones who take advantage of the benefits available due to big data. Turning that proverbial mountain of data into smart data – that’s the road to true information-based decision-making.
Allan Buitendag is national insurance consulting leader at PwC. He can be reached at email@example.com. Keegan Iles is director of insurance consulting at PwC. He can be reached at firstname.lastname@example.org.
Copyright 2013 Rogers Publishing Ltd. This article first appeared in the October 2013 edition of Canadian Insurance Top Broker magazine