In recent years, Insurtech companies have taken the lead in the digital transformation within the insurance industry. Data science and machine learning played, and are still playing, a decisive role in this transformation by distilling value from data. The traditional insurer will have to invest in these capabilities in order to remain competitive in the future, gain efficiency benefits and deliver added value to the consumer. Translating data into valuable actions will therefore become an essential competence within each insurance company.
This is, of course, easier said than done. Where do you start this transformation? What are the most valuable business cases? And what are the best practices? During the online round-table 'The power of data' organized by Triple A and Building Blocks, these questions and more were answered on the basis of practical examples presented by three insurance data experts.
Bart Kling, Head of Pricing & Data Analytics at Inshared emphasizes in his presentation that the use of predictive models should always fit the business proposition in order to be successful. Inshared strives to keep premiums low and to return as much of the earned premiums as possible to their policyholders. This is possible if there is money left over from the premiums reserved for claims. To achieve this, the analytics team at Inshared uses models to fight against fraud. By means of network analysis, individuals are linked to each other on and suspicious patterns can emerge. Based on these identification of possible fraud, the human agents can focus their efforts more effectively.
It is important that you automate where you can and be personal where you need to be, says Bart Kling. Models thus support and simplify the work of their fraud experts, for example by taking manual desk research out of hands. And the human employees focus their efforts on the most suspicious cases detected by the algorithm. In this way, models and humans form a golden duo in decreasing fraud and thereby reducing costs. Which leads that more of the previously earned premiums can be returned to their policy holders.
The possibilities for the use of data science within the insurer's value chain are diverse. Sophie Heethuis, Consultant Data Analytics at Triple A - Risk Finance introduces the customer value model in her presentation. Based on various data sources and models, the expected contribution of a consumer to the portfolio performance of the an insurer can be determined. For this, the expected loyalty of the insured and the expected profitability are taken into account. Sophie shows that these insights about customer value can then be applied in different ways. For example, it can be used to monitor portfolio performance. Do my marketing actions have the desired result? Do you attract the right customers that are profitable for your business?
In addition, it is also possible to steer on portfolio performance by using selective premium increases where the customer value is in proportion to the premium paid. There are also cases where data analysis can significantly contribute to the customer experience. For policyholders with a high customer value, claims can be handled more tolerant, thus boosting customer satisfaction. In addition, extra attention can be paid to policyholders with a high customer value who are likely to churn, by predicting so called 'next best actions' that prevent the customer from switching brands. It became clear that there is a range of applications within the entire value chain.
Many insurers and underwriters see the relevance of data science and a data-driven customer journey, but not everyone is putting it in practice yet. Erwin van Oosten, Captain Commercial Organisation at Building Blocks sees several reasons for this, but also offers solutions. The first reason mentioned is that data science still feels complex to people who don't deal with it on a daily basis. It is therefore important to show that it does not necessarily have to be complex and that it fits within – and delivers value to - the daily operations of the business of insurance companies. By working pragmatically and experimenting, a solution can be tested relatively quickly and be optimized and further developed over time.
But which business case is best to start your data science journey with? Erwin van Oosten explains that it is crucial to start thinking from the business perspective, so what are the most important KPIs you want to improve? And what factors contribute the performance of these KPIs? You want to have insight in the biggest bottlenecks of your customer journey, and thus in where most value can be delivered. From these data driven insights the most valuable business cases can be detected.
That insurance companies, intermediaries and underwriters can gain much value with their data is shown by the examples and concrete successes in practice. So do you as an insurer want to be successful with data science and machine learning? Then start thinking from the business, look for the most valuable business cases, align these with your strategy and apply the insights step-by-step throughout the value chain. In this way you guarantee long-term success and short-term results.
Download the presentations of the online-roundtable here.