Controls Charts for Forecasting Vs Actual
- May 21, 2018 at 1:24 am #56004
Being relatively new to LSS, im still trying my hand at the huge range of tools we have at our disposal.
One objective we have been asked to look at for our manufacturing department is our planned hours to machine products, compared to actual time taken. We are looking to create the output of this exercise into a KPI, for us to monitor and improve on.
Products are typically no more, then one off and vary in size from one part to the next. The data is gathered through route cards, whereby the manager will allocate the planned hours initially and the operators will then note the actual time taken non the same route card. The initial thought was to create control charts for each operative, as they have their own skill set, control charts as this will tell us if we are under/over planning their hours.
However i would like to use this information to understand Cp/CpK values for the overall department and highlight it as a KPI. The bit im confused about is should i be taking an average of all the individuals data, to create a departmental Cp/CpK value, as im not sure of any other way of working out the departments Cp/CpK values? Also with this approach we could potentially see all results balance out in affect to show good figures so still unsure if this is the right approach?
Any help will be appreciated, and have attached some sample data i have been playing with to sound out the idea.May 21, 2018 at 6:17 am #202565
@tgroves-exta I can understand doing a regression on actual vs planned. Should be a 45 degree line 1:1.
The Control chart and Cpk I don’t understand. We are going to do a little heresy here. Cpk is in general a non value add number. For the most part, particularly this case, it is a number that has little or no meaning. Quality people love this number for that reason. It lets then talk in terms that doesn’t mean anything to anyone except another Quality person. If you want to do the Capability study to communicate with management – not an issue. From your other posts management seems to be disengaged so I am not sure why you would bother.
The control chart is an odd question but you made the statement the things we make are one off. Plotting time will be and issue if everything is different. Maybe plotting the difference between planned and actual by operator or by machine type or something like that would be interesting to look at. If you are doing planned time vs actual time (plotting the difference) you might try to capture the data in a way that would allow you to display it in a Multi-Vari chart.
If you want to swap idea about this you can email me at firstname.lastname@example.org. You probably don’t want to repeat my remarks about Cpk out loud. It will probably get you an emotional response.May 22, 2018 at 7:47 pm #202571
what problem statement are you attempting to tackle?May 29, 2018 at 9:01 am #202603
Apologies for the late response.
The problem statement i should have started with (DMAIC) “There is no way of monitoring how accurate the forecasting of the machine shop workload is. The knock on affect of this, is to the work plan schedule”
As you can imagine this will create slippage on agreed delivery dates. hope this clarifies what we are looking at.
I am still awaiting data to come through for this, to start playing with.
RegardsJune 1, 2018 at 12:37 am #202606
Cp/Cpk is an attempt to replace a control chart with a number. It is a statistic that changes over time.
You must be logged in to reply to this topic.