Hi all,

I've been reading the forums for a short time and you guys are always very helpful to question. I'm hoping you guys can help. I have a couple of questions relating to Reporting using the data from Symposium. I have been put into the position of centralising our Contact Centre's stats and am loving every minute of it ;). Our Contact Centre provides Customer Support for Technical, Sales and Billing enquiries for an Internet Service Provider.

The first one is probably an easy two part question. It may seem a little silly, I apologise in advance. Here goes.

- What formula is a good starting point to use when calculating Service Level? The SL target the powers at be have set is 80/60. We currently use (Number of Calls Answered in 60secs/Total Calls Answered.) I've been told to use (Number of Calls Answered in 60secs/(Total Calls offered - Number of Calls Abandoned in 60secs))

- Our Tech Support and Sales deptartments run 24/7 - makes life easy for my reporting. Our Billing team only work business hours. Outside of hours and on weekends customers get presented with a message advising they have called after hours. Some of these callers hangup (abandon) some of them hold on until the IVR terminates the calls. I believe the that when the IVR terminates the call this is record in the CallsGivenForceDisconnect field of the database. Can anyone confirm that this is correct?

Thanks for your help, I think I've given enough information for you guys to help, but if there is something I've missed please let me know.

Dan

I can remember having what i thought was an issue with disconnects.

Turned out it was because i was routing the calls to Zurich out of hours but that this was recorded as disconnected.

What i found with symposium though is that if you have full access to the system its fairly easy to set up a test line, write a little script that does what the live lines do overnight, i.e. play message and then disconnect, throw some test calls at it and then run a couple of reports to see what is recorded in the database.

Service levels, depends on what is agreed really, i have also used the formula (calls ans within service time + calls abandoned within service time) / calls offered.

Definition: - The number of local and incoming calls given force disconnect treatment for an application. What this means in real terms is that wherever you have "Disconnect" as a line in your script, calls meeting the condition will show under "CallsGivenForceDisconnect". A very common use is at the end of a "office closed" message.

As to service level calculations. This depends largely on what your are trying to show. It can be as simple as Calls Answered< Threshold/Calls Answered.

HTH

William, I'm not in a position at the moment to test by creating my own script. By saying "at the moment" I don't think it will ever happen. :) I am in bright and early this morning to test live.
Why do you use (Calls Answered in Threshold + Calls Abandoned)/Calls Offered? What are you trying to show here?

Ian, thanks for the definition, it sounds like what I'm after. Is there a valid reason to include the Calls Abandoned in the calculation? Or maybe the question shoudl read, what would be a valid reason for including/removing the abandoned calls?

Thanks for the response so far.

Dan

We only used calls abandoned within threshold.

The reason we did that was that we argued with our client that it was unfair to penalise us for calls we may have answered within the threshold but were abandoned before we had chance to.

((Answered+Abandoned)-(Answered over threshold + Abandoned over threshold)) /(Answered+Abandoned)

In my personal opinion the most literal definition of service level

HTH

DaveA

Dave, can you tell me the reasoning behind that formula?

Dan

The formula given by Dave is the standard calculation used by Symposium (and Meridian MAX).

The logic is:
- You want to measure the calls that are contained within your time threshold (the time in which you say you'll answer x% of calls) against your total calls.
- You factor in abandoned calls as you are wanting to look the service given to everyone who called and not just those who had their calls answered.
- The service level is reduced by having calls answered or abandoned after your threshold level (the point by which they should have been answered).
- Calls that abandon before your threshold do not impact on your servive level.

As previoulsy mentioned, the calculation you use is dependant on what you are trying to show combined with a good portion of personal preference.

HTH

Hi all, we use ((Ans - Ans aft Thres) + (Aban - Aban aft Thres))/Offered.

All similar if you ask me.


Does anyone use Microsoft Query to access the SQL database of symposium?

A.S.

Just about to start using this after permission granted by BT last week that this is an approved method of ODBC(even though Nortel reference it in their own user guides).

Made one successful test connection before getting bogged down in day to day matters but was able to bring back a full set of data for a given date from dAgentSkillsetStat.

I must be busy this is my first visit to the site in a week and first post in I dont know how long.

Thanks for the explanation Iain, it does help.

Andy, I only use this method to get data. I don't have access to Crystal Reports, so Excel and MS Query are my friends. I find the Standard Reports in Symposium hard to look at, so I get all my information from the database for what I want to see.

Dan

Further to Dave's formula (admittedly from some years ago!), which is consistent with the calculation used within the Nortel setup:

((Answered+Abandoned)-(Answered over threshold + Abandoned over threshold)) /(Answered+Abandoned)

what are the advantages/disadvantages to using that formula for service level calculation over the following:

(Answered within threshold) / (Answered+ Abandoned) - Abandoned within threshold

The comments I've read here reflect my own view that short abandons need to be removed from the calculation, but it's only upon starting to use a Nortel system that I've seen the first calculation used. Previously, I'd only ever encountered the second calculation.....