The 5 Why Questions, Like All Roads, Lead to Rome

Today Patrick Shumaker from Gemba forwarded a great example of asking “Why?” persistently until the root cause is found. Why are U.S. standard railroad gages 4 feet 8.5 inches? Rome. You can find the full story on this bulletin board.
The “ask why 5 times” approach is also great way to find one of the seven wastes, the waste of processing. Often very hard to spot, but very easy to eliminate once you do, the waste of processing is an otherwise good process that simply isn’t needed or does more than needed. Typically the process is done “because we’ve always done it that way” or because of a long-forgotten or long since changed requirement.
I ran into that recently when one part still had to go through a batch oven process while another one 20 feet away did not. After going through the 5 why questions it turned out that they knew the oven process was not necessary, but it would be too much hassle from their customer to get the specs changed… So they were smart enough to design this waste of processing out of new parts, but the cost of removing it from older parts was too high due to another waste of processing, paperwork approvals.

2 Comments

  1. Chris Nicholls

    March 14, 2007 - 4:00 am

    Hi Jon
    Thanks for all your posts recently. My comment on the Railroad Gauge…………
    Isambard Kingdom Brunel the greatest Engineer of all time tried to change the railroad gauge from 4ft-8.5ins to 7ft because he concluded that trains with this wide gauge would run smoother. He actually constructed a number of major railroads in Britain using the 7ft gauge. But the increased size of track and rolling stock etc. increased the cost compared to the 2 Roman horses’ bums gauge used by his competitors. So the Kaizen lesson here is don’t change anything unless the improved method produces the lowest cost product.
    Best wishes
    Chris

  2. Sonu

    April 1, 2007 - 4:36 am

    Is zero defects possible? If so, what are the steps.
    We are dealing with around 300 parameters each having minimum 25 to 30 parameters to be met.
    We find it difficult to maintain zero defect for all parameters. Any thoughts???