Criteria to raise a Nonconformance based on KPI values

[Mike S.]

My experience doesn't match yours. I'm neutral about red-amber-green, but I have seen it in place in some extremely sophisticated companies with highly intellegent and educated leaders. If it works for them, don't knock it.

But does it work for them, or, does it work for them as well as if they would put some effort into understanding things better?

Not too long ago I was at a plant belonging to a multi-billion dollar aerospace company where the #2 guy in the company raved to their customer (an aerospace prime) about their red-yellow-green andon lights posted above each bulletin board, which was located in an alcove just off the main hallway. The idea was everyone who walked by could just look at the lights and see if things were going great, so-so, or badly (green, yellow, red), thus no need to bother with walking a few steps over and actually looking at any details on a chart!

After raving about his idea and how the corporate VP (his boss' boss) loved the idea so much that he was gonna see to it that every plant in the corporation copied the light idea, the VP took a breath and the customer asked a question: "I see the light here is green, but what about this data (pointing to a piece of paper on the bulletin board) showing how for 3 weeks in a row you are behind schedule in shipping this product?"

The VP got a funny look on his face, stared at the data for several seconds, conferred briefly with a subordinate who was also on the tour, and then explained that manpower had been taken off that job and moved to a hotter job because they are in a very competitive area for aerospace manufacturing and manpower is hard to find, but they have 23 unfilled employee requisitions out there right now and HR is doing their best to fill them.

So the light was green, the VP thought all was well, but all was not well, not well at all.

That's a great way to convince your customer to give you more business, right?

 

[Jim Wynne]

I was in a plant several years ago where they had a video monitors at several machines that were supposed to monitoring the SPC data. They too had red, green and yellow things to indicate...well they indicated something. I asked two managers and an operator a few questions about how these things worked in terms of the data and got three different conflicting answers.

 

[RoxaneB]

After raving about his idea and how the corporate VP (his boss' boss) loved the idea so much that he was gonna see to it that every plant in the corporation copied the light idea, the VP took a breath and the customer asked a question: "I see the light here is green, but what about this data (pointing to a piece of paper on the bulletin board) showing how for 3 weeks in a row you are behind schedule in shipping this product?"

Without knowing what the "green" was reporting on, I'd offer that maybe it's not the colour-system that's the issue, but rather the metric being reported.

 

[RoxaneB]

I was in a plant several years ago where they had a video monitors at several machines that were supposed to monitoring the SPC data. They too had red, green and yellow things to indicate...well they indicated something. I asked two managers and an operator a few questions about how these things worked in terms of the data and got three different conflicting answers.

Is the issue with the red/yellow/green method or is communication/training/awareness also at play there?

 

[Bev D]

Red yellow green can help people visualize or see the situation better. IF - and this is a big IF - the colors are assigned according to the statistical rules that properly represent differentiation of a signal from random noise AND the data are appropriately defined and measured/assessed AND the data are accurately input into the system. In the case of the Andon - as explained, it is not weakness of the colors, but the lack of changing the color as the situation changed - in other words inaccurate data into the reporting system.


In circling back to the OP's original question, we need to all remember that are 4 ways to meet a goal
1. Hope that random variation plays in your favor
- If the goal is 'small' enough you might get lucky and the random variation in the system will be within the goal during the assessment time.
2. Manipulate the system
- Negotiate the goal
- Negotiate the relevant measurement period
- Negotiate constraints: what is in your control and beyond your control. in other words, don't include results that "weren't within your control"
3. Manipulate the data
- Censor data that is “not fair”
- Censor data that is outside your ability to control
- Rationalize the data
- Lie
4. Make actual changes to the system
- the goal must match the business need, not be arbitrary, inspirational or punitive.
- must have the resources in time, people, money
- must have a method for achieving the goal
- must have a plan and strategy to meet the goal

This is why I cringe when people ask about 'raising an NC' for a missed goal. It is a complex system and requires true knowledge, profound knowledge.

 

[Jim Wynne]

Is the issue with the red/yellow/green method or is communication/training/awareness also at play there?

If we assume that the purpose of the R-Y-G stuff is to facilitate awareness of what's going on with a process, and no one seems to understand the underlying R-Y-G significance, there's a problem that training won't help. The only place I can think of where these signals can be really useful is in a real-time situation where measurements are automated.

 

[Mike S.]

Without knowing what the "green" was reporting on, I'd offer that maybe it's not the colour-system that's the issue, but rather the metric being reported.

The green was to indicate that everything with the production of that product was going fine relative to the parameters being reported on and no intervention was required. But in reality there was a problem lasting 3 weeks and yet the green light shined on. The customer who had never seen the charts before noticed the problem almost immediately because he took a few minutes to look at data and not a stoplight.

 

[Bev D]

Mostly I would agree with Jim.
When it comes to Andons, I have seen them work well. But they do require a true Toyota Production System culture. In this culture everyone knows, understands and expects that work will flow and when a problem is discovered, it is addressed and notification (via stopping the line) is made immediately. "Problems are precious" is the mantra.
I've seen too many 'leaders' implement things like Andons 'hoping' that they will 'do the trick' without doing the hard work to change the culture. In a non TPS culture, turning the Andon on is seen as a sign of failure, so the light wont' be turned on...

 

[Jim Wynne]

I've seen too many 'leaders' implement things like Andons 'hoping' that they will 'do the trick' without doing the hard work to change the culture.

Everyone wants to go to heaven but no one wants to die first.

 

[Bev D]