Jul 305 min read

InsurTech Ohio Spotlight with Matt Hite

Matt Hite is an Application Development Manager at Encova Insurance, a super-regional carrier providing business, auto and home insurance. Matt was interviewed by Andrew Daniels, Co-Founder and President at CrashBay and Founder and Managing Director at InsurTech Ohio.

Matt, can you define observability, and tell us how it differs from monitoring?

“Observability is the ability to infer the internal state of the system based upon the external outputs. It seeks to explain how the application is performing without needing to directly inspect the internal workings of the code. That said, the better observability platforms actually provide deeper insights into the internals of an application, which helps IT quickly fix problems and restore service.

Foundationally, observability uses three types of data to provide visibility: logs, metrics and traces. Logs are structured or unstructured data that the system outputs. Metrics are structured data such as CPU, memory or network utilization or custom data points that we want to track. Traces are the end-to-end journey. With a trace, I can follow a payment or quoting process from start to finish and see what happened along the way, i.e., what services were called or what database queries were executed. These traces, plus the data from logs and metrics get correlated to paint the complete picture. That’s observability.

Monitoring is a component of observability. I need to do monitoring to have observability. For example, I monitor system memory and CPU to add system health information into my observability view. All this monitoring creates a lot of data. The magic happens when technology can take this monitoring data to learn the norms, ignore the noise and pinpoint the problems.”

Why is observability important?

“We used to have systems that were standalone and monolithic, especially in the insurance industry. You could easily install a monitoring platform on them, and you would have the full view of the application. But, with the API revolution, things got difficult. Our new systems are highly distributed, a web of connectivity, and sometimes outside our control with the cloud. We may now have tens, maybe hundreds of systems communicating in a single process. If the payment system is running poorly, do I understand the dependencies and the enterprise impact? We need observability solutions to collect all that information and analyze it in one place. That’s why it's important.”

How do you use this in your business today?

“We use it to understand how our users interact with our system, and we match customer feedback we receive through our online feedback form with the observability data we're collecting. If we get bad feedback, we can compare that against the observability data and troubleshoot the issue. We can also look to see if other users are experiencing similar problems.

A really good use case of this is First Notice of Loss (FNOL). We used our observability tools to improve the success rate of reporting claims. We found that some users had a poor experience in our FNOL application. There were unclear questions and some buttons were causing the users to have to restart. By analyzing our observability data, we were able to improve our application and customer experience.”

How can this data be used to help improve business processes?

“There are multiple ways observability can help business processes, but first, we need to start talking about system health in terms of the service-level objectives and service-level indicators. Instead of just reporting the technical issues of the system, we need to measure system health by measuring the ability to complete business processes successfully and within performance expectations. We gain the support and prioritization to improve systems when both IT and the business can pinpoint the business processes that need to improve.

The second area is experimentation. When new features are released to our customers, we can use observability data to compare if the new features are improving the business process in terms of ease of doing business or performance. We can compare the business process metric data before the change versus after the change. If a feature is not performing as expected, we can roll back and try something different. By continuously monitoring and getting feedback quickly, we can have rapid development and iterative improvement of digital products and services.

Third is customer experience. Observability tools provide metrics that provide detailed and overall satisfaction scores for our applications. Error counts, high visually complete times or the number of rage clicks provide indications of poor user experience. Additionally, observability tools allow us to understand how our users are interacting with the systems which enables us to change layouts and system flows. By receiving customer experience insights and personalizing the experience, we aid the business and gain favor with our customers.”

What are some other use cases for this in the insurance industry, and how do you bring the business side on board to what you're doing?

“Customer relationship management (CRM) systems are all about knowing your customer. Beyond storing just policy, billing or claim information in your CRM system, we can populate it with observability data to elevate the customers’ experience and usage patterns to business users. This usability data can aid customer service calls or spur conversations at agency visits.

Observability can play a role in fraud detection by looking for anomalies and suspicious activities and correlating those interactions to gain information. Additionally, in the claims space, improving workflow processes and identifying bottlenecks can streamline operations and decrease the time to resolve claims.

We can bring the business on board with observability by sharing the power of observability. We should not keep these tools in an IT-focused space. Provide access to business users and show the business service level indicators. There are many business insights that observability can bring to light. Showing the dashboards, the real-user monitoring and the user experience will get your business interested in the data.”

What's next for technology innovation in this space?

“I have two things here. Observability platforms are now including application security monitoring in their toolbox. At the most basic level, they're doing static-code analysis or analysis of the running system to detect if you are using third-party libraries with known vulnerabilities.

Then you have real-time threat protection, which is the next level. Because these observability tools are in our running code, they can observe the code and the data flowing through it. Knowing how third-party libraries have been exploited in the past, they can detect and stop a similar attack in real-time. It's like putting active security guards around your home versus a passive security system. Hardening your applications by adding application security monitoring is a smart choice.

The second is AI. We're collecting a lot of data through observability. We talked about logs, traces, metrics and distributed systems that are all communicating at one time and creating this large pool of data. These tools are already using AI to detect anomalies, yet sometimes the data can be hard to find or visualize and share. I believe AI’s natural language processors will soon allow us to more easily gain insights from our observability data, generating dashboards and metrics in chatbot fashion. This can unleash even more business power by making the data more easily accessible.”