Breaking Through the Analytics Wall at DigIn 2018

May 21, 2018
Mark Rusch's picture
Mark
Rusch
VP, Insurance

Last week, I along with several of my colleagues attended the DigIn Conference, an exciting event that brought together insurers from around the country. While at the conference, we shared our newly released insurance video and also sponsored a session in which we gave a talk that focused on how insurance companies can “break through the analytics wall.” The topic of the talk originally came to us after realizing that many  insurance companies want their analytics experience to be like something out of a science fiction movie, but they’re currently struggling to see results with even basic analytics.

Clearly, this presentation resonated deeply with a large number of attendees. In addition to a very engaging Q&A session after the presentation, we had people coming up to us for the rest of the conference to share that they have the exact same problems that we outlined. Namely, these issues revolve around the fact that companies have taken the step to deploy some predictive models, but they’re static, offer no consistency, and take too long to provide any insight to the employee working with the customer. I heard from one attendee that their company actually turned off their fraud models because it was so out of date, difficult to use and they weren’t seeing accurate  results (too many false positives).

To address this problem, many attendees told me that they’ve started introducing one-off solutions, to varying degrees of success. The problem is that models just produce numbers, but you need some explanation to contextualize the data behind it in order to share the model output with the customer or agent. To progress and mature, these companies need a more modern infrastructure that embraces the full life cycle of model management, deployment, refresh, and monitoring.

With an advanced platform with embedded analytics, not only do companies get insights and recommendations to the end user at their point of work, but they’re then able to use the platform to handle some of the workload. Companies can begin testing the waters, automating decisions that have, for example, a confidence level of 92% or higher that the decision is correct so let the system complete the transaction. The ability to throttle your business process with these kind of analytic levers is a great future direction that is also within the realm of possibility.

Stay tuned for more on how you can break through the analytics wall. We had such a great response to our presentation, both during the Q&A and at our booth for the rest of the conference, that we’re  turning our presentation into a whitepaper. It’ll clearly lay out the problems that many insurance companies are facing when it comes to deploying analytics, and it sets a path for the way forward.