- This topic has 3 replies, 1 voice, and was last updated 18 years, 3 months ago by Carlo Trobia.
21st April 2002 at 07:00 #34311JavierGuest
I have many samples of qty of simultaneous calls in an interval, BUT not # calls per hour OR average call duration… can I make some kind of aproximations so I can calculate BHT?.
Any suggestion will be welcome.
Thanks.-21st April 2002 at 07:41 #34312AlexandraGuest
I have the same problem but in my case we are trying to determine the optimum number of lines for a CallCenter
Alex31st May 2002 at 23:25 #34313RommelGuest
Javier…I see that happen all the time and really if you are going to use some sort of calculation you just need more information. You can _assume_ hold times of 3,5 mins, but then again that would be business level traffic. Take the calls in the interval and change that interval to 1hr. Multiply that by 210 seconds. Divide by 3600 and you will have # of erlangs. Its in accurate but it gets a number.
Alexandra…you have to do the same thing but you will have to ask them the average hold time….you need to different call centers have different hold times….operator services have an avg hold time of 9seconds and avg speed to answer of 2 seconds…something like that. While a call center taking travel trips may have an average hold time of 35 minutes with an avg speed to answer of 2,5 minutes. You need their metrics, or you will get the calculation wrong. BTW…dont forget to get the number of agents.1st June 2002 at 12:32 #34314Carlo TrobiaGuest
Also, it’s diffused best practice to consider the peak hour traffic to size any telecom system.
Therefore, sample of the quantity of calls should be considered in their BH (Busy Hours)occurrence.
Normally, telecom operators do disclose some information about their traffic, like the total monthly traffic in minutes, in “per subscriber” rate or such.
This _can_ be valuable starting point to populate _any_ model.
Again: every model should show the sizing, behavior, performace of a system under stress, assuming that this is the most delicate criterion to watch.