Warranty Return Rate - How do you calculate Warranty Return Rate?

[Warranty_Guy]

Typically the return calculation is based on the customer matrix of measurments. Example would be $10 spent on 10 vehicles with 10 claims would be a $1 dollar repair cost per vehicle,
The repair count would be based on a minimum number of vehicles SOLD. example would be 1000 vehicles sold and one repair = 1 r/1000 vehicles sold
If you try and normalize or place a ppm value it will not work. The customer is focused on vehicles as one part and you may supply two or more parts per vehicle and that would not be the same calculation.
I can give you more information if needed. Please respond back to this and i can contact you

[gvalenti9]

Warranty Guy, Are you saying to calculate repair costs as a % of sales? This metric is an interesting one. I've seen instances where the return rate is really quite high, but the repairs costs as a % of sales are low. This does say something. But, you have to be careful in how you interpret this metric. If there is an inexpensive part that fails, this metric will look good even through the return rate is high.

[Helmut Jilling]

In Reply to Parent Post by Warranty_Guy

Typically the return calculation is based on the customer matrix of measurments. Example would be $10 spent on 10 vehicles with 10 claims would be a $1 dollar repair cost per vehicle,
The repair count would be based on a minimum number of vehicles SOLD. example would be 1000 vehicles sold and one repair = 1 r/1000 vehicles sold
If you try and normalize or place a ppm value it will not work. The customer is focused on vehicles as one part and you may supply two or more parts per vehicle and that would not be the same calculation.
I can give you more information if needed. Please respond back to this and i can contact you

I think you need to look at several methods: incidents, costs, frequency, if you want to improve it. A single metric won't give the whole picture. I would think the customer will begin with incidents as a percentage. And, perhaps dollar cost.

[Warranty_Guy]

You really cannot track warranty as a percentage of sale, the cost of the repair is so non normal that you would never show or pick up a trend on your failed part. I would only and only use the customer matrix to track my performance. We did attempt once to rationalize the warranty by placing a ppm number to it and the Customer had big issues with that. The Typical methods used
IPTV=Incidents per thousand vehilces (GM)
CPV= Cost Per Vehicle (GM)

CPU= Cost per unit (FORD)
R/1000= Repairs per one thousand vehicles sold (Ford)

EPUS= Expense per Vehicle Sold (DCX)
C/1000= Conditions per 1000 vehicles sold (DCX)


these in my opinions are the only matrix that should be used in defining feild failures. if you start to normalize your data and split it out by part not by vehicle system you quikcly end up managing the data and not the problem.

That is not to say that we don't use ppm in the internal Six Sigma or other process for improvement. The only challenge is trying to convince the customer that there systems is not robust enough to really define process shifts or failure rates.

P.S We have improved on some components without doing anything to improve the process. That just shows you how the system is so senstive to sale shifts part mark ups and dealer labour costs. As we all know the original intent of the systems were to get the dealers paid, it has evolved into what we use today

We should really look at getting a Warranty Forum started, there is a lot of good sharing that can be gotten

[xfngrs]

In Reply to Parent Post by Warranty_Guy

You really cannot track warranty as a percentage of sale, the cost of the repair is so non normal that you would never show or pick up a trend on your failed part. I would only and only use the customer matrix to track my performance. We did attempt once to rationalize the warranty by placing a ppm number to it and the Customer had big issues with that. The Typical methods used
IPTV=Incidents per thousand vehilces (GM)
CPV= Cost Per Vehicle (GM)

CPU= Cost per unit (FORD)
R/1000= Repairs per one thousand vehicles sold (Ford)

EPUS= Expense per Vehicle Sold (DCX)
C/1000= Conditions per 1000 vehicles sold (DCX)


these in my opinions are the only matrix that should be used in defining feild failures. if you start to normalize your data and split it out by part not by vehicle system you quikcly end up managing the data and not the problem.

That is not to say that we don't use ppm in the internal Six Sigma or other process for improvement. The only challenge is trying to convince the customer that there systems is not robust enough to really define process shifts or failure rates.

P.S We have improved on some components without doing anything to improve the process. That just shows you how the system is so senstive to sale shifts part mark ups and dealer labour costs. As we all know the original intent of the systems were to get the dealers paid, it has evolved into what we use today

We should really look at getting a Warranty Forum started, there is a lot of good sharing that can be gotten

I know this is an old thread, but I think this applies. Does anyone know how GM calculates their IPTV currently?

[Statistical Steven]

I like the rolling period to get the average warranty cost per unit. The 12 month window makes sense if you have a 12 month warranty. I would have the window be equal to length of the warranty.

[No_tengo]

I'm also interested to know how GM calculates their IPTV and if we can see this metric in their GM Supply Power portal.

[George_Ji]

In Reply to Parent Post by Statistical Steven

I like the rolling period to get the average warranty cost per unit. The 12 month window makes sense if you have a 12 month warranty. I would have the window be equal to length of the warranty.

Agreed,reasonable and more accruate.