Faster Than You

Posted by stephens on Sep 2, 2009 in Automated Transactions, Engineering Decisions | 0 comments

My pal Eric Kam has an excellent post about estimating dies.

He is correct in that most shops do not know their actual costs to build a one-off stamping die.

Eric is also correct that the estimate of costs by these shops has a high degree of variation.

I think the intellectual argument he suggests is valid, but the culture of the die business will get hung up on the actual results. The problem is the die business estimates by essentially trial and error guesswork.

Two estimators five steps away from each other in the same shop cannot agree on the estimate, yet these die shops like to argue the validity of the cost from software.

My approach would be much simpler.

I would challenge the estimating department to a challenge. I would call it the “Faster than You” challenge.

Here is how it works:

Let’s say the die shop has five estimators. Give everyone in the estimating department the same product / press / process data to quote to their way.

Show up with five users of the estimating software. Give them the same product / press / process data.

Have everyone start estimating at the same time and clock the results. Without even being there, I know what the result will be:

• Each of the die shop estimators will finish at different times with the slowest taking 50 – 100% longer than the fastest.
• Each of the die shop estimators will arrive at a different cost. The difference between the high and low estimates will be at least 30%.
• Each of the software guys will finish at roughly the same time with the slowest taking 5 – 10% longer than the fastest.
• Each of the software guys will arrive at exactly the same cost. There is no variation in cost given the same product / press / process input.

My argument is two-fold:

1. The die shop’s own estimators cannot agree on the cost using what amounts to trial and error guesswork as quoting method;
2. Since the die shop lacks a structured process, the estimating times are as unpredictable and random as the cost estimate.

The focus should be on eliminating variation in cost and process time. Solve that problem and the software sales will follow.