There is no one-size-fits-all solution for insurance business intelligence needs. But, there’s a four-part approach and set of technologies that can help.
In your role as an analyst for the Commercial Property line of a Property & Casualty Insurance company, your line manager has come to you for insight on what the loss ratio looks like if policies that meet certain criteria are discounted.
Your analytical mind is thrilled at the prospect of a new investigation. However, you understand very well what is about to happen next.
The Insurance Business Intelligence (BI) Map
You will schedule a meeting with IT to request the data you need. That meeting will end with IT promising to see what they can do to provide the requested data.
It will take more time and involve more complexity than anticipated to get the data. The data set will take several iterations before you have what you need to answer the question.
You will likely become “shadow IT,” collecting data from sources you have outside IT. You will spend late nights and early mornings completing the pivot table and charts to meet your deadline.
Once that’s done, the data and methods will live in one of your spreadmarts, which you then continually tweak and update, often long after the need for that data set has passed.
Perhaps your colleagues hear you have done this and come to you to assist them with a similar task for their line, which results in more iterations with IT and a fork of the spreadmart.
For many insurance companies, this process plays out over and over with each new analytical question: Analysts spend most of their time collecting and curating data sets and little time doing the actual analysis at the core of their job description.
Meanwhile, dozens of spreadmarts with conflicting versions of reality are brought into competition in high-stakes C-suite meetings.
IT vs. The Business
The source of much of the tension when it comes to analytics in the modern enterprise is the relationship between IT and “the business.” The dilemma:
IT is often dealing with multiple back-end systems on varying versions of varying database platforms and must ensure security and integrity while extracting that data and providing it in a common format to “the business side.”
They often feel the business side has unrealistic expectations of what can be achieved given the limitations IT faces.
On the other hand, the business often feels that IT wants to be tail-wagging the dog: setting the rules for what the business can do rather than supplying the IT needs.
Experience with data requests has many practitioners reflexively asking for ALL DATA for ALL TIME. The mindset behind that: “If I don’t ask for everything now, it will cost a lot more later.”
These tensions are exacerbated by the clash of legacy systems with modern analytics, such that every conference everyone attends has at least one presentation of what someone accomplished with a modern BI tool and a seemingly pristine data set that your company can’t possibly hope to attain.
BI of the Past
There are three overarching principles we must balance when implementing a modern architecture to serve the analytical needs of the enterprise:
- All data has some value.
- Not all data has equal value.
- Data collected is not automatically available for analysis. (i.e. Just because you have data in a source system does not mean analysts can access it.)
Until recently, business intelligence often served the lowest common denominator. That means the data available for analysis was hyper-governed and tightly controlled. Only data determined to have a high value was available for analysis. Adding new fields to an existing analysis took exorbitant time and effort.
There is good news: you are not alone. Even better news: there is a solution.
The Insurance BI Modern Analytics Platform (MAP)
The cold, hard fact is that there is no one-size-fits-all solution for modern insurance BI. There is, however, a four-part approach and an evolving set of technologies that can provide a golden path to a BI solution that serves the entire enterprise well.
1. The data must be available.
In modern enterprises, data can be captured in any number of systems for any number of purposes. The modern architecture must be flexible enough to capture data from any conceivable source, even those that do not currently exist. The sufficiently modern BI architecture will support the efficient capture of all data from all source systems so that it is available for analysis when needed. This could mean that all data is captured at the source, or that all data is efficiently capturable at the source as needed. The Modern Analytics Platform supports either use case.
2. The data must be accessible.
Having all that data is the first hurdle, but you still need to ensure that the data is easily accessible for everyone who needs it, using whatever tools they will use now or in the future.
3. You must have a semantic layer where all your data governance decisions are made concrete and objects, their attributes and metrics are defined in clear business terms.
This approach will ensure ease of use for full-time analysts as well as other business users that need to access data for ad-hoc analysis.
4. You need a BI toolset that allows users to capitalize on the features above to more quickly create and share meaningful analytics.
The BI tool should be easy to use so that analysts spend less time learning the tool and more time exploring data. The insurance BI tool should also make it easy to share visualizations to the appropriate audience, so that insights are quickly and easily communicated in a secure fashion, regardless of the size of the data or the expertise of the user. A corollary to the above is that the BI tool should also make it easy to communicate analytics success to the organization. That blockbuster dashboard you’ve created should be easily shareable across the organization, as should the data set that you used, so that others can build upon the charts you’ve already created.
In conclusion, the Modern Analytics Platform is not a single prescription fit for all insurance companies.
Rather, it is a set of experience-backed approaches and modern tools that allow for flexibility to meet the needs of a wide range of modern enterprises.