Limited visibility into user experience hinders efforts to improve digital health platform

Cambia Health Solutions is a not-for-profit healthcare company dedicated to making the healthcare experience simpler, better, and more affordable for the nearly 4 million people it serves through its regional health plans.

Cambia’s health plan websites are used by millions of members to perform critical tasks such as finding a doctor, managing medications, accessing mental health resources, and reviewing benefit information. The member experience team strives to provide simple, intuitive experiences through highly available, responsive applications.

However, the team faced challenges in understanding how users were experiencing the site. They lacked appropriate frontend insights to effectively understand user journeys and improve mean time to resolution (MTTR) when members reported incidents, such as login, account recovery, and other authentication issues. Without comprehensive data, identifying the root cause and troubleshooting these problems was time-consuming and difficult.

They also had limited insight into website performance, including key metrics such as Core Web Vitals (CWV) that measure page load speed, interactivity, and visual stability. Without this visibility, the team couldn’t identify when and where users were experiencing frustration with site performance. Cambia wanted to improve overall platform performance and establish meaningful service level objectives for availability and platform service reliability. However, its existing tools didn’t provide the end-to-end visibility needed to measure, monitor, and improve these metrics effectively.

Cambia searched for a solution that could provide deep insight into the user experience, help quickly identify and resolve issues impacting members, and enable engineers to proactively improve platform performance and availability.

Adding Real User Monitoring to complete the observability picture

Cambia has been a Datadog customer since 2018, initially using APM traces and service logging to gain insight into its infrastructure. “When I joined Cambia, we were already using Datadog,” says Kevin Bigler, Senior Software Engineer. “We were still feeling things out, but the impact of Datadog was spreading throughout our organization.”

As the team matured its observability practice, it realized it was missing a critical piece: frontend insights into the actual user experience. About three years ago, the team introduced Datadog Real User Monitoring (RUM). Initially, RUM was used primarily for performance insights. However, the team quickly discovered its value extended far beyond simple metrics. “Before we were logging around key user flows like password and username recovery, but that only shows half the story,” explains Bigler. “There’s a big blind spot when you’re only looking at the server-side metrics. While that telemetry is incredibly invaluable, it’s incomplete. We didn’t really discover that until we implemented RUM.”

Adding RUM introduced a crucial human element to Cambia’s observability plan, capturing real user experiences through CWV and other performance signals. “It gives a picture of what someone is going through as well as correlates with associated under-the-hood telemetry,” says Bigler.

“The information you get feels more like a story and gives you a step-by-step understanding of how a user progresses through our sites.”

The value of RUM became immediately apparent when the team was troubleshooting authentication issues, a critical pain point since sign-in is the gateway to all other website functionality. When members reported sign-in problems, engineers would review logs only to find a contradiction: users reported failures, but backend logs indicated success.

During one on-call rotation, Bigler encountered this scenario and, after exhausting traditional troubleshooting options, turned to RUM. He discovered a React frontend error, preventing the rendering altogether, that backend logs couldn’t capture—a clear signal demonstrating the need for full end-to-end visibility.

“That type of error was really inconsistent, so we struggled to reproduce it, but because we had the RUM data, we had a lot of information that we wouldn't have otherwise for this really tricky issue,” says Bigler.

Today, when incidents arise, Bigler’s team looks at RUM first to see the complete user story before drilling down as necessary.

“RUM, in the context of the Datadog stack, is super helpful and insightful for our team day-to-day,” he notes. “Combining RUM with other types of metrics has been a great value-add.”

End-to-end monitoring delivers faster incident resolution and improves confidence

Cambia now has complete end-to-end visibility into its stack. By combining APM tracing with RUM, the team connects frontend and backend monitoring to get full application performance visibility. This approach helped them identify and fix a long-standing authentication issue that backend monitoring alone couldn’t detect.

RUM gives the team the user context they need to quickly debug and resolve frontend issues, resulting in an approximately 20% decrease in authentication MTTR. “This gives us confidence in understanding members’ experience, which is critical given one of our main values is empathy for our users,” says Bigler.

Ultimately, using RUM means less guesswork. Instead of speculating based on incomplete backend data, the team knows exactly what members experience when using the platform. Engineers can see the complete user story—from initial page load through authentication and navigation—and quickly identify where problems occur.

Looking ahead, Bigler is excited about combining Bits AI with RUM data. “Building the whole story even with all the information can be difficult,” he explains. “Bits AI can help with the heavy lifting and correlation, especially with day-to-day operations and incident response. RUM data feeding Bits AI will give us clearer insights.”

Bigler also appreciates the partnership between Datadog and Cambia Health Solutions. “I feel like our voice is heard, and we have that direct communication with Datadog to get things that we need or even just share ideas. There’s a give and take and a close relationship there that I’m happy about,” he adds.