Bench Marks

[dllohio]

What are industry standards for call center abandonment rates and speed of answer.

We are a hospital call center and currently our standard 80% of calls answered in 30 sec or less, equaling an 80% service level.

Our abandonment rate is 10% or less.

It this in line with other call centers?

[Moe4525]

One approach you can take toward identifying the typical abandonment rate is to use an Erlang "C" calculator. The variables you will need for the 1/2 hour are; Volume, Handle time, # agents staffed, Service level goal. With this information and the Erlang calculations you can do some "What if" scenarions. Good Luck.

-Moe

[Michcall]

Industry standards is a misleading term. Metrics like service level, abandon rate, speed of answer etc. are functions of the type of call center it is, expectations of your calling population, and financial impact.

For example, I have one emergency response center with a 90/10 service level, but a companion support group that doesn't handle emergency calls is 80/20. For most of the clients/centers I deal with, 80/20 seems to be a fairly common practice. If it's hospital information or switchboard that you're dealing with, a lower service level is acceptable. If it's emergency room, then I would expect a higher service level.

Your abandon rate of 10% seems high based on what I'm used to----typically 5% or less.

Dr. Jon Anton of Purdue University does benchmark studies on call centers across industries. This might help you undersatnd what others in health care are doing. Here is his website: http://www.benchmarkportal.com/newsite/index.tml Dllohio's suggestion of using Erlang is a good one, although it doesn't tell you industry standards, it will help you understand how changes to service level will impact your staffing plan.

[JoeB]

I agree with Michcall – the frequently sought after “standards” are misleading and potentially risky. Each call center needs to determine the appropriate metrics for their business, service performance targets, and customer needs and expectations. Purdue University’s benchmark data is always interesting, but because it averages many data points of mostly unverified information, it has to be used prudently. See the following for discussions about abandon rates and benchmarking practices.

www.DiagnosticStrategies.com/benchmarking.htm
www.DiagnosticStrategies.com/abandonment.htm

As for the use of Erlang C to predict abandonment, you will need to understand the sensitivity of your callers to wait time – again, no “industry standard” here. Despite common belief and the (incorrect) use of average figures, our research has shown that this in an exponential function that is unique to each audience. Analysis of the ACD data and use of appropriate WFM can adequately predict abandonment for various staffing and service levels.

[n_meghanath]

I totally agree with Michcall, looking from all perspective, it will be expectation of the calling population and the kind of service that you are offering, However in any kind of service even if it is a low ARPU I would feel that the abandon should never go above 5%. About Service level and ASA you might like to consider which one to set more important since there is a mix in it. you may say you have answer 80% of the calls in 20 seconds but what about those 20% calls? 30 sec? 40sec? 2 min? Another example would be your ASA is 6 sec for a given time period of 30 min with 6 calls. now lets say you were able to answer all the 5 calls in 1 sec each but the last calls had to wait for 35 sec. the ASA would give 6 sec as the ASA whereas it doesn't hold even near to true for 5 calls neither for the 6th call.

Regards
Meghanath

[JoeB]

Meghanath is of course correct. This is but one of numerous examples of how incorrect use of averages yield incorrect results and possibly false sense of acceptable performance. To get a complete view of the individual customer experience, you need to know how many calls will be routed to the ACD queue and how long they will wait (average as well as maximum wait). These are the parameters that you need to forecast the number of abandoned calls.

Unfortunately, most industry benchmarks continue to report simple averages that vary by industry and hardly describe the individual caller experience.

Joe
www.DiagnosticStrategies.com