Forecasting the call volume

Six Sigma – iSixSigma Forums Old Forums General Forecasting the call volume

Viewing 5 posts - 1 through 5 (of 5 total)
  • Author
  • #48972


    Dear Sir/Madam
    Dear Sir,
     I want some help about forecasting.

    I have day wise Nov’07 Total calls, Emergency calls, and Dispatches. How to forecast 01st Dec’07 Total calls, Emergency calls, and Dispatches
    I have 31st December 2005 and 31st December 2006 Total calls, Emergency calls, and Dispatches. How to forecast 31st December 2007 calls.
     Sorry for taxing your time
     Thanks & Regards



    Hi Nalini,
    My name is Tarak and i myself pretty much involved with workforce management and prediction and would suggest you to go to the following link


    Six Sigma guy

    Crystal ball is good software for volume forecasts- which uses monte carlo simulations.
    Also the below article would help


    Michael Schlueter

    Dear Nalini,
    I fear your data base is too small to do reasonable predictions. E.g. we do not know about seasonal variations in this business. Neither do we know about annual variation and so on.
    If you can’t obtain more information probably your best estimate is using the average value of your data and assign it to whole Dec’07 or 31st Dec’07, respectively.
    Unless we know about the inner dynamic, the inner forces of this call-business we can’t really say anything with enough confidence.
    E.g. just assume this call-business works like flames do. So a call generates many new calls, which generates more new calls, until the whole system breaks down.
    Or assume waiting times are intentionally increased day by day. Probably more people get discouraged using this call property, until hardly anybody continues calling.
    Statistics probably can’t hold your answer, while insight and understanding might do.
    Hope this helps, Michael Schlueter


    Mark Chockalingam

    Hello Nalini,
    I agree with Mike.  The data is too small to forecast for December.  Typically in the call volume forecasting situation, the day of the week and the month of the year are important factors.  Although November daily data will give you some clues about the day of the week effect, it will miss the calendar seasonality posed by December.
    I work for a forecasting consulting firm.  In our experience, the day of the week plus holiday interventions plus the monthly seasonal dummies are very critical in developing a daily sales forecast.
    For modeling, any statistical software package capable of ARIMA would definitely be helpful.  However, you may want to ensure that the software can produce ARIMA with interventions. 
    For data, you need to have at least one year’s worth of data to do justice to capture all the calender seasonalities and interventions.
    If you are hard pressed, one solutin in your situation will be to use the day of the week average for November and scale this up for expected DEcember growth to come up with the December forecast. 

Viewing 5 posts - 1 through 5 (of 5 total)

The forum ‘General’ is closed to new topics and replies.