Failure is not an option … or is it?

How many times have you heard, or even said “Failure is not an option!”  ?

That’s a great movie line, but a dumb way to run in business. If we don’t take risks and try alternatives, there is no progress. Failure is an option, but only if we understand the parameters. Risk taking without a plan, without a mitigation strategy, and without a high probability of known outcomes is not an option.
So how do we option failure? It is simple – use the scientific method. First start with a hypothesis. If x, then y. Next comes the plan to execute x, and finally the mitigation plan if y does not occur. Even if y does occur, it is important to review the whole system and ensure that in the execution of y, other negative consequences did not materialize.
It is easier to look back and identify business failures, but not so easy to identify successes. That’s where metrics come in. Use metrics to measure your successes. Use failures to build your knowledge base. Failure is an option, but make it under controlled circumstances.

I have been involved in many incremental changes, some of them have not worked (been failures), but most of them became the new normal.  All of them provided data and information necessary to make informed decisions.  Speeding up of equipment or lines is a change that almost everyone in manufacturing has been involved in.

• On single line equipment it is as simple as increasing the speed and holding it for long enough to evaluate the product quality and necessary support actions (refilling packaging, product, or other supplies).  At a certain point, it becomes obvious that the speed gains are off-set by the limit of quality or refilling supplies.
• On a multiple equipment line, the complexity of finding the optimum running speed can take days, maybe even month.  Often, when by the time the reliability engineer is called in the line is so out of whack (technical term) that it takes significant research to determine what the speeds were, the last time the line ran reliably.  Slowing down a line in these instances is usually the answer to increasing overall equipment effectiveness (OEE).

Steps for implementing continuous improvement (CI):

1. Start with a stable system.  Results must be repeatable and sustainable to create a baseline.
2. Determine metrics of the system.  These include not only a metric of the change you want to implement, but whole system metrics to ensure that the change did not cause negative side effects.  This includes how data is captures, the formula for the metrics, and how often the analysis will be performed.
3. Possible metrics
   • Overall Equipment Effectiveness (OEE)
   • Cost per unit
   • Waste/scrap value
   • Cost of energy
   • Labor usage
4. Determine the cost / benefit analysis for the change.  This includes the disposition of the product during the experiment.  Is it saleable product, can it be used in rework, are there special disposal costs?
5. Create a written CI experiment plan.  Try to have as few variables as possible.
   • What is the cost of the experiment
   • What is the expected gain from the experiment
   • How long before the return on investment (ROI) will be realized – assuming the CI project is successful
   • How will the decision for final implementation (new normal or return to base state) be made; including the timing for the decision
     • Is there a plan for early termination should the negative results be clearly evident
     • Clearly define who is responsible for decision making
     • Specify date results will be implemented
   • Clearly state product disposition
6. Get the CI plan approved by leadership, including funding
7. Create a written process deviation plan
   1. Post at machinery if possible
   2. Have face to face interaction with each machinery operator to inform them of the plan and their specific duties to the plan
      1. Data collection
         1. What
         2. How often (frequency)
         3. Where (make it easy for the operators to collect/report data)
      2. Tagging of materials
      3. Escalation process (with specific names and contact information) if they need to inform regarding problems/questions encountered
   3. Create clear tagging process (tags, material storage areas) for all material that needs to be quarantined
   4. Start/stop time of deviation
8. Analyze results quickly and get approval from sponsor for decision
   1. Create an executive summary of the project
   2. Create an implementation plan if new process is to be implemented
9. Post results to all stakeholders (operators, management, support functions)
   • If CI is the new operating process
     1. Update process documentation, including date process is to become effective
     2. Train all operators on the new process
   • If new process will not be implemented, clearly communicate that to all stakeholders
10. File all relevant documentation including executive summary
    1. In product folders
    2. In equipment folders
11. Thank operators for their help
12. Keep monitoring data, and devising new CI projects

The key to any successful continuous improvement activity is

• Baseline the current state
• Determine the changes to be made
• Allow time for the changes to become the new normal, and then evaluate the data (OEE, or other measure) to decide if you are going to institute the change, or go back to baseline.
• Make a decision
• Implement decision
• Document, document, document
• Communicate, communicate, communicate

What has been your experience in risk taking?  Did you implement the change, or pull back to original state?  Why?