In our first “Data-driven Insurance” series article we addressed the “opportunities and potentials” that arise from the application of data science in the insurance industry. The customer journey is a central point within this context. Data science allows insurers to gain a deep understanding of their customers through targeted analysis of customer data. As a result, insurers can better tailor the customer experience to each individual customer. This requires designing the customer journey accordingly and integrating data science into the process steps.
In the insurance market, this trend is already evident and providers are increasingly applying data science. Machine learning, for example, can help to settle claims within seconds, as demonstrated by Lemonade Insurance Company. [i] Pioneering companies also increasingly use voice recognition services in telephone sup- port and chatbots in online support.
When designing a customer journey, it is essential that data science represents the key element and is applied at those points where added value can be created. On the way to becoming a data-driven insurer, it does not suffice to just consider individual use cases within the customer journey, but rather you need to consider the entire customer journey from a top-down perspective.
This leads to two central questions that insurers need to address in this context:
- What should a data-driven customer journey be like to improve customer experience?
- How can data science be applied to improve internal processes?
Process design for all aspects of data science
Insurers have various options to improve the customer journey. Be it offering a tailor-made insurance product or contacting the customer for relevant products only—improvements in these areas only have a selective influence on the entire customer journey. The challenge, however, is to not only integrate data science into the individual steps, but also to ensure that these steps communicate seamlessly with each other. Intelligent robots, for example, can automatically assign insurance contracts to the clerks responsible. This is also known as intelligent process automation (IPA).
Integration must therefore take place both vertically and horizontally. Vertical integration can be defined as the application of data science within a single process step. One example is personalised premium calculation. Horizontal integration means applying data science at the interface of the individual process steps. This leads to the definition of a data-driven process: a process is data-driven if its design and functionality are determined by data science.
The following three steps are necessary to integrate this process into the current process landscape:
- Evaluation: potential processes are evaluated and a check is made as to whether the necessary prerequisites for a successful transformation have been
- Definition: the process is designed together with all parties
- Implementation: the necessary IT infrastructure is created, models are implemented and results are further
Successful implementation benefits not only the customer, but also the insurer. Compared to a non-data-driven process, the use of data science ties up fewer resources, and automation eliminates potential human error sources. Automation also enables decision-making primarily based on rational criteria, thereby eliminating emotional factors.
Using data science for process improvement
If the business is restructured, it is essential to identify bottlenecks in the current processes. In our previous article, we already mentioned that the amount of data collected doubles every two years. This data not only contains customer and business data, but also so-called metadata. Metadata also includes information on processing or lead times. This data can be evaluated through the targeted use of analysis software, which allows to monitor lead times, to identify influencing factors that slow down processes and to eliminate weak points.
It is essential to fulfil the necessary prerequisites in order to ensure successful transformation. These include the following areas:
- Staff / know-how
- Data / IT infrastructure
In the following sections, we will focus on the topics of staff / know- how and organization. The data / IT infrastructure topic will be addressed in a separate article in our series.
It is essential to build up the necessary knowledge in order to fully exploit the opportunities offered by data science and to use them productively. To this end, job positions must be created for experts with the necessary know-how in the fields of mathematics, statistics and computer science. These experts enable the creation of models and the use of the latest available technologies.
Current developments within the German labour market, however, indicate that over the coming years, the search for data science experts will prove to be quite difficult. Our extrapolation reveals a shortage of approximately 45,000 trained data scientists on the labour market in 2025. [ii]
In the course of the customer journey, the customer interacts actively and passively with various organisational units. Previously, employees from the individual departments often evaluated the use of data science within the customer journey independently. At present, we can observe a trend in several DAX-listed companies towards setting up data science departments that assume a central management role for data science projects. All competencies in the area of data science are bundled in these departments. Their experts provide advice to the individual departments and support during the implementation.
The use of data science leads to faster, more individual and smarter processes, from which both the customer and the insurer will benefit equally. Customers can thus be offered tailor-made services, and insurers benefit, among others, from cost-efficient processes. In our next article, we will discuss in more detail how data science can be used specifically in marketing and sales.
[ii] Sources for extrapolation: Bundesagentur für Arbeit (German Federal Employment Agency): Beschäftige nach Berufen (employees by occupation) (June 2017), Fachkräfteeng- passanalyse (skilled labor shortage analysis) (January 2018) and Destatis Statistisches Bundesamt (German Federal Statistical Office): Prüfungen an Hochschulen, Fachserie 11, Reihe 4.2 – 2016 (university exams, subject series 11, series 4.2 – 2016)