Fail Fast

We have all heard the expression “eat your peas”.  Simply stated, it means to do the tough tasks first, and then enjoy the rest of the process.   This applies to business as well.  Doing difficult tasks first eliminates uncertainty and clears the way for easy success later.  Cutting your losses early has the same effect, freeing up time and resources for other important tasks.

Cumulative effect of failure

Most people feel pretty good about performing at the 98% level – 2 failures per 100 attempts.  For an office worker processing 20 orders a week, that would be 2 errors for the entire week.  Not bad, right?  Probably not, if viewed by itself.  Now imagine the order processing is a 9 step process, involving 9 different workers or departments, each with the same 98% accuracy rate.  The table below shows the cumulative effect of a 2% failure over the 9 steps.  The cumulative effect is 83% acceptance rate, or a 17% failure!

Table 1:  cumulative effect of failure % good 98%           $/step $100
Process step 1 2 3 4 5 6 7 8 9
Acceptable parts 98.00 96.04 94.12 92.24 90.39 88.58 86.81 85.08 83.37
cumulative cost $100 $200 $300 $400 $500 $600 $700 $800 $900
cumulative cost of failures (out of 100) $200 $792 $1,764 $3,105 $4,804 $6,849 $9,231 $11,939 $14,963

Even worse, if each step costs $100 and each the errors are not caught until “final inspection” at the end of the process, the rejected processes cost of each of the failures is $900, or a total of $14,963 for 100 orders processed.   An effective “Check Do Check” program, where each process auto-inspects and errors are not passed forward for further processing, keeps failure costs at a modest $1,800.

Defect rates as low as 2% or higher are not uncommon in business, especially in office environments and processes with high variation.  Years ago silicon wafer fabrication had failure rates 30% or higher (I only hope it has improved)!  Long processes with such high losses force companies to build extra product, or “just in case” inventory, with all the associated costs of both time and money.  So of course all efforts must be made to control processes and eliminate failures (using both Lean and Six Sigma methods).  But some processes remain problematic, either due to difficult materials or a lack of process control.  What then?

The Importance of Failing Fast

Let’s consider a very difficult stage in the process, with low success rates.  Often the most difficult task, the one with the highest opportunity for failure, is performed last, the thinking being this gives us the best opportunity for success “having done everything else correctly”.  The problem with this thinking is that, though comforting, failing at the tail end of a long process is far more expensive, as measured in both time and money.  To illustrate this, Table 2 below shows the same success rates for processes 1-8, but step 9 has a success rate of only 60%.  At process step 9, 40% of products fail, at a cost of $900 each.  Cumulative failure goes from 17% in the above example to 49% as a result, and cost of failures jumps to $44,059.  Cumulative costs of “Check Do Check” are somewhat under control at $5,600.  Where should we perform this difficult task?

Table 2:  Hardest process last % good 98%   critical process (step 9) 60% $/step $100
Process step 1 2 3 4 5 6 7 8 9
Acceptable parts 98.00 96.04 94.12 92.24 90.39 88.58 86.81 85.08 51.05
cumulative cost $100 $200 $300 $400 $500 $600 $700 $800 $900
cumulative cost of failures (out of 100) $200 $792 $1,764 $3,105 $4,804 $6,849 $9,231 $11,939 $44,059

The Process shown in Table 3 has the critical, high failure process first.  The failure rates are the same, and, unchecked, the cumulative cost of failure remains $44,059.    However, if the product is inspected after step 1, cumulative losses can be controlled to total $8,830, saving a whopping $35,229 or 80% the cost of failure!

Table 3: Fail Fast – hardest first % good 98%   critical process (step 1) 60% $/step $100
Process step 1 2 3 4 5 6 7 8 9
Acceptable parts 60.00 58.80 57.62 56.47 55.34 54.24 53.15 52.09 51.05
cumulative cost $100 $200 $300 $400 $500 $600 $700 $800 $900
cumulative cost with inspection at step 1 $4,000 $4,240 $4,593 $5,054 $5,619 $6,283 $7,042 $7,892 $8,830

How is this possible?  This is due to the fact that we add value to each processed part, be is good or bad.   Doing the most problematic process early and inspecting parts at that time ensures we cut our losses early and don’t throw good money after bad by further processing reject parts.  An added benefit of cutting losses early is an improved ability to plan for time and resources – the relatively higher reliability of remaining processes helps control “just in case” inventory and the associated costs.

Lean is all about speed.  Deming’s PDCA pushes experimentation (and its associated failures) to the lowest level possible.  Toyota Kata is based on rapid experimentation and failing fast.  I’ve just put some numbers to it.  Fail Fast.  Eat your peas first!

Copyright 2013 Paul Yandell.  All rights reserved.



One thought on “Fail Fast

  1. Pingback: GROWING AS A LEADER | paulyandell

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s