Effective Meetings for trend analysis - Corrective and preventive action

[little__cee]

Our corrective/preventive action procedure states:

Process failures are analyzed in the Combined Management/Supervisor Meeting on a monthly basis to evaluate for trends and the initiation of Preventive Actions.

At the present time, all managers create a Pareto chart for their department and discuss the information at this monthly meeting. The charts are hung on a bulletin board after the meeting and I create one summary chart with all of the data on the individual charts.

We're overhauling the system, in part because "process failure forms" sound so negative and changes needed to be made there, so this seems like a good time to improve the meeting process too.

How do others conduct effective meetings for trend analysis? Right now, someone reports "this happened x number of times and it cost the company xx dollars - any questions?" (oversimplification but that's basically it)

I'm thinking I'd like to have everyone follow the same reporting format but not sure if Pareto charts are really the way to go here. I want to know things like "has this happened before? how often? what is the next step to make sure that this doesn't happen again" but really at this point I'm stuck as to what to do next. Any suggestions would be appreciated. :

 

[Wes Bucey]

I'm not sure about what makes the most effective visual aid for your particular meeting, but I do know that the meeting is most efficient by distributing IN ADVANCE a detailed agenda and pertinent data which will be discussed and perhaps voted on for action to be taken.

One person should be assigned to act as recording secretary to make note of the action items and distribute them to all participants and other interested parties as soon as possible after the meeting.

(I have even attended some meetings where the recording secretary takes the minutes and instantly distributes them via email to participants so they are waiting and available when the participants reach their desks after the meeting.)

The reason for advanced distribution of detailed agenda and discussion data is to allow participants to be prepared to actually discuss the action items rather than take time from the meeting to inform themselves about the data. (Some folks have different rates of absorption of information - why make quick study wait for slow study or slow study be left out of discussion because he doesn't have total knowledge of topic?)

 

[Al Rosen]

Our corrective/preventive action procedure states:



At the present time, all managers create a Pareto chart for their department and discuss the information at this monthly meeting. The charts are hung on a bulletin board after the meeting and I create one summary chart with all of the data on the individual charts.

We're overhauling the system, in part because "process failure forms" sound so negative and changes needed to be made there, so this seems like a good time to improve the meeting process too.

How do others conduct effective meetings for trend analysis? Right now, someone reports "this happened x number of times and it cost the company xx dollars - any questions?" (oversimplification but that's basically it)

I'm thinking I'd like to have everyone follow the same reporting format but not sure if Pareto charts are really the way to go here. I want to know things like "has this happened before? how often? what is the next step to make sure that this doesn't happen again" but really at this point I'm stuck as to what to do next. Any suggestions would be appreciated. :

I like pareto charts, so I'm biased. Anyway, why not just use $$ as a basis for the pareto chart. That way the problems that cost the most are higher on the "hit list".

 

[Steve Prevette]

Pareto charts are a fine start but they won't tell you if there are trends. To tell if there are trends (is this going up or down or staying stable) a fine tool is the control chart.

Our corrective action organization uses a combination of SPC and Pareto charts for their trend analysis. Our procedures for doing so are on the internet: Worksite Analysis - Trending Safety Information

 

[Jim Howe]

Pareto Charts

Pareto charts are a fine start but they won't tell you if there are trends. To tell if there are trends (is this going up or down or staying stable) a fine tool is the control chart.

Our corrective action organization uses a combination of SPC and Pareto charts for their trend analysis. Our procedures for doing so are on the internet: Worksite Analysis - Trending Safety Information

Steve I agree but sometimes a Pareto is all you need to show a trend. I was recently tasked with tracking scrap cost. I went back into the data for the past four years and produced pareto's for each year up to the current year. It revealed a definite trend for obsolete material. Not to sure a control chart would work in the above example but I am open minded an always looking for better ways of doing trends analysis.

 

[Bev D]

A simple method I've used is to combine the above methods: Create an initial Pareto chart that documents some base period (say the last 3 months). I typically use occurrence rate as the importance factor for th epareto, because it is an inherently more stable value for trending than $ or customer satisfaction. If $ or Cust Sat are aslo important I simply have the value added as text comment to the Pareto.
THEN, I have the managers report the trend for those 3 months in control chart format (typically using only the average chart for the display). As time goes on the manager updates the trend chart so the group can understand if teh problem is stable or gettign worse or if actions taken were effective.

Periodically, I reset the baseline Pareto once corrective actions have caused a substantial chagne to the pareto ranking...

This tend to keep the meeting focused on the top chronic items (What are we doign about them?) and allows reprioritaization when a true "new item" is large enough to worry about. It also shows when a top item has been reduced to the insignifcant level.

Teh format of the graphics is to place the pareto on it's side (Bars horizontal to the left) with the control chart for that bar to the immediate right of the bar...it make the pareto kinda big and the chart a little smaller than what we may be used to but it is still very visible and getst he point across.

 

[Steve Prevette]

A simple method I've used is to combine the above methods:

That sounds good. A key with Pareto Charts is to make sure that during the time interval the Pareto Chart covers the process is stable. If there is a significant trend in the middle of the Pareto Chart time interval, there are now apples and oranges on the Pareto chart.

 

[little__cee]

Example?

Teh format of the graphics is to place the pareto on it's side (Bars horizontal to the left) with the control chart for that bar to the immediate right of the bar...it make the pareto kinda big and the chart a little smaller than what we may be used to but it is still very visible and getst he point across.

If its possible, could you please post an example of what you're describing? I think I can picture it and know what you're talking about, but I might need more coffee to create one myself. A visual would be great if its not too much trouble. Thanks.

 

[Caster]

Ford QOS Ideas

...I'm thinking I'd like to have everyone follow the same reporting format but not sure if Pareto charts are really the way to go here. I want to know things like "has this happened before? how often? what is the next step to make sure that this doesn't happen again" but really at this point I'm stuck as to what to do next. Any suggestions would be appreciated. :

Ford QOS suggests using 4 charts

• Trend Chart (better to use SPC per Steve Prevette)
• Pareto Chart (for the current period)
• Problem Solving Summary (what are you doing)
• Paynter chart (I've never used one?)

If you supply Ford, they have 2 really excellent presentations on how to do this on their supplier site.

AIAG CQI-5 Business Operating System also is really good, well worth the money.

I suspect a google on Ford QOS will get you al the free ideas you can handle.

We have been about 18 months building our metrics system. What gets measured gets done from what I see. Unfavorable trends in front of executives get action FAST. And it's really nice to see some of this quality stuff work like it is supposed to.

Good luck