How do you perform a GR&R for system software generated values?

E

Elbarax

Good Day!

I would like to ask how do you perform a GR&R for system generated values. Say, in a call center, the Handling Time for 1 call is timed by the system. The question is, is it ok to trust the system that when it says 2 minutes its really 2 minutes?

If we will perform GR&R for this, how do we go about it? Its not like in manufacturing where we have a standard gauge.

Appreciate your replies :)
 

Bev D

Heretical Statistician
Leader
Super Moderator
Re: How do you perform a GR&R for system generated values?

For these types of measures a traditional repeatability study is inappropriate. however a measurement system analysis is.

you can check the software to see how it 'measures': what is the start signal, what is the stop signal? is it the correct definition? is it actually correct? if you measure the time does it match the system time? (this is liek calibration) is there work that is not included? does the time that is measured reflect the actual Problem you are working on?

I do this for inventory counts in ERP systems, on-time delivery measures, all cycle time measures, etc. I never accept that the system is correct. it may result in the same values each time, but that doesn't mean it is correct.
 
Last edited:
E

Elbarax

Re: How do you perform a GR&R for system generated values?

Gee thanks for that insight! I was wondering how can I apply the traditional GR&R for systems. So do I sample out say 5 calls per hour for the next 8 hours and validate the actual time?
 

Bev D

Heretical Statistician
Leader
Super Moderator
Re: How do you perform a GR&R for system generated values?

not sure you will need that many, but its good to get a diverse set: some short, some long, some during long hold times, or other times that will stress the system.
 
E

Elbarax

Re: How do you perform a GR&R for system generated values?

oh ok now its Crystal! Thank you very much!
 
Top Bottom