Simple changes to the installation procedure can provide significant returns.

The network and infrastructure are architected, the product is developed, load testing of the infrastructure is complete, the marketing campaign is underway and customers are signing up at a blistering pace. You want your triple-play installation to go smoothly and don’t want to send a technician to the house for installation issues that can be detected automatically. The ideal scenario is a self-installation by the customer. No one has to be home, there is no truck roll and costs are kept to a minimum.

Unfortunately, the technical competence of customers and the number of unsuccessful installations has kept the percentage of self-installations relatively small. Most carriers still send technicians for at least one triple-play service installation, and some are sending a technician for every new service turn-up. Truck roll costs vary from carrier to carrier but range up to $500. If a second truck must be rolled just to correct the initial installation, 10 months’ revenue from a $50 service is wasted.

For those carriers advanced enough to provide self-installation of a VoIP service, the installation is usually considered complete when the customer hears a dial tone and successfully makes their first phone call.

But what if the quality of calls made on the line is poor? It’s well within the capabilities of a service provider to go one step further and immediately establish call quality, allowing the provider to address any quality issues right away.

Imagine an installation scenario in which customers are instructed to dial a number within the provider’s own network. The phone rings, and a friendly welcome message is played to the customer, thanking them for signing up for the service and letting them know the installation is complete. From the customer’s perspective, this phone call is a pleasant introduction to the service.

From the provider’s perspective, the call serves a specific technical function. The provider can measure the voice and signaling quality of that initial phone call, scoring it and comparing it with thresholds established within its network.

If the call quality is acceptable, the quality report is automatically logged as a baseline for that customer, and no action is required. If the call quality is unacceptable, an automated alert is sent to the NOC to take action.

The purpose is to identify installations that are a problem and take action before a truck roll, or before the customer returns the gear within their “30-day trial” period. This increases customer satisfaction and ensures the stickiness of the service.

Measuring the quality of a customer’s service...
Measuring the quality of a customer’s service starts with a simple
phone call and results in correlated data that guarantees a
consistently high quality of experience.

For those installations where a technician is on site, the scenario can be used in the same way. The majority of carriers today only require a technician to check for a dial tone before calling the installation complete. A simple phone call from the customer’s home phone can validate the installation, provide the test call results to the technician, and give the technician confidence to move on to the next installation or stay to fix a problem. This simple test is done without any custom equipment, with limited technician training and for a minimal investment.

The signaling and media metrics used to determine an acceptable call are based on the standards in the industry:

  • Post-Dial Delay (PDD) – The time it takes, after you've dialed, for the phone you're calling to ring.
  • Post-Pickup Delay – The time that elapses between answering the call and receiving the first packet of media (conversation).
  • Call Setup Delay – The time from receiving the request for the call until the final acknowledgment from the caller that the call setup has been successfully completed.
  • Mean Opinion Score (MOS) – Voice quality based on the G.107 E-model.
  • Packet statistics – Packet loss, delay and jitter.

All of these statistics can be measured for every installation call, and the results are consistent across the user base.

Tens and hundreds of thousands of installations make for interesting data. Turning that data into useful information is unlocking a key to new business intelligence. Now, installers can be tracked for completion time between installations. Direct employee installations can be compared with contracted installations. Quality can be compared by time of day, region or any temporal dimension. Information can be presented to multiple levels of the organization and used to validate the success of the process.

Post installation, carriers need to identify problems proactively and give technicians tools that allow them to actively troubleshoot VoIP issues. Service providers need to know more than service availability; they need to understand when quality is degrading and address problems before the customer service calls start.

The same system that answers the calls from customers can be used to generate and measure calls when a problem is detected. The message to the customer is different, but the information about the quality of the signaling and media is the same. When a customer complains about poor voice quality, the customer service representative uses the automated system to generate a test call to the home. The customer answers and hears a message explaining that the call is being measured and that the issue is under investigation. The results can be compared with the baseline call and can be used as a starting point to further troubleshoot the issue.

In addition to these troubleshooting scenarios, 24x7x365 monitoring of VoIP service quality can be achieved through a number of standards that allow silent test calls to the home and reporting of quality from devices in the field:

  • RTCP XR (RFC 3611) – Defines how any network device can report media quality statistics over RTCP.
  • NCS Loopback (PacketCable standard) – Allows an EMTA to act as a mirror for test calls.
  • SIP Media Loopback (IETF draft) – Defines how an endpoint (ATA, IP phone) acts as a mirror.
  • End-of-call voice quality reports through various protocols, such as SIP and H.248.

The concept of IP Media Loopback testing is simple. A device in the network makes a test call to the endpoint at the customer premises and asks the receiver to circle back every packet it receives during the test. Performance metrics are shared between the source and the mirror (loopback device) using RTCP or RTCP XR, making the data available to the correlation and analysis engine. Test calls are silent, and flexible scheduling policies can be defined for loopback sessions that meet the needs of network operators and don’t interfere with customer calls. Since RTCP is already part of the CPE device, there is no need to have a technician sent to the customer’s site or to download software agents. Service providers can automate this process for proactive, periodic monitoring or can run on-demand tests as needed.

Simple changes to the installation procedure and minimal training can provide significant returns, especially in an economy when every dollar counts. Limiting the repeat truck roll through the final installation call is the first step to assuring the installation. The result is unleashing an entirely new set of data for analysis and taking a step toward the ultimate goal of increased customer satisfaction.