ASA vs. Service Level (Your feedback requested)

[CASR]

Our current measurement of call response time is measured by a service level of 90% of calls answered in 20 seconds. We are replacing our current ACD system and we will now be able to get at the ASA, before we could only get it the next day. We are in the financial industry.

I was hoping people could tell me:
1. What you do measure Svc Level or ASA?
2. Why?
3. What is your goal?
4. What is your industry?

Thanks much,

Cathy

[kala]

We are a utility company. We measure both Service Level and ASA. Service Level is looked at more closely because it is reviewed by our PSC. You asked the question why do we look at SL. To answer that question 1) our PSC looks at it 2) we feel like it is a good measure to indicate what kind of service we are providing to our customers. Our goal is 85% of calls answered within 20 seconds.

[gkbansal]

We measure both Service level and ASA.

1) Service levels gives us a percentage of how many calls we handled out of the total calls with in a set threshold of time, like in our case it is 30 seconds.

Although we do measure ASA, personally I feel it is redundent because service level is a better indicator of how you service your customers, where as Average speed of Answer can be deciving, even a too low ASA say 20 seconds does not tell how many customers you handled below 20 seconds and how many above 20 seconds. For example, assuming you handled 10 calls, 5 calls at 10 seconds and 5 at 30 seconds, your ASA in this case would be 20 seconds, but if your threshold is 20 seconds for service level, your service level is only 50%.

2) Our goal is 80% in 30 seconds.

3)We are a major Newspaper circulation customer service.

[JoeB]

As pointed above, SL gives you the upper limit for wait time by the majority of the callers. If you meet your SL target of 90/20, you know that 90% of the callers wait no more than 20 seconds for a response. However, this doesnÂ’t tell you how long the remaining 10% wait.

ASA gives a better idea about the experience of the entire population instead of the select 80% or 90%.. However, ASA alone doesn’t give a complete answer either, because even with a “good” average, some callers may still experience unacceptable wait time.

What many ignore is that neither metric gives you a good enough idea about the number of abandoned calls, which for some businesses are a key business metric. For this reason, ACDs use a combined metric that includes SL and abandonment. See http://www.diagnosticstrategies.com/service_level.htm for a detailed explanation.

Therefore, a good ACD report will include SL adherence and abandonment, as well as averages and longest wait for the SL group and the entire population. Incidentally, the EasyErlang resource calculator (http://www.diagnosticstrategies.com/EasyErlang.htm) forecast all these parameters: SL, average wait for SL calls, ASA, time in queue, abandonment rate and many others.

Joe Barkai
www.DiagnosticStrategies.com
jbarkai@DiagnosticStrategies.com

[remy]

I agree with Joe B, however there is ons thing more to consider, with standards as high as CASR says the number of or percentage of abandoned call is probably very low. In the case that you won't fullfill your targets for your servicelevel this measuring the number of abandoned calls is very important, otherwise (if your standard is as high as yours) it won't be such a big issue.
In respons to gkbansal, if you state that ASA is redundant if you're measuring servicelevel I would have to disagree. The thing is that you ignore a portion of your calls by doing that.
F.E.
In theory it is possible that all calls are answered in exactly 31 seconds. this means your ASA is 31 seconds. However your servicelevel will be zero!!!!!! A customer won't be bothered by waiting one second longer but your servicelevel will be zero. And an ASA of 31 is not that bad eitehr
Second example: Imagine that during the day one has an enormous waitingline for just a short period of time (an hour or so), All the calls in the that period will have a very long asa and a will fall outside your servicelevel norm. Over the day as a whole an hour of calls falling outside servicelevel won't matter if the usual servicelevel during the rest of the day is okay. Instead of scoring 92% you'll get 90%. However ASA will increase with a lot more than that and so will the abandoned calls and max delay.

Hope this stresses that looking at just one of the metrics just isn't enough. To get a complete picture you have to look at the whole package

[Steve Turton]

Lots of good points by all.....

But I agree, You can't just look at one report.


Steve
www.telax.com