The 5 Commandments of Data for Insurance
In October 2019, Carpe Data hosted its first inaugural insurance data conference, Data in Paradise. In his keynote session, CEO Max Drucker shared five central tenets that guide the company’s product and innovation strategy. You can watch the full session video here.
1. Everything we do must drive automation
This is the first and most important commandment. According to IBM, 90% of all data that exists today was created in the last two years alone, and it’s estimated that the total amount of worldwide data will reach 40 zettabytes (or 40 trillion gigabytes) – this year. Yet this vast collection of data has no meaning if we cannot analyze and extrapolate its patterns at scale.
If we are to contribute to the operational efficiency of claims and underwriting life cycles we must provide relevant data and mechanisms like artificial intelligence (AI) by which it can be interpreted.
2. All data must be consumable via system-to-system interface
For all that data and AI can do to foster efficiency within the insurance ecosystem, that data means nothing if it can’t be delivered effectively to its end user. From our perspective, all data should be system agnostic (or system interpreted) and allow for any current ecosystem to process and interpret results.
A natural precedent for this simple digestion of information is in the delivery method. A system-to-system integration— through direct API, as a top choice— not only guarantees swift, secure data delivery, but further streamlines the processes that depend on it.
3. All data must predict insurance outcomes and illuminate a deeper understanding of risk
Whether you’re searching for new data sources to refine segmentation, reduce fraud, or improve your overall customer experience, it’s not enough to simply source quality data and stop there— each data element should be offered with an intent in mind. The only way to know that is to test, analyze, and interpret until that information has a guaranteed insurance outcome.
Customer feedback is a crucial aspect of this data dance: what is predictive on its own may not be significant in a larger context, so whichever data provider you partner with should be enthusiastic to receive feedback and work with your team to provide what is truly valuable.
4. Data must be provided for a reason and should be managed to perform accordingly
Simply put, data that is not immediately relevant to our clients’ needs or to the greater insurance ecosystem should not be considered, and we remain vigilant against providing extraneous data for its own sake, or data that has not been proven to be significant (see Commandment Three). Furthermore, if data has been proven to predict an insurance outcome, it is our responsibility to maintain a standard of quality for that data.
5. The data we provide must be unique, and therefore additive to insurance
Creativity plays a crucial role in discovering new predictive elements and is one of the reasons we provide wholly unique perspectives to insurers. We’re constantly testing new theories against existing standards, which requires a framework of flexibility when investigating emerging sources. This framework guarantees that the excitement of the discovery does not overshadow our purpose to find new, unique and additive elements for insurers.
Of course, not every experiment ends in success. But 40 trillion gigabytes of data demand that we think differently about what is relevant to our clients, to their customers, and to insurance at large.
For more on the 5 Data Commandments, check out the recording of Max’s session at Data in Paradise 2019.