Data Collection Plan

The Data Measurement Plan:- Use full dates (month, date, year)
- Data Collection Plans are used in the Measure Phase- Use explanatory title
of Lean Six Sigma projects.- Consider lowest common denominator on metric
For each performance measure (Y), update a data(Minutes vs. Hours; Inches vs. Feet)
collection plan to:- Test and validate your design (try it out)
- Include MSA (Measurement System Analysis)- Don't change form once you've started, or you'll be
measure plan (Gantt chart, MS-Project plan is Optional)"starting over"!
- Add Financial measure plan if separate fromAs you set up Check Sheets:
performance Y- Prepare a spreadsheet to compile the data.
- Add any Time Study or other data collection plans- Think about how you'll do the compiling (and who'll do
for Value Stream Mapit).
- Sample Size Calculation- Consider what sorting, graphing or other reports you'll
- Input, Process and Output Metrics (confirm that thewant to create.
right process metrics have been chosen and logical- Continuous or Discrete data?
trade-offs have been made in determining what to- Adequate level of Discrimination and Accuracy?
measure.)- Adjust check sheet as needed to ensure usable
- Is this project driven by Customer specifications? Ifdata later, but don't make data harder to collect.
so, how do you know that the specifications satisfyWho Will Collect the Data?
customer critical requirements?Considerations:
- What was the process for determining the metrics in- Familiarity with the process
the Data Measurement Plan?- Availability/impact on job
- What trade-off's were made is determining the finalRule of Thumb - If it takes someone more than 15
set of metrics for which to gather data?minutes per day it isn't likely to be done
The extent to which these have been addressed or- Potential Bias
executed form the basis of evaluating whether or not- Will finding "defects" be considered risky or a
to allow the project to proceed with the actual data"negative"?
collection.- Benefits of Data Collection
Key Question: Does the data currently exist?Will data collection benefit the collector? Be Sure They
Existing Data...
- Taking advantage of archived data or current- Give input on the check sheet design
measures to learn about the Output, Process or Input- Understand operational definitions (!)
- This is preferred when the data is in a form we can- Understand how data will be tabulated
use and the Measurement System is valid (a big- Helps them see the consequences of changing
assumption and concern)- Have been trained and allowed to practice
New Data- Have knowledge and are unbiased
- Capturing and recording observations we have notNarrow Potential Key Process Input Variables (KPIVs):
or don't normally capture- Have the potential root causes been narrowed?
- May involve looking at the same "stuff," but with new- Was a Cause and Effect (C&E) Matrix used? If
Operational Definitionsso, what were the results?
- This is preferred when the data it is readily and- How were the KPOVs (Key Process Output
quickly collectable (it has less concerns withVariables) rated?
measurement problems)- Did people who operate the process, technical
Existing vs. New Considerationsexperts, and supervisors collaborate to produce the
- Is existing or "historical" data adequate?C&E Analysis?
- Meet the Operational Definition?- Have you characterized the variables (controllable,
- Truly representative of the process, group?uncontrolled [noise], etc)?
- Contain enough data to be analyzed?- Was a Pareto Chart used to select potential Key
- Gathered with a capable Measurement System?Process Input Variables (KPIVs) from the C&E
- Cost of gathering new dataMatrix?
- Time required to gather new data- How many KPIVs do you have at the beginning and
The trade-offs made here, i.e. should the time andend of C&E Matrix?
effort be taken to gather new data, or only work with- Are there any potential KPIVs which need immediate
what we have, are significant and can have aBaseline capability and MSA?
dramatic impact on the project success- Are these potential KPIVs monitored in the
Check Sheets:workplace?
- The workhorse of data collection- Which process steps stand out as especially
- Enhance ease of collectionsignificant in the C&E Matrix?
- Faster capture- Is there any process step that the team feel can be
- Consistent data from different peopleeliminated or combined?
- Quicker to compile dataKey Steps:
- Capture essential descriptors of data "Stratification- Fill in the Output measure Y.
factors"- Fill in the key stratification questions you have about
- Need to be designed for each jobthe process in relationship to the Y.
How will Data Be Collected:- List out all the levels and ways you can look at the
1. Select specific data & factors to be includeddata in order to determine specific areas of concern.
2. Determine time period to be covered by the form- Create specific measurements for each subgroup or
(Day, Week, Shift, Quarter, etc.)stratification factor.
3. Construct form- Review each of the measurements (include the Y
- Be sure to include: (clear labels ; enough room; spacemeasure) and determine whether or not current data
for notes; test the form)exists.
Tips:- Discuss with the team whether or not these
- Include name of collector(s) (first & last)measurements will help to predict the output Y, if not,
- Reason/comment columns should be clear andthink of where to apply the measures so that they will
concisehelp you to predict Y.