ICMI Contact Center Forum

HomeContact Center Best Practices Forum

Looking to share and discover contact center best practices while networking with peers? This space is geared to serve the needs of contact center leaders looking for strategic direction, benchmarking data, or general advice.

Forecast Accuracy Percentage Calculation Messages in this topic - RSS

Guest


7/13/2007
Guest
As per COPC standards, which formula is used to calculate Forecast Accuracy percentage?

- Yogesh Rathod, Orbit Consultancy
0 mark as "accepted answer" link
Guest


7/27/2007
Guest
Hello Yogesh,

The formula for Forecast-to-Actual Percentage is: (forecast-actual)/(actual) This is the most commonly used formula in determining forecast accuracy.

I hope this information is helpful to you.

- Mohamed Haqqi, CookDoor, Egypt
0 mark as "accepted answer" link
Guest


8/10/2007
Guest
Hi Yogesh,

Your question is very interesting. We are certified COPC and there is no standard for this purpose. However, COPC recommended to us the following formula: (actual-forecast)/(forecast).

I hope this helps.

- Steve Lacroix, Desjardins Group
0 link
Guest


8/10/2007
Guest
There are two measures of forecasting accuracy:

Normal Method and New Method

Normal Method considers only the day end total and performs the calculation. But it doesn't consider any interval level forecasting accuracy. Hence it doesn't represent the correct picture. For example, consider in one interval we received 10 calls more than forecast but in another interval we received 10 calls less than forecast -- the Normal Method will balance the effect (+10-10), so day end forecasting accuracy will look good.

Normal Method:

1-(ABS(Actual Calls-Forecast Calls))/Sum(Forecast Calls)

(Note: ABS means take the absolute value, so negative numbers are treated as positive numbers)

Here Actual Calls and Forecast Calls are day end values.

New Method:

1-(Sum(ABS(Actual Calls-Forecast)))/Sum(Forecast)

Here Actual Calls and Forecast Calls are for each interval and the same is added to calculate the day end forecasting accuracy.

This gives a correct picture of the whole day.

- Naresh Kumar, HP
0 mark as "accepted answer" link
Guest


9/6/2007
Guest
I agree that forecast accuracy should be evaluated at the interval level; however, we take it a step further and measure the forecasted workload to actual workload at the interval level. If you don't forecast your AHT appropriately, a good volume forecast can be meaningless, especially if you work with high handle times. Something to think about...

- Amy Hugunin
0 mark as "accepted answer" link
Guest


11/16/2007
Guest
Hello Yogesh,

According to the COPC standards you need to calculate the accuracy on the interval level and on the monthly level. The interval level is called the schedule accuracy, which measures the accuracy per interval (the prime intervals, or the intervals that contain most of your offered calls over the day). The second level is the staffing accuracy, which means you forecast the number of calls for a particular month. That helps you to make your recruitment decisions. For both levels, the accuracy is calculated by dividing the actual number of offered calls received over the forecasted number of offered calls (Actual/Forecasted).

I hope this helps.

- Green
0 mark as "accepted answer" link
Guest


4/7/2008
Guest
Hi, I was looking at the answers posted and from my point of view the formula is what I would call the percentage error of the forecasting.

I was looking for which would be the correct way of calculating accuracy, which in my understanding should always result in values between 0% (0 accuracy) and 100% (perfect accuracy). But I understand that a value of, for example, 120% accuracy is meaningless, as could be the case using the formulas posted by other answers.

Maybe it's just a question of interpretation of what accuracy really is...

- Nicolas
0 mark as "accepted answer" link






Powered by Jitbit Forum 8.3.7.0 © 2006-2013 Jitbit Software