Six Sigma
Click here for free tools and know-how materials from the Mud Valley™ strategy & brand marketing community.
Question: why is Six Sigma currently the most favoured of the quality processes?
Answer: because it focuses not only on what to do, but how to do it:
- Six Sigma uses pareto principles to focus you ruthlessly on those areas that will make most difference to your business outcomes. The resultant financial reporting emphasises sales & profits growth, cost reduction or cash flow improvements
- Six Sigma is a heavy user of simple tools and techniques to drive you through the process. If you complete the tools, you are likely to complete the project relatively successfully
- the tools require data. To complete the tools, you must have data, so your decisions will be data-driven
- the process requires that processes, tools and techniques are “replicated” – projects build on the methodologies and findings of previous projects; new tools are developed to address specific issues, and rapidly shared and improved
Key action points
Six Sigma is a relatively prescriptive process. To complete a project, you must go through each step of the process in turn, and use each of the mandated tools.
The DMAIC steps are:
- Define – use the Project Charter to agree what the projected is intended to achieve, what is in and out of scope, and what the benefits are planned to be. The project charter should be supported by a Stakeholder Analysis, and be distilled into an Elevator Speech (to be able to explain to someone what the project is about and its benefits in the course of an elevator ride)
- Measure – use a Process Map to feed the Cause & Effect Matrix (C&E), and the Failure Mode & Effects Analysis (FMEA), to explore the issues. Use Design of Experiment protocols to define the methodology to measure the current state (“Initial Capability”). Use Measurement System Analysis (MSA) protocols to ensure that your measurements are accurate before scaling up your data capture
- Analyse – use Multivariate Analysis to explore the interdependencies within the data, and to predict where changes will need to be made. Any other statistical analysis tools can also be used
- Improve – define the new Process Map, make the changes, and measure the Final Capability to ensure that the process has indeed been improved, and will deliver the benefits required by the Project Charter
- Control – use a Control Plan to mandate how the process should be managed in order to “sustain the gain”. This will include ongoing measurements
Six Sigma came out of a desire to improve manufacturing processes (the name refers to errors measured in parts-per-million), and is easiest to apply to large-scale repetitive processes, where there are lots of data capable of easy, accurate and frequent measurement. It can be applied effectively to more social processes, such as marketing, but this area is much more problematic as the frequency and validity of measurement becomes a significant hurdle.
In more detail………….
For some, Six Sigma is a born-again business religion, which is tiresome, and can often lead to undue complexity as its disciples develop more and more tools and techniques in their striving to reach the promised land.
The promised land in Six Sigma terms is a stable process that delivers the required benefits consistently. The big Six Sigma questions are whether the process is under control and, if so, whether it is delivering the required benefits.
If the process is stable and under-performing, then the process will have to be re-engineered to achieve the desired results. If you do not change a stable process, you will simply carry on getting what you have always got.
To know whether a process is under control or not, you have to use Statistical Process Control (SPC) analysis to understand what are the recurring outputs of the system, and what are “special causes” – random fluctuations.
One of the Six Sigma refrains is that “two data points do not make a trend”. Just because this month’s numbers are better than last month’s, it does not necessarily mean that you are getting better. Maybe next month’s will be even worse than before. SPC charting will allow you to map your results over time, and to define a corridor of results which are the core outputs of the system. You can then judge whether the process is under control, and whether it is improving, deteriorating, or remaining fundamentally at the same level of performance throughout.
SPC mapping potentially has radical implications for the accurate interpretation of financial results. Typically, we judge performance day-by-day, month-by-month, quarter-by-quarter and year-by-year. What we do much less often is to define a performance corridor and ask whether our performance is statistically improving, declining or stagnant.
If the process is out of control, then it must be mapped to bring it under control. If people are doing the same job differently, it is impossible to tell what the capability of the process is, and therefore whether the process itself is under-performing, or whether the people applying it are not using it to its best advantage.
The five DMAIC steps, and the tools that are used
The encouragement to “replicate” tools and techniques leads to their proliferation, however these are the basic tools:
- Define
The purpose of the define stage is to clarify:
- the problem the project is designed to address
- the scope of the project (which should be as narrow as possible to avoid “boiling the ocean”)
- what the project, as designed, can realistically be expected to deliver
- which project is the most urgent, or will provide the best return on investment (Six Sigma projects tend to be resource-hungry)
The basic tools that are used are:
- the Project Charter, that represents the “contract” of the project
- a Stakeholder Analysis, to explore the potential benefits and to whom
- a Project Hopper, as a means of listing all the potential projects and prioritising them
- an Elevator Speech, to help communicate the benefits of the project, and gain buy-in to it
At the end of the define stage, you have decided which projects to focus on, and what resources will be required.
- Measure
The point of departure for the Measure phase is a Process Map. There are two types of process map – the “is map”, which charts how a process is actually being deployed, and a “should map” which says how it should be deployed. At this stage you should stick to the “is map”.
From here you develop a Cause & Effect (C&E) Matrix. This maps the relationship of each process step to the required output (in Six Sigma parlance, the “X” – input step – against the “Y” – outcome). You score each process step for its influence on each of the required outputs, and select the few process steps that are likely to be most critical to achieving the results you require.
You then take these few process steps into the Failure Modes & Effects Analysis (FMEA), to explore:
- what could go wrong
- what happens when it does go wrong
- how often it goes wrong
- how effective your processes are in responding to the failure
- what is likely to be the “root cause” of it going wrong
What you are looking for here are steps in the process that frequently go wrong, cause significant damage when they go wrong, and where you have ineffective provision for rectifying the situation when it goes wrong. The questions then are “Why does this happen, and what can we do about it?”
The FMEA is extremely tedious and frustrating to complete, but also at the very heart of the success of the project. Once you know where the problems lie, you can measure these problems in depth.
Use Design of Experiment protocols to define your research methodology. Before investing significantly in data capture, use Measurement System Analysis (MSA) protocols to ensure that the data is valid and reliable. Statistically, “validity” is the accurate measurement of whatever you intend to measure, and “reliability” means that you will get the same results if the same situation should arise again. For social research, you will have to include statistical significance tests to accommodate the acceptable margin of error in your findings.
- Analyse
The core tools for the analysis stage are:
- Statistical data quality tools to ensure that the data is reliable. These will include tests for normal distribution, seeking to detect skew (where the data are lopsided), or kurtosis (where the data are unnaturally bunched). Typically, you are looking for a normal distribution bell-curve in the shape of your data
- Multivariate analysis tools. These analyse the cross-impacts of several inputs on your desired outcome, and recognise the complexity of the changes you will have to make. It will not simply be about pulling one lever to effect a global change relevant to all situations
You can also use any other data analysis tools that you happen to know and trust, such as correlations, multiple regressions, factor analyses, cluster analyses, multi-dimensional mapping, conjoint analysis etc..
- Improve
In the improve stage, you define your “should” process map – how the process should look to be optimally effective.
You can use all the data you have to model specific scenarios.
You then develop an action plan to drive the changes. This forms the basis of your control plan.
You then measure the “final capability” to ensure that the improvements you have made deliver, in reality, the intended effects. It is, of course, entirely possible, that you do not improve the process as much as you hoped, or indeed that as one outcome improves another deteriorates. If this is fundamental, you need to repeat the entire DMAIC process. This can be disheartening, but you will already know far more about the process than you did originally.
- Control
The Control process is about:
- Ensuring that you continue to make the changes you planned to make
- Tracking your measurements to ensure that the process continues to deliver the desired outcomes
- Modifying the action plan to fine-tune the results
The Six Sigma process requires a great deal of discipline, but it has been reported by its users to generate spectacular business improvements. Gains in cash flow and profits of over $1 billion a year are claimed to have been achieved by some of the global companies who have adopted it, such as GE, Motorola and 3M.
Click here for free tools and know-how materials from the Mud Valley™ strategy & brand marketing community.
For further information, please contact enquiries@mudvalley.co.uk.
© 2006, Mud Valley ™ brand marketing community.
Related answers
Business scorecards
Change management
Strategy C&E Cause & Effect matrix
Overwhelming case for strategic DIY
Statistics
Get intimate with Mud Valley tools training!
