What's Happening?

Big Data and the Intermediary

SHARE THIS POST

By Precious Nduli, Head of Marketing and Technical Marketing, Discovery Insure

 

Technology is a cog in the machinery of insurance, where the intermediary still has a valuable role to play. 

Big data, machine learning, and artificial intelligence have become buzzwords in the insurance industry, and for good reason. In 2018, global spend on artificial intelligence increased by 44% to $35 billion. If we were to believe Moore’s law, this will increase exponentially, as these techniques pervade into essentially every other industry.  

Today, these techniques are already being applied in every major industry. For example, machine learning and artificial intelligence are useful for fraud detection, predicting when customer relations are deteriorating, image classification, dynamic pricing (such as how Uber prices can change depending on how many customers there are), and market research and sentiment analysis. 

In the insurance world, machine learning has been applied in quoting, the settlement of claims, chatbots, and a variety of other uses. But you may still wonder how this can help you, the intermediary, directly. This article aims to shed some light on that. 

Firstly, big data enables the effective usage of machine learning and artificial intelligence. Machine learning techniques are useless without data, and, the more data a machine learning model has, the more accurate it can be.

 

What is big data? 

Big data refers to large sets of information that are too large to be analysed without the aid of a computer. This could mean anywhere from a few thousand points of data to billions and trillions. It is estimated that Google processed at least 200 petabytes of data every day in 2019. This is the equivalent of 200 million gigabytes a day (far too big for a human – or even a single computer – to handle).

All large insurers have big data, too. This is in their clients’ and claims records, as well as any other data the insurer may wish to keep. Insurers that offer telematics devices have far superior data. Instead of gathering information only when the client interacts with the insurer, data can be gathered from clients continuously. This enables far superior models to be built such as those used in insurance pricing but also in the creation of client and intermediary solutions.

Below are some ideas that could impact you directly and are being implemented in the industry right now.

 

Using big data to target the right clients

An insurer can use a sophisticated model to look at clients’ rating factors and recommend the most valuable insurance cover based on their circumstances. 

One way of doing this is by using machine learning to analyse all the insurer’s customers. An insurer may have some insight into how valuable their products are to certain population groups. It can gain this insight by looking at take-up rates, lapses, claims experience, and engagement in any of the insurer’s initiatives. A large insurer could have millions of such data points from their interactions with clients. It can then feed a machine learning algorithm with this data to predict the value of each product to each customer.

For instance, household contents insurance might be more valuable to a parent with many possessions than a bachelor with very few possessions. 

After the value of each product has been calculated, the model can recommend the most valuable product to each customer at any point in time. Intermediaries can then use this tool to sell the most appropriate policies or benefit enhancements to their customers. 

For example, the model may recognize that most people in a particular region who take out buildings insurance lapse quickly. This might be because such clients get many different offers for buildings insurance while they live in that region. Buildings insurance may thus not be particularly valuable to these clients in comparison to other covers.

Instead of spending time selling buildings insurance, it might be more worthwhile for both the client and the advisor to sell a different product that may be valuable to the client for many years to come.

At the same time, this recommendation tool makes servicing easier, and that comes with a whole host of auxiliary benefits. These include:

  • Significantly lower lapse rates in clients that are serviced regularly
  • Increased customer satisfaction, loyalty, and brand support
  • Increased word of mouth, which fuels growth even further.

Not only can the model recommend which products are most appropriate, but it can also supply campaign mailers, so that the intermediary has an even easier time communicating with their clients.

It is important to note that a tool for recommending benefits cannot replace a financial advisor. A financial advisor can offer tailored solutions that take into account all kinds of details that a model could not. For example, a financial advisor could consider how financially savvy a client is before recommending a complex product.

Thanks to our Vitality Drive programme, and the huge sets of data we can analyse, Discovery Insure will be releasing a model similar to that described above later this year.

 

Using group-wide data to price more accurately

Big data can have applications outside of just recommending products and benefits. It can also be used to give the right clients rewards for being loyal.

One way of doing this is by giving each client a score based on:

  • Other products that they have with the same company
  • Their engagement with the company’s initiatives or rewards programme
  • Other behavioural variables, such the client’s risk habits.

We call this score a select score. For example, if a client already has life insurance and a bank account with the same company, they could get a high select score. This is because we have seen that clients with more than one Discovery product are better risks, and their premium should reflect this.

This score can be used to give clients discounts to make the insurer’s value proposition even more attractive for both clients and advisors. This acts as an integration benefit and encourages clients to enjoy the full range of benefits that the company offers.

The level of discount can be set by looking through millions of points of data and using an algorithm to quantify the impact of the discount on the client, intermediary, and insurer.

Thanks to our data from Discovery Group, Discovery Insure has been able to implement the above system while also taking into account multiple variables, including wellness choices, credit and retail risk profile, risk attitudes and benefit choices.

Using machine learning and artificial intelligence is one of the easiest and most useful ways of getting insight into a huge amount of data. This usage of big data is becoming increasingly relevant to both clients and insurers. Every day, people are finding new ways to analyse and implement big data to their advantage. 

This is only a tool, and it covers only one step in the cycle. Intermediaries play a different role in the cycle. Our philosophy is that value can be attained by combining the usage of big data with the role of the intermediary, thereby giving intermediaries the tools they need to provide the best, most relevant service to their clients.