Establish metrics

Choose your metrics based on the objectives you identified. Below are some metrics to consider based on different goals:

Goal: Improve contact center operations

• Call volume by day
• Hours of operation
• Average handle time / wait time
• First call resolution
• Interactive Voice Response (IVR) resolution rate
• Reason for the calls (for example, new claim filing, appointment scheduling, claim status inquiries, etc.)
• Additional issues spotted during the call

Goal: Improve claim throughput

• First payment promptness
• Nonmonetary determination time lapse
• Average time for claimants to respond to requests
• Rate of continued claims with no issues

Goal: Improve user satisfaction

• Claimant feedback scores
• Expand claimant self-service options

In an ideal state, your tools will be configured to surface the data needed to evaluate your program’s health. Your benefits system will be able to provide data about how long claims are pending adjudication, and your call center software will track volume and other success metrics. In practice, existing tools, especially legacy tools and mainframes, are not always configured to get at data and metrics easily.

Suppose your system does not have the metrics you need readily available. In that case, it’s worth considering adding some scope to your claims status project to access relevant data. Evaluating what kind of effort it would take to surface the needed metrics would also be valuable.

Here are a few questions to consider:

• Is the data in question being tracked and recorded (if permitted under state law)?
  • If, for example, the reason for inbound calls is not being recorded, the first step would be to set up a process and supporting tools for contact center representatives to log the reason for the call.
• Can data be extracted manually for analysis?
  • For example, if the data in question resides in a relational database, you may be able to pull it with ad-hoc SQL queries. This is not an ideal long-term process for tracking metrics, but it can be a good option to support data extraction.
• Is there an existing business intelligence tool and/or data lake that a particular metric just needs to be added to?
  • If you already have some reporting architecture, adding new fields may be low-hanging fruit that you could include in your project scope.

Once you have access to data, the next step is establishing baselines. A baseline is just a snapshot of a metric before the project starts. For example, if your goal is to reduce call wait times, and the current wait time is 15 minutes, that is the baseline you will want to improve.

These baselines can help to refine the project goals by framing the problem and allowing realistic goals. They will also allow you to measure the project’s success and have a story to tell at the end.

Continue to collect data and measure metrics throughout the duration of the project. If you do incremental releases, continual measurements will help you evaluate the success of each iteration and pivot as needed. Ongoing measurements will also help you identify fluctuations in metrics like seasonal patterns and factor those during the final analysis.

If your data collection or extraction processes are manual, determine a frequency to pull the data (at least once a week is preferable).

If your data is available to a business intelligence tool, make sure you have dashboards set up to show the metrics in question as both snapshots and over time.

Once the project or a project milestone is complete, compare your baseline metrics to the current metrics. Make sure to account for any extraneous factors like seasonal patterns.

If you can measure an improvement of the objective metrics, congratulations! You now have a great story to tell stakeholders and constituents about how the program has improved.

If the metrics do not appear to be impacted by the project rollout, that’s also a valuable finding. Not every project will be successful. Data is essential because it allows an analysis of whether objectives were met and why or why not. Learning from implementation work in this way will help your next project to be more successful.