a new look at THE TOP 20 Contact Center Metrics

Nov 01, 2006 10:30 PM  By

The evolution of a simple call center into a multichannel contact center doesn’t happen overnight. You may need to add or upgrade technologies, and certainly staff skills will need to expand as customer contacts begin to include e-mail and Web chat in addition to incoming phone calls.

It’s also important to rethink what performance measurements are important for this new breed of operation. Are the measures of performance that served you well in the call center the same ones that will determine how well the multichannel contact center is working? You can organize contact center standards into three categories: service, quality, and efficiency. We’ve put together the top 20 metrics in these categories.

SERVICE MEASURES

The most important measures of performance in the contact center are those associated with service. Some of these measures are the same for both the old-fashioned call center and the modern-day contact center, while some need to change slightly to reflect the new types of transactions.

1) BLOCKAGE

An accessibility measure, blockage — busy signals — indicates what percentage of customers will not be able to access the center at a given time due to insufficient network facilities in place. Most centers measure blockage by time of day or by occurrences of “all trunks busy” situations. Failure to include a blockage goal allows a center to always meet its speed-of-answer goal simply by blocking the excess calls. As you can imagine, this damages customer accessibility and satisfaction, even though the contact center appears to be doing a great job of managing the queue.

The contact center must also carefully determine the amount of bandwidth and e-mail server capacity to ensure that large quantities of e-mails do not overload the system. Likewise, the number of lines supporting fax services must be sufficient.

2) ABANDON RATE

Call centers measure the number of abandons as well as the abandon rate, since both relate to retention and revenue. Keep in mind, however, that the abandon rate is not entirely under the call center’s control. While abandons are affected by the average wait time in queue (which the contact center can control), a multitude of other factors also influence this number, such as individual caller tolerance, time of day, availability of service alternatives, and so on.

Abandon rate is not typically a measure associated with e-mail communications, as e-mail does not abandon the “queue” once it has been sent, but it does apply to Web chat interactions.

3) SELF-SERVICE AVAILABILITY

More and more contacts are being offloaded from contact center agents to self-service alternatives. In the contact center, self-service usage is an important gauge of accessibility and is typically measured as an overall number, by self-service methodology and menu points, and by time of day or demographic group. In cases of Web chat, automated alternatives such as FAQs or use of help functions can reduce the requirement for the live interaction with a Web chat agent.

4) AND 5) SERVICE LEVEL AND AVERAGE SPEED OF ANSWER

Service level, the percentage of calls answered in a defined wait threshold, is the most common speed-of-answer measure in the call center. It is typically stated as X percent of calls handled in Y seconds or less. Average speed of answer (ASA) represents the average wait time of all calls in the period.

In the contact center, speed of answer for Web chat should also be measured and reported with a service level or an ASA number. Many centers measure initial response as well as the back-and-forth times, as having too many open Web chat sessions can slow the expected response time once an interaction has begun. The speed of answer for e-mail transactions, on the other hand, is defined as “response time” and may be depicted in terms of hours or even days, rather than in seconds or minutes of elapsed time.

6) LONGEST DELAY IN QUEUE

Another speed-of-answer measure is how long the oldest call in queue has been waiting: the longest delay in queue (LDQ). A number of centers use real-time LDQ to indicate when more staff need to be made immediately available.

Historical LDQ is a more common measure, to indicate the “worst case” experience of a customer over a period of time. Historical LDQ is measured in two categories. One is the longest delay for a customer whose transaction was finally handled by an agent (longest delay to answer), and the other is the longest delay for a customer who finally abandoned the contact (longest delay to abandon), as might be the case in a Web chat scenario.

QUALITY MEASURES

Perhaps a more significant indicator of customer satisfaction than the “how fast” measures outlined above is “how well” the contact was handled.

7) FIRST RESOLUTION RATE

The percentage of transactions completed within a single contact, often called the “one and done” ratio, is a crucial measure of quality. It gauges the ability of the center, as well as of an individual, to accomplish an interaction in a single step without requiring a transfer to another person or area and without needing another interaction at a future time to resolve the issue. The satisfactory resolution of a call is tracked overall in the center, as well as by type of call and perhaps by time of day, by team, or by individual.

You should likewise track the one-contact resolution rate for e-mail transactions and Web interactions. The resolution rate will likely be lower for e-mails, as it generally takes multiple messages between two parties to resolve a matter to completion.

8) TRANSFER RATE

The transfer percentage is an indication of what percentage of contacts have to be transferred to another person or place for handling. Tracking transfers can help fine-tune the routing strategies as well as identify performance gaps of the staff. Likewise, tracking e-mails that must be transferred to others or text chat interactions that require outside assistance helps to identify personnel training issues or holes in online support tools.

9) COMMUNICATIONS ETIQUETTE

One of the critical factors that affect the caller’s perception of how well a call was handled is simple courtesy. You can monitor the degree to which telephone communications skills and etiquette are displayed via observation or some form of quality monitoring.

E-mail and Web chat etiquette should also be observed. Standard wordings that employees should follow in both types of communications should be carefully observed, reviewed, and recorded.

10) ADHERENCE TO PROCEDURES

Adherence to procedures such as workflow processes and call scripts is particularly important so that the customer receives a consistent interaction regardless of the contact channel or individual agent involved. In the call center, adherence to processes and procedures is typically measured for individuals through simple observation and the quality monitoring process. Adherence to processes and procedures such as written scripts and preapproved responses is also important for e-mail and other channels of contact.

EFFICIENCY MEASURES

Executives in every type of organization are concerned with how well its resources are being put to use. That is especially true in the contact center, where more than two-thirds of operating expenses are related to personnel costs.

11) AGENT OCCUPANCY

Agent occupancy is the measure of actual time an agent is busy on customer contacts compared with available or idle time, calculated by dividing workload hours by staff hours. Occupancy is an important measure of how well the call center has scheduled its staff and how efficiently it is using its resources. If occupancy is too low, agents are sitting around idle with not enough to do. If occupancy is too high, agents may be overworked.

Agent occupancy rates often reflect the randomness and unpredictability of incoming calls. In those instances, the desired level of occupancy may lead managers to pull agents away from processing e-mails to answering phones, or vice versa. Because Web chat interactions are essentially random events like incoming calls, the same measures of occupancy apply here as in an incoming call scenario.

12) STAFF SHRINKAGE

Staff shrinkage is the percentage of time that employees are not available to handle calls. It consists of meeting and training time, breaks, paid time off, off-phone work, and general unexplained time where agents are away from their stations. Staff shrinkage is an important number to track, since it plays an important role in how many people will need to be scheduled each half-hour. The same measures of shrinkage that are used for call center calculations also apply to the multichannel contact center.

13) SCHEDULE EFFICIENCY

Workforce management is all about getting the “just right” number of people in place each period of the day to handle customer contacts. Schedule efficiency measures the degree of overstaffing and understaffing that result from the scheduling design.

Measure schedule efficiency for responding to the randomly arriving Web chats just as you measure it for responding to incoming calls. Since e-mails typically represent sequential rather than random workload, the work fits the schedule, and therefore overstaffing and understaffing measures are less relevant.

14) SCHEDULE ADHERENCE

Schedule adherence measures the degree to which the specific hours scheduled are actually worked by the agents. It is an overall call center measure and is also one of the most important team and individual measures of performance, since it has such great impact on productivity and service. Schedule adherence is a critical measure in the multichannel contact center as well. Specific hours worked is less of an issue in a group responding to e-mails rather than real-time demand of calls and Web chats, but it is still relevant in processing the work in a timely manner, especially if response-time guarantees exist.

15) and 16) AVERAGE HANDLE TIME AND AFTER-CALL WORK

A common measure of contact handling is the average handle time (AHT), made up of talk time plus after-call work (ACW). To accommodate differences in calling patterns, you should measure and identify it by time of day as well as by day of week. AHT is also important regarding the other types of multichannel contact workload. It’s harder to calculate, however, given the difficulties of measuring how long it takes to handle an e-mail or a Web chat transaction. An e-mail may be opened and put aside for varying amounts of time before it is completed. Likewise, a Web chat session may appear to take longer than a phone call, since a Web agent typically has several sessions open at once.

17) SYSTEM AVAILABILITY

Slow response time from the computer system can add seconds or minutes to the handle time of a transaction. In the call center, system speed, uptime, and overall availability should be measured on an ongoing basis to ensure maximum response time and efficiency as well as service to callers. For example, if the interactive voice response (IVR) typically handles 50% of calls to completion but is out of service, more calls will require agent assistance than normal, causing overtime costs, long delays, and generally poor service. Often this will be a measure of performance that resides in the IT department, but it is also a crucial measure of contact center performance.

18) CONVERSION RATE

The conversion rate refers to the percentage of transactions in which a sales opportunity is translated into an actual sale. It can be measured as an absolute number of sales or as a percentage of calls that result in a sale. You should track and measure conversion rates for incoming calls as well as outgoing calls, e-mail transactions, and other Web interactions.

19) UPSELL/CROSS-SELL RATE

Many companies measure the upsell or cross-sell rate as a success rate at generating revenue over and above the original order or intention of the call. It is becoming a more common practice, not just for pure revenue-generating contact centers but for customer service centers as well. Although more prevalent regardng telephone calls, it is also an appropriate measure of performance for other communications channels.

20) COST PER CALL

A common measure of operational efficiency is cost per call or cost per minute to handle the call workload, both in a simple call center and in a multichannel contact environment. This cost per call can simply be a labor cost per call, or it can be a fully loaded rate that includes wage rates in addition to telecommunications, facilities, and other costs. In setting cost per call, it is critical to define the components being used and to use them consistently in evaluating how well the center is using financial resources over time. This metric is commonly used to compare one company or site to another in benchmarking, but that’s not a good practice, as the components included and the types of contacts will often vary.


Penny Reynolds is a founding partner of Nashville, TN-based the Call Center School, a contact center consulting and education company.