The IASSC Certified Lean Six Sigma Green Belt (ICGB) certifies that you understand Lean Six Sigma process improvement and can support, or lead, smaller improvement projects using data. The whole exam is organised around one method, DMAIC, the five-phase improvement cycle of Define, Measure, Analyze, Improve and Control, and the most reliable way to study is to learn each tool inside the phase it belongs to. The part that catches most people is the statistics in the Measure and Analyze phases, so a good deal of this guide is spent making those concrete rather than scary. This guide is a full self-study course. It walks through each DMAIC phase in depth, teaches the statistics you actually need at Green Belt level, explains where Green Belt stops and Black Belt begins, and turns it all into a study plan. It is original teaching material only. It contains no real or simulated exam questions, and you should study against IASSC’s published Green Belt Body of Knowledge and confirm the current format on the IASSC certification page.
Chapter 1: Exam overview and how to use this guide
What the Green Belt actually measures
The Green Belt measures whether you can apply the Lean Six Sigma method and its everyday statistical tools to reduce defects and variation in a process, typically as a part-time contributor to improvement projects, often under the guidance of a Black Belt. It is a practitioner-level credential: deeper than an awareness course, but not yet the full statistical and leadership depth of a Black Belt. The IASSC route is vendor-neutral and exam-only, meaning there are no prerequisites and no mandatory training, so it suits disciplined self-study against the published body of knowledge. Because several bodies certify Six Sigma, IASSC, ASQ and others, with different exams and requirements, the first thing to settle is that you are preparing for the IASSC body of knowledge if you take this route, and to check which body your employer or industry actually prefers.
The format and what it implies
The exam is 100 closed-book questions in three hours, proctored, a mix of multiple-choice and true/false, with a required score of at least 70%; some forms add up to ten unscored questions. IASSC structures the body of knowledge into the five DMAIC phases as major sections of roughly equal weight, so you can expect a meaningful and fairly balanced share of questions from each phase rather than everything concentrating in one. That balance is a planning gift: it tells you not to neglect Control just because the statistics in Measure and Analyze feel more demanding, because each phase carries real marks. Three hours for 100 questions is comfortable on time, so the challenge is breadth and the statistics, not the clock.
What Lean Six Sigma is, and how to use this course
Lean Six Sigma combines two ideas: Six Sigma, which uses data and statistics to reduce variation and defects so a process performs consistently, and Lean, which removes waste and improves flow so a process delivers value efficiently. DMAIC is the structured project framework that puts both to work. Read the five phase chapters in order, because DMAIC is a sequence and each phase feeds the next, then use the statistics chapter to consolidate the quantitative tools that span Measure and Analyze. Treat the bold terms as tools you must be able to name, place in the right phase, and say when to use. The exam rewards knowing when a tool applies, not just what it is, so throughout, anchor every tool to the problem it solves and the phase it lives in.
Chapter 2: Define
The Define phase sets up the project: what problem are we solving, for whom, how big is it, and what does success look like. Getting Define right prevents the most common cause of failed improvement work, solving the wrong problem or one that does not matter to the business or the customer.
Project selection, the charter, and the problem statement
A Lean Six Sigma project starts by being selected sensibly: it should address a real, measurable problem tied to business priorities, with a process that exists and can be measured, and a scope small enough to finish. The central management document is the project charter, which captures the problem statement, goal statement, scope, business case, team, and timeline, and acts as the agreement between the team and its sponsor (often called a champion). A good problem statement is specific and data-based, describing what is wrong, where and when, and how big the gap is, without guessing at causes or solutions, because naming a cause too early biases the whole project.
Voice of the Customer and critical-to-quality
Improvement is defined from the customer’s perspective, so Define includes capturing the Voice of the Customer (VoC), the expressed needs and expectations of those the process serves. Those needs are then translated into critical-to-quality (CTQ) characteristics: specific, measurable requirements the output must meet. The CTQs are what make “better” concrete and measurable, and they connect directly to how you will measure performance in the next phase. A frequent exam theme is that vague customer wishes must be turned into measurable CTQs before you can manage them.
Lean foundations and high-level mapping
Define also draws on Lean enterprise concepts and a high-level view of the process. The SIPOC diagram (Suppliers, Inputs, Process, Outputs, Customers) is a classic Define tool that frames the process at a glance and clarifies its boundaries and stakeholders. Underpinning everything is the Lean idea of value versus waste: value is what the customer is willing to pay for, and waste is everything else, often remembered through categories of waste such as defects, overproduction, waiting, and unnecessary motion or transport. Recognising the SIPOC and the value-versus-waste distinction at the Define stage is the kind of recall the exam expects, and it sets the scope that the data work then sharpens.
Chapter 3: Measure
The Measure phase establishes the facts: how is the process actually performing today, measured reliably and expressed in numbers. This is where the statistics begin, and it is one of the two phases where most candidates lose marks, so it deserves careful study.
Process definition and data types
Measure starts by defining the process in detail (for example with a detailed process map or value stream map) and deciding what to measure. A foundational distinction is data types: continuous (measured on a scale, such as time, length or weight) versus discrete or attribute data (counts and categories, such as pass/fail or number of defects). This matters because the data type determines which statistics and which later hypothesis tests are valid, so misclassifying data leads to choosing the wrong tool. You should also be comfortable with basic descriptive statistics, measures of central tendency (mean, median, mode) and of spread (range, variance, standard deviation), and with how a distribution’s shape is summarised.
Measurement system analysis
Before trusting any data, Six Sigma checks the measurement itself through measurement system analysis (MSA). The logic is that if your gauge or measurement method is unreliable, the numbers it produces cannot be trusted, and you may “fix” variation that is really just measurement error. MSA assesses qualities such as accuracy (closeness to the true value) and precision, and a common technique is a Gage R&R study, which separates the variation due to the measurement system (repeatability and reproducibility) from the true process variation. The exam expects you to understand why MSA comes before drawing conclusions from data, not as an afterthought.
Process capability
Once you have trustworthy data, you assess process capability: how well the process meets its specification limits. Capability indices such as Cp and Cpk compare the spread (and centring) of the process against the allowed specification range, with Cpk accounting for how off-centre the process is. The famous “six sigma” level corresponds to an extremely capable process with very few defects (the methodology’s name comes from aiming for processes so capable that defects are rare, often expressed as defects per million opportunities, DPMO). For the exam, understand what capability indices tell you, that a higher value means a more capable process, and that capability is meaningful only once the measurement system is validated and the process is stable.
Chapter 4: Analyze
The Analyze phase finds the root causes of the problem using data, separating the vital few causes that really drive the defect or variation from the trivial many that do not. This is the second statistics-heavy phase and, with Measure, where the exam concentrates its quantitative difficulty.
Patterns of variation and root-cause tools
Analyze begins by examining patterns of variation and generating candidate causes. Qualitative root-cause tools sit alongside the statistics: a cause-and-effect (fishbone or Ishikawa) diagram organises possible causes into categories, the 5 Whys drills from a symptom to a deeper cause, and a Pareto chart applies the 80/20 idea to focus on the few categories responsible for most of the problem. These tools structure thinking and point the statistical analysis at the right suspects, and the exam expects you to recognise each and its purpose.
Hypothesis testing
The quantitative heart of Analyze is hypothesis testing, a structured way to decide whether an observed difference or relationship is real or just chance. You state a null hypothesis (typically “no difference” or “no effect”) and an alternative hypothesis, then use a test to produce a p-value, and compare it against a chosen significance level (alpha, often 0.05): a p-value below alpha leads you to reject the null, suggesting the effect is statistically significant. At Green Belt level you should understand this logic and recognise the common tests and when each applies depending on the data: for example tests comparing means (such as t-tests and the idea of ANOVA for comparing several groups) for continuous data, and tests such as chi-square for categorical data. The body of knowledge also includes working with both normal and non-normal data, so you should know that the distribution of the data influences which test is appropriate. The exam rewards choosing the right test for the data situation more than hand-calculating it.
Correlation and regression
Analyze also studies relationships between variables. Correlation measures the strength and direction of a linear relationship between two continuous variables, summarised by a correlation coefficient, with the important caution that correlation does not prove causation. Regression goes a step further by modelling the relationship: simple linear regression fits a line predicting an output (the response) from one input (the predictor), and the Green Belt body of knowledge also includes multiple regression, which models an output from several inputs at once. The boundary worth knowing for the exam is that Green Belt covers correlation and both simple and multiple regression, but it does not include design of experiments (DOE), the planned experimentation that is reserved for the Black Belt. Knowing this line keeps you from over-studying DOE for a Green Belt exam and clarifies exactly how far the Analyze and Improve statistics go.
Chapter 5: Improve and Control
The final two phases turn analysis into action and then make the gains stick. They carry real exam weight, so do not let them fall off the end of your study after the heavier statistics of Measure and Analyze.
Improve: from causes to verified solutions
The Improve phase generates, selects, pilots and implements solutions that address the root causes confirmed in Analyze. The discipline here is that improvements should follow from the evidence, not from opinion: having identified which inputs drive the output, you change those inputs and verify the effect. Green Belt Improve leans on regression (simple and multiple) to understand how inputs relate to the output, alongside Lean improvement techniques to remove waste and improve flow. Crucially, the body of knowledge keeps design of experiments out of scope at Green Belt, so Improve is about understanding relationships and piloting changes rather than running structured factorial experiments. A pilot is emphasised because testing a change on a small scale first reduces the risk of a full rollout that does not work as hoped.
Control: sustaining the gains
The Control phase ensures the improvement lasts after the project team moves on, which is what separates a real, durable gain from a temporary fix. The signature tool is statistical process control (SPC): control charts plot a process measure over time with statistically derived control limits, distinguishing normal common-cause variation from unusual special-cause variation that signals something has changed and needs attention. Control also relies on a control plan, a documented description of how the improved process will be monitored and maintained, including what to measure, how often, and what to do if the process drifts. Lean controls such as standard work, mistake-proofing (poka-yoke) and visual controls reinforce the gains by making it harder for the process to slip back. The exam expects you to know that SPC distinguishes common from special causes and that a control plan and standardisation are what make improvements stick.
Seeing DMAIC as one loop
Pulling the phases together, DMAIC is a single problem-solving loop: Define the problem and its scope from the customer’s view, Measure the current performance with trustworthy data, Analyze the data to find the vital few root causes, Improve by changing those causes and piloting solutions, and Control by monitoring and standardising so the gains hold. Many exam questions are essentially “which phase does this tool or activity belong to?”, so the most valuable mental model is a clear map of each tool to its phase, with the statistics concentrated in Measure and Analyze.
Chapter 6: The statistics you actually need
Because the Measure and Analyze statistics are where most candidates struggle, this chapter consolidates them into a working level of understanding rather than leaving them scattered across phases. The goal is to recognise and reason, not to become a statistician.
Describing data and judging the measurement
Start with the basics that everything builds on: know the data types (continuous versus discrete), the measures of central tendency (mean, median, mode) and spread (range, variance, standard deviation), and the shape of distributions, especially the normal distribution and what it means for a process to be normally distributed. Pair this with the habit of asking whether the data can be trusted at all, which is the job of measurement system analysis and Gage R&R. A surprising number of mistakes, in practice and on the exam, come from analysing numbers without first confirming the data type or the integrity of the measurement system.
Capability, inference, and relationships
On top of the basics sit three families of tools. Capability (Cp, Cpk, DPMO) answers “how well does the process meet specification?”. Inferential statistics (hypothesis testing, p-values, alpha, and choosing the right test for normal versus non-normal and continuous versus categorical data) answers “is this difference or effect real, or just chance?”. Relationship tools (correlation, simple and multiple regression) answer “how are these variables related, and can I model the output from the inputs?”. For each tool, the exam cares most that you can pick the appropriate one for the situation and interpret what its result means, for example that a low p-value suggests a real effect, that a Cpk below a target signals an incapable or off-centre process, or that correlation alone does not establish cause.
Where Green Belt draws the line
A clear sense of scope keeps your preparation efficient. At Green Belt you are expected to handle descriptive statistics, MSA, capability, hypothesis testing (including awareness of non-normal data), correlation, and simple and multiple regression. You are not expected to cover design of experiments (DOE), full or fractional factorial designs, which belong to the Black Belt. Studying to this boundary, deep enough on the included tools to choose and interpret them, without straying into DOE, matches what the IASSC Green Belt body of knowledge actually tests and is the most efficient use of your time.
Chapter 7: Study plan and timeline
With the phases and the statistics mapped, the plan is about pacing the content so the Measure and Analyze statistics get enough time without letting Define, Improve and Control slide, since each phase carries roughly equal weight.
A phase-by-phase plan
Most candidates need roughly 50 to 80 hours over six to ten weeks, with people new to statistics at the higher end. A workable shape follows DMAIC: week one on Define (project selection, charter, VoC and CTQs, SIPOC, value versus waste); weeks two to three on Measure (process definition, data types, descriptive statistics, MSA and capability), giving the statistics real time; weeks four to five on Analyze (root-cause tools, hypothesis testing, correlation and regression), the other statistics-heavy phase; week six on Improve and Control (piloting, SPC, control plans and Lean controls); and a final week or two on full-length timed practice and weak-area review. Study each tool in the context of its phase so you learn when to use it, and treat the Measure and Analyze statistics as the part to over-prepare rather than hope to recognise. To turn this into dated weeks for your own start date, use the free study-plan generator.
Study against the IASSC body of knowledge
Because IASSC publishes its Green Belt Body of Knowledge in full and builds the exam directly from it, that document is your definitive scope: use it as a checklist and make sure you can speak to every listed topic at the level described, no more and no less. This is also what keeps you from drifting into Black Belt material such as DOE. If you expect to lead larger improvement projects later, it is worth knowing how the next level differs; the deeper statistics and the leadership step up are covered in the Lean Six Sigma Black Belt study guide, and the wider exam directory can help you see where process improvement sits alongside delivery credentials.
Chapter 8: Final preparation, exam day, and format
Final preparation
In the last week or two, shift from learning to full-length, timed practice against the published body of knowledge, treating each session as a diagnosis. Break your results down by DMAIC phase and look hardest at Measure and Analyze, since the statistics there are where marks most often leak, but confirm you are also solid on Define, Improve and Control given their roughly equal weight. Review the reasoning behind every miss: was it a gap in knowing a tool, or in knowing which tool fits the data situation? Aim to be scoring comfortably above the 70% pass mark on fresh questions before you book.
Confirm scope and the certifying body
Two checks protect easy marks. First, make sure you have been studying the IASSC body of knowledge specifically, since ASQ and other bodies differ, and confirm with your employer or industry which certification they expect. Second, hold the Green Belt scope boundary firmly in mind, especially that design of experiments is not included at this level, so you neither waste time on DOE nor get rattled if a question’s framing tempts you toward it. Confirm the current exam format and any administration details on the IASSC certification page, since logistics such as proctoring can change.
Exam day and format
On the day, the exam is 100 closed-book questions in three hours, proctored, requiring at least 70%, with some forms adding up to ten unscored questions. With around 108 seconds per question, time is comfortable, so the discipline is accuracy and tool selection: read each question for the DMAIC phase and the data situation it implies, choose the tool that fits, and interpret results rather than over-calculating. Flag anything uncertain and return to it rather than stalling. Prepared with each tool mapped to its phase, the Measure and Analyze statistics genuinely understood, and the Green Belt scope boundary clear, the exam rewards exactly the structured, data-based thinking the method is built on.