DMAIC| Measure Stage
What is it… DMAIC Measure Stage, in this second phase of the DMAIC six-sigma improvement cycle we understand where we are today. A cornerstone of a DMAIC project is they are data bound. That doesn’t mean to say we have all the data in the world at our disposal, rather we utilise the data we have and seek out data for our purposes. Critically, the measure stage provides the baseline for our project, thereby placing a stake in the ground for the projects staring point . This is referred back to in order see what difference the improvement project has made. The measure phase is often detailed and requires significant thought and investment – however the investment will pay off here and throughout the six-sigma project.
How long should the Measure Stage take?
As a guide the Measure Stage should last between 2-4 weeks. If you have an overly dispersed team or restrictions on freeing up operational colleagues then this may take a little longer.
What are the key steps…
Step 1 | Process Measurement – This is where the project begins to build it’s hypothesis. In doing so it structures the data collection on step 2. Here we need to consider how we will measure the problem.
There are 3 key areas to ensure robust process measurement for the measure stage of DMAIC:
- The measure set based on KPI’s developed by what the customer deems as Critical to Quality (CTQ).
- Lean measurements
- Process Cycle Efficiency (PCE)
- Average Lead Time
- Types of Data (there are three recognised data types with Contextual added as the fourth data type)
- Continuous Data – For when you are looking to ‘measure‘ something and by nature then contains numeric measurement data.
- Count Data – For when you are looking to ‘count‘ something.
- Attribute Data – For when you are looking to ‘classify‘ something and is binary in nature.
- Contextual (the 4th data type) –This data type provides the much needed context to a set of numbers and avoids management by spreadsheet. An example could be a lower sales performance in May contributed to by a system outage not recorded in the data or a short term reduction in FTE. However, resist the urge to ignore what quality data is telling you if others are trying to bend the view with a non-fact based approach.
These steps help us to be focused and channel our efforts. Just as you can have issues with organisations or projects having no data the same can be said when an organisation or projects has too much data. Getting this right will significantly reduce the effort and time in collection whilst being more clinical with the data produced.
Step 2 | Collect Process Data – Now we set about collecting the data we want. For this we need to consider:
- The Data Collection Method, covering:
- What type of data is required?
- What are the critical factors of the process based on the KPIs?
- Where can the data be collected from? Systems – which sources; The Operation
- Who will collect the data?
- Spread of data required? Is the process open to seasonal or intra-day, week, month fluctuations?
In addition to the above we like to capture 3 things:
- What is Readily available;
- What Could be available (we don’t have but can collect); and
- What is not readily available (so we will need a plan B)
These 3 things help with the next sub-step of planning the data collection
- A Data Collection Plan, covering:
- Who will collect the data?
- When will the data be collected?
- Data Sampling, covering:
- Sample sizes required – and the level of confidence we are willing to accept
- Minimum sample size
You can use our Sampling Calculator to understand the sample size required for your activity.
A general rule of thumb is to gather data points across 12-13 weeks (13 weeks as it makes a full quarter). This avoids overreacting to singular data points. However, this will need to be tailored to your project. Items to consider are; seasonality; peaks and troughs; exceptional activity (outliers) all of which impact teh validity of your baseline.
Often we can over complicate data collection. At InvisibleConsultant we believe in data led decisions. However, not becoming data blind. We need to collect what is reasonable for the activity we are setting out. This often comes with experience (in other words from learning from past mistakes). Think about the term analysis paralysis – that is something we want to avoid; the aim is to be intelligent rich, not data rich and intelligence poor.
Step 3| Data Quality Check – An often missed step within the data collection of any business improvement step – or indeed any operational delivery. DMAIC qualifies this as Measurement System Analysis (MSA). Which is simply ensuring for each data point the numbers are accurate or an accurate representation of the situation.
There are many organisations and projects where this check is never completed. Even expert Business Intelligence Teams and consultancies are guilty of providing data as the de facto position without validation. If all data is taken at face value the consequences can be as severe as or worse than taking a decision ignoring data.
A MSA drill down can help to test the validity of the data. Gauge R&R is a six-sigma technique not covered here which may be applicable to your industry. This technique is most commonly used in technical environments such as car manufacturing.
Step 4| Process Behaviour – Having reviewed the plan, collected and assessed the data we will now begin to have an understanding of the way of our process behaves. In this section we conduct 1st Pass Analysis where we see what has happened in the past with our process. This is not the analysis stage, that comes after the measure stage, however we want a baseline and to understand historical movements.
Step 5| Process Baseline – Now we have completed the above steps we will have a baseline for our process. This baseline will be referred back to, to assess the improvement the project has achieved.
Step 6| Gain Approval – to move onto next Stage
At the end of the Measure Stage the following will have been completed or identified:
|Focused Problem Statement||Process Mapping||Gemba Walks||Baseline Process Capability||KPI’s|
|Hypothesis Developed||Data Collection||1st Pass Analysis||VoC specification limits set||Calculated Sigma Levels|
Key Deliverables of the Measure Stage
- Focused Problem Statement
- Detailed Process Map
- Baseline Data and Process – Six-Sigma Levels
- Hypothesis Developed
When assessing Process Capability be mindful that internal targets (specifications) may not represent the Voice of the Customer. For instance a contact centre may strive to answer a call within 40 seconds whereas the customer may be comfortable with 120 seconds or require response within 20 seconds. Understanding which it is will have a material impact on the projects direction.
DMAIC Measure Stage FAQ’S
What is '1st Pass Analysis'?
This activity can help to show a historical view of the process and highlight any positive or negative trends.
What is an hypothesis?
An hypothesis is a theory offered to explain a situation or a possible situation. At the initial stage it is just that; a theory, which is not proven or dis-proven.
The principle use of hypothesis in Six-Sigma is for structured data gathering and to align stakeholders on measurement/analysis and the results.
What is Process Capability?
Process Capability is for the set of KPI’s which measure the processes performance against the VoC (Voice of the Customer) specifications.
What is Takt time?
Takt Time is a lean measurement and is referred to as the drumbeat of a process, as it represents the speed of customer demand. It is a load balancing mechanism and is calculated on a process accordingly:
Takt Time = Available Work Time / Customer Demand
The full | Six-Sigma A-Z Glossary
Join Our email list
Join the InvisibleConsultant email list today for support direct to your inbox. Don’t delay Level up Your Capability today.
Though we hope you will find our material helpful, reliance on this material and any related content is at your sole risk. InvisibleConsultant.co.uk assumes no responsibility for any errors or damages arising.