What does big data mean for microinsurance?

There are 24.2 million adults in Tanzania. About 15 million experienced an insurable event in 2013. Less than 3 million of those reported to have insurance and even less (200,000) actually used insurance to manage their risk. In other words, 14.8 million adults had to rely on other risk-coping mechanisms.

Insurance is important, but hard. Microinsurance in developing countries is even more important, but even harder. Low incomes require low premiums and therefore small policies. The lack of information in these markets makes it harder to design and viably deliver these products. 

So why are we excited?

Big data is providing new information in these markets that can be used to take on this very challenge. Big data refers to large or complex datasets. We live in an era where 2.5 quintillion bytes of data are generated daily (that is enough data to fill 57.5 billion 32GB iPads!) An increasing number of sectors are putting big data to use, from healthcare to retailers to sports. The analysing of this data has made a new generation of market intelligence possible: it helps decision-makers and service providers get to grips with the features, preferences and propensities of existing and potential clients; and informs them of the most effective means and modes of communicating with customers. Insurance is no exception. The industry is moving beyond structured data for reactive business models to incorporating unstructured data for preventative business models. ‘Big data and analytics’ was the most popular insurance industry buzzword in 2015 (Kumbla, 2016).

Whilst developed countries have a superabundance of data, emerging markets have typically suffered from a lack of traditional information, particularly for low-income consumers. This lack of data available to providers – or information asymmetries, in economic lingo – has meant financial exclusion for many. For example, without data on financial records to allow credit scoring, low-income customers struggle to access formal credit. However, providers are now starting to leverage digital data (mobile phone usage, social media, etc.) to develop new scoring models that enable extension of previously unviable credit offerings.

Similarly, big data is also opening doors to reach currently uninsured populations. The potential benefits span the entire microinsurance product lifecycle. For customer acquisition and distribution, analysis of behavioural data can provide indicators of consumers’ propensity to take up insurance offers and to continue paying premiums, resulting in improved targeting of the sales and distribution effort – and therefore lower acquisition costs. For example, tracked call records, where calls clustered outside typical working hours could indicate a level of formal employment and therefore more stable income, meaning increased ability to pay regular insurance premiums. In addition, preferences and behavioural traits revealed through social media interactions, could inform which distribution channels or marketing strategies to prioritise, and make for a more tailored product design.

By applying sophisticated algorithms to big data, providers can also better understand and model the risks they are underwriting and be more efficient at premium pricing. This could include learning what diseases are more prevalent in various locations, by analysing outpatient data; or optimising price and collection frequencies, by assessing financial behaviour and affordability for a specific target audience.

In claims processing, big data can help speed up the process and reduce assessment costs by quickly identifying fraudulent claims through better profiling and predictive modelling. For example, where claims are logged via voice or video calls, data techniques and voice recognition software can look for certain words and changes in tone of voice that can be used as indicators of potentially fraudulent claims.

Nigel Bowman of Inclusivity Solutions, a third-party solutions provider in the mobile microinsurance space, explains the biggest potential for using big data in the microinsurance arena lies in getting the distribution and sales channel right and reducing churn. By understanding customers’ propensity to buy and tendency to stick with the brand, microinsurance operations can improve customer relationship management thereby reducing costs and increasing the ability to scale. Ultimately, this could make for a more sustainable industry.

However, there are also a number of potential challenges and pitfalls. Even though increasing volumes of data are available, it isn’t always accessible. In the case of mobile datasets, this requires mobile operator willingness to share the data and raises considerations regarding data protection of personal information. There can also be regulatory restrictions in some markets on what data can be shared. Furthermore, it can be daunting to figure out where to begin. In order to extract maximum value it is important to start by asking the right questions and to have a clear goal as to what you want big data to answer. And of course, the results of analyses can only ever be as good as the data on which they are based, so data quality is key.

There are also concerns that using big data to select which risks to cover could lead to certain individuals being excluded if they are considered too risky. This ability for more differentiated risk selection, which is arguably the biggest benefit of big data for traditional insurance, would go against the grain of microinsurance. Microinsurance is meant to be inclusive and protect the previously unprotected from risks that could devastate their lives.

So what’s the verdict? It is clear that big data holds much promise for expanding access to microinsurance in developing countries such as Tanzania if the challenges can be bridged. Now let’s see how it’s applied and who the early innovators will be.

To better understand the benefits and potential pitfalls of using big data within decision making, Insight2impact is conducting a scoping survey of data and analytics applications within financial service providers across developing countries. One particular area of focus is the role of big data in insurance, and specifically microinsurance (report due early 2017). Cenfri is further considering this topic as part of its thematic work on understanding the implications of insurtech for inclusive insurance (report due early 2017).