Define the system and unit of analysis
Specify process, population, product, time window, boundaries and aggregation level. Without a defined unit of analysis, measures and comparisons may represent different phenomena.
TECHNICAL GUIDE · STRUCKEL TOOLS
DAT-01 · Continuous improvement
Data Collection Plan
Defines what to measure, why, where to collect, how to sample, who measures and which criteria make the data fit for decision-making.
dados · CTQ · amostragem · especificaçãoFormulários Struckel
01
MANAGEMENT QUESTION
The problem the tool must make decidable.
Before opening a form, define the question that must be answered. A management tool is useful when it reduces ambiguity, exposes assumptions and makes clear which decision will be made from the evidence produced.
DAT-01What data are needed to answer the project question, what is the operational definition, and how will data be obtained while controlling bias and measurement error?
Continuous improvement requires separating opinion from evidence and turning a broad problem into a sequence that can be observed, tested and revised. The tool creates value when it starts from an explicit current condition, applies criteria before concluding and ends with a verifiable decision — not merely a completed document.
Use it when you need to reduce variation, solve recurring problems, test changes, prioritize causes or preserve learning beyond the meeting in which it was produced.
02
METHOD ARCHITECTURE
Technical construction: from operational definition to evidence.
A robust plan starts by defining the expected outcome and the starting condition. It then organizes actions, owners, deadlines, criteria and resources. Order matters: when action precedes diagnosis, the plan becomes a task list without a theory of results.
Define criteria before evaluation
Translate abstract concepts into operational definitions, scales, classification rules and data sources. Criteria should exist before the result to reduce retrospective interpretation.
Preserve traceability
Link each conclusion to data, observations, assumptions and owners. When a score is used, retain its components; when a hypothesis is used, record how it will be tested.
During review, look for actions without owners, dates disconnected from dependencies, costs without assumptions and goals without measures. A good plan lets you ask “how will we know it worked?” and find the answer in the document.
03
WORKED EXAMPLE
How the method behaves in a real situation.
To measure “service time,” define the exact start/end events, exclusions, automated timestamp source, sample size and stratification by channel/shift. Without an operational definition, analysts may measure different phenomena.
The example is intentionally specific. Numbers illustrate the reasoning and are not universal benchmarks: limits, scales and targets must be defined for the process being analyzed.04
MEASUREMENT AND INTERPRETATION
What to observe so the document does not become mere form filling.
Measures should serve reasoning. Whenever possible combine outcome, process and contextual evidence. For time series, look for patterns and variation rather than only point-in-time before/after comparisons.
01unit of analysis
02operational definition
03sampling plan
04data source
05measurement quality/MSA
Consistency test
Ask whether another person, using the same definitions and evidence, would reach a similar interpretation. If the answer depends heavily on individual judgment, refine the operational definition, scale or data source.
05
LIMITATIONS AND GOVERNANCE
The tool supports decisions; it does not replace judgment, context and validation.
Established method/tool in professional and academic literature; exact implementation should be tailored to context.
More data do not compensate for poor definitions or an inconsistent measurement system. Validate that the data represent the phenomenon before sophisticated analysis.
Define an owner, review frequency, official data source and condition for reopening the decision. Link actions to deadlines and indicators, hypotheses to tests, and risks to treatment and follow-up.
DefinitionEvidenceInterpretationDecisionReview
06
METHODOLOGICAL BASIS
References for deeper study of the method.
The references below support the methodological family or principle used. Struckel’s digital implementation operationalizes application, documentation and traceability; it does not change the conceptual limits of the original sources.
IHI_MFI
Open reference Langley GL, Moen RD, Nolan KM, Nolan TW, Norman CL, Provost LP. The Improvement Guide: A Practical Approach to Enhancing Organizational Performance. 2nd ed. Jossey-Bass; 2009.
MONTGOMERY
Montgomery DC. Introduction to Statistical Quality Control. Wiley.
USE IT IN PRACTICE
Turn the method into a document ready for real work.
Struckel Forms offers a free way to structure this methodology with guided fields, contextual help and print/PDF preparation. Use the tool to organize reasoning; use consulting when the challenge requires design, implementation, facilitation or executive follow-up.