Five of Today’s Most Misunderstood Terms
1 Sep, 2003
By: Brad ClevelandLike most professions, the call center industry has an abundance of unique terms and acronyms. But the call center world seems to have done an especially notable job of confusing things –what can you expect from an industry that can''t even decide on whether it''s... well, a call center (a.k.a., contact center, interaction center, help desk, et al.)? Here are five of today’s most misunderstood terms:
First-Call Resolution
First call resolution is the percentage of calls that do not require any further contacts to address the customer’s reason for calling; the customer does not need to contact the call center again to seek resolution, nor does anyone within the organization need to follow up. Sounds simple enough.
However, there is wide variation among call centers on how first-call resolution is actually calculated. The bugaboo is that the many possible definitions of “resolved” and “total calls” impact results significantly.
Some of the definitions of “resolved” include:
• Caller states, upon being asked, that his/her reason
for calling was resolved
• Agent has no follow-up work to do as a result of the
call
• Agent does not need to transfer the call
• Agent resolves all of the caller’s concerns that fall
within the call center’s defined responsibility
• One of the call tracking codes designated to count
as “resolved” is associated with the call
There are also differences in methods of measuring total calls, including calls answered, calls answered plus calls abandoned, calls offered, calls answered that meet certain criteria (e.g., omit wrong numbers, calls with invalid data from a call tracking system, calls handled entirely by IVR or calls that the call center is not authorized to resolve, etc.).
Ideally, first-call resolution should be defined from the customer’s perspective as an issue resolved on first contact (the caller doesn’t have to contact the center again or vice versa), even if escalated or transferred during the contact. Given the wide variation in use, this measure’s greatest value is as a relative measure over time.
Erlang
The challenge here is the sheer number of terms based on “erlang.” Some folks refer to “Erlang” when they really mean “Erlang B” or “Erlang C.” Others cite generic “Erlang tables” (there’s no such thing). And some simply wonder, “What the heck is an erlang?” Here’s the scoop:
• A.K. Erlang was a Danish engineer who worked for the Copenhagen Telephone Company. In 1917, he published the formulas, Erlang B and Erlang C, widely used in the call center and telecom fields today.
• Erlang B is a formula widely used to determine the number of trunks required to handle a known trunk load during a one-hour period. The formula assumes that if callers get busy signals, they go away forever, never to retry (“lost calls cleared”).
• Erlang C is a formula used to calculate predicted waiting times (delay) based on three things: the number of servers (agents); the number of people waiting to be served (callers); and the average amount of time it takes to serve each person. Erlang C is widely used in workforce management software programs, as well as low-cost call center staffing calculators.
• An erlang is one hour of telephone traffic in an hour of time. For example, if circuits carry 120 minutes of traffic in an hour, that’s two Erlangs. (Historical note: after A.K. Erlang’s death, a standards committee decided to commemorate his work by referring to an hour – a common unit of measurement in telecommunications engineering – as an “erlang.”
So, there’s A.K. Erlang (a person), Erlang B (a formula), Erlang C (a formula) and an erlang (a unit of measurement). There’s no such thing as “the erlang formula” (be more specific) and no such thing as “an erlang table.”
Cost Per Call
There are various ways to calculate cost per call (i.e., determining which factors to include in staff costs, how to allocate equipment, how to value the building), but the basic formula is to divide total costs by total calls received for a given period of time, usually a month. The problem with cost per call is in interpretation.
A climbing cost per call is often a good sign. For example, process improvements may result in fewer calls than would otherwise be necessary (e.g., eliminating the need for customer callbacks, improving the IVR and coordinating with other departments to eliminate problems that generate calls). As a result, the fixed costs (in the numerator) get spread over fewer calls (in the denominator), driving up cost per call. Similarly, a dropping cost per call often indicates problems in forecasts or processes.
Average Speed of Answer
Average speed of answer (ASA) reflects the amount of time callers spend in queue, waiting to reach agents. It''s available from virtually any ACD. It''s widely used and reported. And – it''s highly misleading.
ASA is mathematically sound, no problem there. The problem is in interpretation. Let’s say you expect 250 calls in a given half hour, and anticipate an average handling time of 3.5 minutes. If you want to achieve an ASA of between 10 and 15 seconds, you’ll need 34 agents (according to Erlang C, 34 agents will produce an ASA of 12.7 seconds).
But what really happens to individual callers? The following illustration is based on QueueView: A Staffing Calculator, developed by ICMI; similar low-cost Erlang C calculators are available from ACD vendors and software companies:
Number of Calls Still Waiting This Many Seconds:
Immediate
Answer <5 10 15 20 30 40 50 60 90 120 180 240
Calls: 185 65 58 52 46 37 29 23 28 9 5 1 0
Scenario: 250 calls in half hour, 3.5 minute average handling time, 34 agents, ASA = 12.7 seconds.
Notice what happens. Sixty-five callers will wait five seconds or longer. In the next five seconds, seven of those callers reach agents, so 58 callers are still waiting 10 seconds or longer. In the next five seconds, six more callers will reach agents, leaving 52 callers waiting 15 seconds or more. And so forth. There''s still a caller waiting at three minutes.
The call center reports an ASA of 12.7 seconds, and everything''s dandy – other than for an obnoxious caller or two who say they waited "several minutes" to get through. Com''on, ASA is less than 15 seconds! Surely no one waited over 30 seconds, tops! Take a deeper look, though, and the real story becomes clearer. In short, ASA does not reflect a “typical experience.”
Occupancy
Also referred to as agent utilization or percent utilization, occupancy is the percentage of time agents handle calls versus wait for calls to arrive; the inverse of occupancy is idle time. For a half-hour, the typical calculation is: (Call volume x average handling time in seconds) / (number of agents x 1,800 seconds).
The terms adherence to schedule and occupancy are often incorrectly used interchangeably. They not only mean different things, they move in opposite directions. When adherence to schedule improves (goes up), occupancy goes down. Further, adherence to schedule is within the control of individuals, whereas occupancy is determined by the laws of nature, which are outside of an individual’s control.
Conclusion
Learning and using existing and new terms correctly is an important and ongoing part of professional development. There are some great rewards when you equip yourself and your management team to use terms correctly – you will communicate more effectively, make better decisions and produce better results.
