Monthly Data Review through Control Charts - Monitoring Complaints

[QE]

Hello Cove members

I need a suggestion. In our monthly data reviews we perform a review of various data streams and trigger actions/investigations.

1. Monthly failures by product and compare them to the relibility goals using Xbar chart
2. Complaints by product using control chart
3. Normalized complaints per 1000000 essential uses
4. Adverse Events using Xbar chart
5. Reportable complaints using Xbar chart
6. Reportable normalized complaints per 1000000 essential uses
using Xbar chart

My question: Is the Xbar chart correct method to monitor the complaints ? kindly advise what chart should be used ?

 

[Stijloor]

A Quick Bump!

Can someone respond?

Thank you very much!!

Stijloor.

 

[Steve Prevette]

You "can't go wrong" with the generic x-chart. I do find that if you are familiar with the Poisson Distribution and the Binomial Distribution, that you can make use of some specialized charts.

For counts of independent events (such as customer complaints), the c-chart can work well. For percent of customer feedback that is of a complaint nature (or on a given product) the p-chart can work well.

The u-chart can be used when you normalize complaint reports per some production population.

If a process that you would expect to be binomial or poisson (that is, independence between individual results) and it does not turn out to be, that can be an important lesson. Dr. Deming related the story of checks for defects at a shoe factory where the defect rate did not fit the binomial (as monitored on a p-chart) showed that data were being falsified.

 

[Bev D]

we utilize the I, MR chart, p, u, c, and p' charts for our customer complaints and run fault rate data.

as Steve said the generic xbar (I, MR) chart does the job for most things. the key point is that SPC is not a precise assessment; it is a (very effective) guide that keeps us from over and under reacting to variation.