Having the correct problem definition is fundamental.
How many times would you expect to solve the same problem? Once I hope. If it reappears then you didn’t solve it the first time; it was possibly solved by one of those instant solutions from a guru, which has been used a hundred times before, which never truly solved the problem, only gave short term improvement.
Pressures of business mean there is rarely enough time to follow the correct methodology and find true root causes of the crisis. Strangely there is always enough time to work on it again and again and again.
Most people accept now that there is a root cause to every problem; the days of saying it’s just one of those things are long gone – the dinosaurs are slowly dying. Managers understand now that there needs to be a culture of systematic improvement, which is the same as a culture of systematic problem solving. Cultures are only defined by the top manager, so the top manager needs to be actively involved in solving the big problems using the same methodologies as the rest of the employees.
The rewards of successful problem solving are sweet. The hard work and disciplines are often the stumbling block to achieving success.
The correct detailed methodology for a problem solving will depend upon the type of problem. I categorise problems into:
Things that go wrong – complaints, breakdowns, …
New levels of performance to be achieved – new designs, improved capability, improved efficiency, lower scrap, …
All methodologies can be based on the six-sigma DMAIC process (originally they were linked to Deming’s PDCA) but I find DMAIC useful and modern.
To illustrate the shortcomings of an existing culture, with respect to problem solving, ask me for one of the two case studies available from RK:QM. Each takes about 45 minutes to run.
1 - The Smolensk Complaint ( 6 managers )
2 - Action Please ( 12 managers )
At each step of the process, there are a variety of common sense tools that can be used. Do not be tempted into thinking that it is important to train people to use lots of tools. That thinking has led many managers to perceive Qualiticians as arrogant; wanting to boast about the number of tools they have and exclude other disciplines from being involved. The important thing is to use common sense and get help when struggling. A good book to help you with tools is Six Sigma and the Quality Toolbox: John BICHENO, Picsie Books ISBN 0 9541 24421
The main steps are:
Define What is this problem? Look from both business and customer viewpoint.
Measue Understand the process, where this has happened and the performance.
Analyse What are the critical factors (root causes) influencing the process?
Improve Develop the improvements for the critical factors
Control Implement the solution and control to hold the gains.
The Team Charter.
Information the team needs to collect at the beginning.
1.1 Why are we doing this?
Impact on the business.
What will the benefits be?
How does this work link to the business strategy?
1.2 Problem definition
Voice of the customer.
Voice of the business
Definition of the problem, how it manifests itself
What are we going to deliver at the end of this project?
What are the targets for improvement?
How will success be measured? Initial definition of Y = ƒ(x)?
1.4 Project scope: SIPOC [supplier, inputs, process, outputs, customer]
What are the boundaries of the project: start and end of process, time limits?
What can we not involve?
What authority do we have?
1.5 Project plan
When are we going to complete the project?
What are the major milestones?
1.6 Team selection
Who are the team members?
What is their role?
How much of their time will be dedicated to the response?
Developing a Problem Definition
An agreed problem definition is vital to provide a consistent understanding, and agreement within the team, about the nature of the problem. This must include the thoughts and aspirations of all those people with an interest in the outcome.
What is happening?
Where does this problem occur?
What are the numbers involved?dimensions, days, % failures, …
When does this problem occur?timings, frequencies, …
What is the impact on the customer’s business and ours?
“The customer is refusing to use the product.”
“Thread deformation at the base of screws, on design type 3792, delivered to Slug tractors on 14 March in our batch number 8AD0205.
The frequency of the defect is 1 in 24 pieces, causing major damage to castings on assembly.”
Tools to help with problem definition
Cause & Effect diagram (Ishikawa diagram, Fishbone diagram)
Definition of the gap between the actual and expected process output.
“What, where, when, … “ does this not happen, when we would expect it to happen.
Voice of the customer: problem in the eye of the customer.
Debate the definition within the team.
Modification of the statement on a flip chart.
A documented record of the final definition.
Laughter is the brush that sweeps away the cobwebs of your mind.
It has been shown that our creativity increases in an atmosphere of fun.
Deciding what to measure in a problem solving is difficult. Do not be deterred, finding the right measures will save time in the end. Taking the short cut and guessing root causes, rather than demonstrating the true root causes with data, will give disappointing results. This is because when guessing, you spend time and money improving features that have little influence on the original problem.
The classic lean-six-sigma formula
Y = ƒ(x)
refers to this model of a process:
From all the ideas we brainstorm about Xs and Ys,
it is necessary to decide which are the critical parameters;
which Xs need an improved control to provide a Y closer
to the customer requirement.
Tools to collect, summarise or illustrate the data.
R&R study ( a tool to ensure data reliability from the gauge )
Tools for analysis
Charts prepared using Minitab
A tool to help us illustrate the relative importance of a collection of parameters which influence the problem. The concept was first used by Vilfredo PARETO an Italian economist 1848 – 1932. Sometimes known as the 80:20 rule. Separates the vital few from the useful many.
A tool to help us illustrate the correlation between the magnitude of the problem (Y axis) and
the parameter we believe is influencing the problem (X axis)
Care must be taken that there isn’t a third parameter that is influencing them both.
A tool to he Before we start producing product in a process, we need a study to ensure the process is capable of meeting the customer’s needs.
If not capable, then we will not achieve sufficient proportions within the specification.
Each picture can indicate actions that will improve the process and reduce waste.
By careful collection of which samples are to be used in the study, other root causes can be identified:
Difference between lines
Head or cavity problems
Design of Experiment (planned trials)
DoE is another way of saying “a well planned trial”; we all know what good planning gives!
The objective is to understand in one trial, the importance of several possible root causes you believe could be influencing the problem.
This tool is an effective use of time and resources; it also allows you to understand permutations of influences on a problem which you could never find with a lengthy series of simple trials.
This is used for business processes as well as manufacturing.
With a good measurement system and an analysis method in which there is 95% confidence, it will be clear which parameters or causes, must be improved or controlled.
Sometimes there will be no time to measure and to provide the confidence that the true root cause of the problem has been found. That’s life. But remember, without measurement there will be a greater risk of the problem not being resolved; then you will need to spend more time working on the issue. “We never have time to do it right first time, but we always have time to do it again.”
A checklist to help find suitable remedies
Can we eliminate the cause?
Can we replace the cause with a more reliable “thing”?
Can we control the cause?
What can be done to make it more reliable?
What procedure needs changing?
Who needs to be trained?
What are the options we have to address the root causes?
Where are the conflicts between the outcomes of an action?
Checking the suitability
Is the remedy acceptable with the customer?
Have all those accountable for the actions agreed?
Have plans to update procedures been agreed?
Have plans for training been agreed?
How is the process to be controlled or monitored;a new QC plan?
Have plans for calibration and R&R% of gauges been included?
Is there an item in the management review meeting?
Is there an engineering solution / policy change involved?
Is it agreed that no action is required?
“People do not object to change, but they do object to being changed.”
Change has two facets, the effect on:
Process; machine, products, procedures,
Culture; the people involved.
Objections to change often claim to be concerns about the process, when the true reason is the social effect.
The Team Charter will explain the need for change. The method of the problem solving will provide the logic for why these actions were selected.
Having those people, who will be expected to change, participate in the problem solving, will help the team understand both the social effects and the process effects.
Get agreement …. It is an art.
Tool for Improvement
If you have two or more alternatives as a potential remedy, a useful logical tool for making the final decision is the Kepner-Trego “Decision Analysis”.
Kepner-Trego Decision Analysis
1. Check that the decision does work
Is there another important cause that we have missed?
To ensure that the problem has been “killed”.
Two classic checks
Review the measures used to find critical parameters at the start of the problem solving, to ensure the performance has not deteriorated.
Interview those people who experienced the problem to check their views.
2. Hold the gains / deploy
Improve the business, not let it decay
1. To avoid a deterioration in the process.
2. To allow others in the business to benefit from the work.
Control the root cause through a procedure, training or audit.
Write up the findings of the problem solving. Show the savings and illustrate the benefits.
Get senior management to issue a process directive for any critical aspect; ask the ISO auditor to include it within the next ISO 9000 audit.
Present the findings at network meetings. Put the report on the Best Practice web site.
A process directive should not be issued without very good reason; perhaps a maximum of 10 in total for a business.
They define in a precise way how some task is to be done; they are mandatory, subject to audit and lead to disciplinary
action if not followed. Only use them for critical aspects in the business.
On-going checks that improvement to a key process is sustained
Check the ongoing performance of an important business or manufacturing process is still satisfactory.
Include a simple measure at management review meetings, with comment upon significant changes, good or bad.
Even if no deterioration has occurred, is the performance still acceptable relative to industry benchmarks.