My calculations give different Service Level then SCCS report does

[mlapchuk]

We are trying to simplify some of our month end reporting that is done through SCCS and as such I am creating custom queries in the backend Access CallStats.mdb database. I have come across a problem though where when I query the database for "CallsAnswered", "CallsAbandoned", "CallsAnsweredAfterThreshold" and "CallsAbandonedAfterThreshold" and use those values to calculate the service level I get a value that is off by up to 6.8% when compared to the value I get when I run the "Application Performance" report in Symposium.

I am using the equation that is listed in the 'reports.pdf' manual which is
[{(CallsAnswered+CallsAbandoned)-(CallsAnsweredAfterThreshold+CallsAbandonedAfterThreshold)}\(CallsAnswered+CallsAbandoned)]*100.
And I have compared the base numbers I use and they are the same as the numbers that Symposium reports give me.

Does anyone have any idea why these two numbers may be different? My calculations always give a higher Service Level then the SCCS report, anywhere from 0.678% to 6.78% higher

 

[bajangsa]

Which table do you query? - You will often get better sla if you query the Skillset stat instead of the Application stat, because the waiting time is shorter in the Skillset table.

 

[mlapchuk]

I had queried the Skillset table. After looking at the report in Symposium I used the dbo_iApplicationStat table instead and all but one of my calculated SLA's matched so that problem is fixed. My next question is why would two tables, listing the same information, have different values for that information?

The numbers for CallsAbandoned, CallsAbandonedAftThreshold and CallsAnsweredAftThreshold have different numbers between the Skillset table and the Application table even when queried over the same time period.

 

[mte0910]

I may be wrong, and if so I'm sure someone will correct me, but what I have seen is that the difference between Application and Skillset is the clock. When a call is handed over to an Application from the Master, the clock starts. If you have something like a message at the start of the application, that takes time to do. Then, you assign the call to a skillset, once queued, the clock starts on the skillset. So, while both may show offered, answered, and abandoned, the application can have more time clocked against it than the skillset.

 

[sandyml]

It's as mte0910 says. You have two different things that need to be measured:

(1) The caller's wait experience, which includes all the time from the beginning of the call to the agent answer. This includes mandatory first greetings ("your call may be monitored or recorded for quality assurance...") and any delays for networking, data exchanges, etc.

(2) Your staffing level and center efficiency. Specifically, how long does a call actually need to wait after queuing? Too long may indicate you need to staff up or look for back up skillsets; no wait may actually mean you are overstaffed.

Basic stuff, I know. I may have had too much caffeine this morning..

 

[bajangsa]

In short:
The Average Delay before Answer you see in the Application Stats are what the customer are experiencing.

The Average Delay before Answer you see in the Skillset Stats are the agents performance.

 

[mlapchuk]

Ah ok. That clears it up a bit.

Thanks