callstats.io gives network administrators detailed visibility into key Amazon Connect call quality metrics, complementing Amazon’s native contact center performance reporting tools.
Amazon Connect can help you avoid equipment expenses and simplify operations. But when you migrate your contact center to the cloud you lose visibility over your communications infrastructure—and more importantly the user experience. While Amazon Connect provides tools and APIs for tracking contact center performance (queue depths, hold times, handle times, abandonment rates, etc.), it provides no mechanisms for monitoring call quality or diagnosing real-time communications issues that can impair customer satisfaction and impact business results--as callstats.io engineers confirmed during our recent Amazon Connect hackathon.
To detect and diagnose call quality problems, Amazon suggests administrators establish a set of manual call problem reporting and troubleshooting procedures. To help agents validate their network connections, it offers a separate connectivity verification tool.
In contrast, callstats.io offers a completely integrated monitoring and troubleshooting solution that incorporates passive monitoring of vital call quality metrics and active network performance testing. callstats.io Monitoring and Analytics for Amazon Connect embeds advanced monitoring and connectivity testing capabilities directly into Amazon Connect agent endpoints. It displays the data in a central dashboard that helps administrators eliminate the guesswork and uncertainty of delivering high-quality user experiences.
Amazon Connect Metrics Gap: Network and Communications Signaling Data
Amazon Connect gathers a wealth of statistics to help administrators evaluate agent and contact center productivity. Administrators can use the Amazon Connect Console, CloudWatch, getMetric/getCurrentMetrics APIs and Kinesis Data Streams to retrieve, view and analyze a wide variety of contact center operational metrics.
But Amazon Connect does not collect or report network statistics like jitter, round-trip time (RTT) or packet loss that contact center administrators normally use to monitor and diagnose voice quality issues. It offers no visibility into the STUN/TURN/ICE protocols used to set-up WebRTC communications sessions between agents and its cloud infrastructure.
Instead, Amazon recommends agents manually report call quality issues through the Amazon Connect Contact Control Panel (CCP), by noting the time of the incident and a corresponding disposition code (e.g. L for latency, C for choppy, W for one-way audio) as described in the Amazon Connect Administrator Guide Troubleshooting and Best Practices.
Agent notations are accessible via the Amazon Connect Streams API and can be written to a database like Amazon Dynamo DB for external reporting and analysis. However, Amazon Connect does not provide any live dashboards or pre-configured reports to help administrators easily analyze the data or correlate problem reports with call records. Complicating matters, the raw data is based purely on agent observations, which are highly subjective.
To help troubleshoot problem reports, Amazon recommends collecting packet loss data for the affected network segments from external sources (routers and network management systems) and correlating this data with the agent reports by timestamp. In addition, it suggests measuring round trip time by conducting a series of experiments that include making recordings of a “loud noise” on the sent and received audio streams and measuring the difference using 3rd party audio analysis tools (see Testing Inbound Calls Using a Soft Phone in Troubleshooting and Best Practices). Executing these procedures and correlating the data can be a time-consuming, manual process.
Amazon Connect Check Connectivity Tool Provides Go/No-Go Verification
Amazon offers a basic web tool to validate an agent’s operating environment and AWS connection. An agent can run the Amazon Connect Check Connectivity Tool to verify their computer is WebRTC-capable and to check the latency to various AWS regions. (The agent must know the region where their Amazon Connect instance is deployed.)
The Check Connectivity Tool does not provide detailed network performance metrics like packet loss, RTT or throughput to help identify or isolate call quality issues. The test results are visible only to the end-user; there is no way for administrators to collect and analyze the data.
Figure: Amazon Connect Check Connectivity Tool
callstats.io Provides Detailed Visibility into Network Performance and Call Quality
callstats.io recognizes there is a direct correlation between voice quality and customer satisfaction. Our products allow Amazon Connect administrators to evaluate call quality and troubleshoot problems using real network performance data, eliminating speculation and uncertainty. callstats.io Monitoring and Analytics for Amazon Connect embeds advanced callstats.io monitoring capabilities into Amazon Connect agent endpoints, giving administrators real-time visibility into WebRTC sessions. (Amazon Connect uses WebRTC for the agent leg of customer calls).
The callstats.io integration complements Amazon’s native contact center performance reporting capabilities, helping network administrators and engineers easily identify, isolate and resolve a variety of Amazon Connect connectivity and audio quality issues including network performance constraints, infrastructure (STUN/TURN/ICE) issues, endpoint configuration errors, and signaling and media protocol incompatibilities.
callstats.io passively monitors WebRTC sessions, reliably collecting hundreds of detailed data points throughout a call, including critical network performance statistics like jitter, RTT and packet loss. Our unique Objective Quality metric helps administrators accurately assess WebRTC user experiences. And AI-driven algorithms automatically identify root causes, helping administrators save time and effort, and avoid conjecture.
The callstats.io product features an intuitive user interface and concise dashboard that make it easy to identify trends and isolate problems. Administrators can monitor inbound and outbound calls and supervisory sessions at a glance, and easily drill down on issues and analyze sessions using built-in event diagrams, graphs and strip charts.
Figure: callstats.io Monitoring and Analytics for Amazon Connect, dashboard screen
Agents can use callstats.io Amazon Connect Dialer Extensions to check their network connection at any time. A simple bandwidth meter lets agents assess anticipated call quality before engaging customers. In contrast to the Amazon Connect Check Connectivity Tool, the callstats.io dialer extensions display detailed statistics like RTT, packet loss and throughput. The callstats.io tool is particularly useful for home-based agents who often rely on unpredictable ISP connections. (callstats.io periodically runs the tool in the background between calls and reports results to the callstats.io dashboard)
Figure: callstats.io Amazon Connect Dialer Extensions
callstats.io’s advanced network performance monitoring and analytics solution can help you improve user experiences, boost customer satisfaction and make the most of your Amazon Connect investment. To learn more, visit our Amazon Connect Integration page, read how a large Amazon Connect customer used callstats.io to isolate an Amazon Connect performance issue, or contact us today.