By Duck Creek Technologies and Belhe Analytics Advisory
A new standard is emerging for how the insurance industry leverages data to fuel and guide their businesses. Insurers are expected to be able to know their customers better than the customers themselves, find opportunities where competitors aren’t looking, and continuously learn and improve their operations. Yet we know that most insurers have not adopted this standard, and the most common question we hear is “where do we start?”
The answer often catches people by surprise. That is because they expect to hear about breaking down data siloes, improving data quality and governance, finding the right talent, or implementing sophisticated technology and tools. The assumption is that high-quality data and processes are needed to make the leap towards this new standard.
Make no mistake, all these are essential pieces to the puzzle, but they aren’t the starting point. The answer to our original question is that the data transformation journey starts with adopting the right attitude. The traits of the right attitude are:
Step 1. Accept – Consciously accepting your data imperfections
We all know that perfection should never be the enemy of the good, but most insurers struggle to realize that what they have is already a good state even if it has shortcomings. For example, your core systems will mostly have good pockets of claims data as well as policy data. Creating a combined view of your customers by combining claims, policy, and exposure data is a function of rightful desire and conscious attempt. We have witnessed that the carriers with “can do” attitudes have created “fuzzy pictures” of customer reality that are “clear enough” to act upon by connecting the dots within the “somewhat ok/good” datasets. Some progressively-thinking carriers are increasingly turning towards external data to supplement the pockets of “good enough” internal data with external data such as demographic, climate, or industry data. Such effort is providing them with insights that are not so commonly available.
Step 2. Ask – Repeatedly asking what the data is telling you
Insurance is a business driven by the experiential wisdom of executives who craft their strategies based on their years of domain knowledge. It has now become more of an industry norm to supplement decision making with data-driven insights. On one hand, the sophisticated pricing models along with predictive models for claims and underwriting purposes have had focused, positive impact on insurance carriers’ profitability, but on the other hand they have raised specifications and expectations for data quality as it relates to analytics applications.
The crying need of the current business environment is to ask relevant questions of the “pockets of good enough” data to gather faster insights. These questions could be based on simple correlations and observation of trends that may matter. For example, COVID-19 has increased the frequency of business interruption claims irrespective of whether they have that coverage or not. The propensity of litigations has also gone up considerably. So questions related to geographical distribution of the trend over the last two quarters of loss of business claims may have a meaningful impact on carriers’ action plans.
So you may want to emphasize the ability to ask questions as they relate to business operations and external drivers, and set up a framework to generate insights that quickly addresses these questions.
Step 3. Act – Diligently acting upon insights you have
Insights are worthwhile if you act upon them. The more relevant questions you ask of the data you have, the more likely your insights are to be actionable. The more you have the mindset of grabbing the opportunity to fix an issue based on newly acquired insights, the more likely you stay competitive in the marketplace. We are observing that new underwriting guidelines stem from conscious use of data – for example, one carrier is cautiously examining quotes with workers’ comp class codes that are demonstrating increasing trend in claim frequency.
Insights should be wisely used, keeping in mind that as we gather more credible data we always have the opportunity to revise our learning – but it may not be opportune to wait for the perfect data to come around to respond to market conditions.
In Conclusion:
Looking inwards first can be uncomfortable, but our experience shows that it’s critical that insurers don’t skip this first step. Progress starts with rethinking assumptions, or insurers risk falling into the same traps they are in today, resulting in most data sitting idle and most of data scientists’ time spent making the data right instead of putting it to use. The carriers that get in the habit of making data useful by accepting imperfections, asking questions of that data to quickly gain the required knowledge, and adopting a balanced mindset to act upon the derived insights, have what it takes to stay competitive in the market.
