12.5 Communicating Statistical Results to Business Audiences

12.5 Communicating Statistical Results to Business Audiences

Quick Summary: This is the final chapter of the CSSC Green Belt Body of Knowledge, and it deliberately steps back from the mathematics of Units 4 and 5. The manual's own framing: the more you learn about statistical analysis, the easier it is to get caught up in the challenge of the analysis and forget that the business is relying on you for answers about its own challenges. Every tollgate, every sign-off, and every funded improvement in a DMAIC project depends on somebody non-technical being persuaded. That makes this chapter a skills chapter, not an afterthought.

---

The Four Obstacles

Obstacle 1: Most people do not care about the method

CSSC states it bluntly. Unless your sponsor is a Six Sigma expert or your leadership committee is well-versed in Six Sigma methods, the people you are presenting to do not care which hypothesis test you used, or what your Chi-Square statistic or p-value was.

Read that precisely, because the exam does. It does not say business leaders are indifferent to your findings. They care a great deal about the conclusions and results of your analysis, and about the results of the project as a whole. What they do not care about is the machinery. The presenter's task is to present the information in a way that ties in with what they care about and does not rely solely on statistical concepts and language.

Obstacle 2: The audience lacks your statistical background

General business staff and leaders are unlikely to understand statistical analysis the way a Six Sigma expert does. That leaves the presenter delivering data whose backing may not be understood. You must be able to explain why the team chose to move in a certain direction or how you know a process has improved — because not everyone will simply take your word that the p-value was appropriate for the decision you made.

Obstacle 3: Getting derailed by the statistics

This happens two ways, and they need opposite responses.

A statistically literate audience member hijacks the discussion. A sponsor or Black Belt in the room starts asking statistical rather than business questions: Why did you use the 1-Proportion test? How did you choose the sample size? Do you really think a beta of 0.2 is sufficient, or should you have gone with 0.1?

CSSC's judgment on these is nuanced. Some are valid — "how did you choose the sample size?" may be worth covering so the audience understands what assumptions the team made and what risks it accepted. But once the analysis is done and decisions have been made on the basis of it, whether the power was set correctly is a moot point. The published preventive measure is structural: keep a Six Sigma-knowledgeable sponsor in the loop throughout the project, so they can give that input in real time rather than at the tollgate.

The audience does not understand the statistics, so the presentation becomes a statistics lesson. The remedy is to walk a line: educate on what is necessary to follow the Six Sigma process, without delivering an in-depth discussion of statistical analysis. CSSC's calibration is concrete — you probably should ensure the audience understands what standard deviation is and why it matters; you probably do not need to explain all the types of hypothesis tests, or in some cases even mention that you used one.

Obstacle 4: Too much information

The temptation is familiar: you have a box plot, a bar graph, a pie chart, a hypothesis test, and several probability-distribution calculations. They all look good and each says something slightly different. Surely they should all go in?

Wrong. Too many ways of illustrating the data overwhelms the audience. Worse, most people in the business will not see the nuances a trained Six Sigma expert sees — so from their perspective you are showing the same data over and over. It is not an effective use of presentation space or meeting time.

---

Why Statistics Stay In

Given all four obstacles, why not drop the statistics entirely? Because statistics are what give your conclusions and results credibility. Whether you are presenting at a Measure or Analyze tollgate or seeking final approval on improvements, the numbers are the warrant for the recommendation. The answer is translation, not omission.

---

The Problem-Statement Test

CSSC gives one decision rule for what goes in the deck, and it is worth committing to memory because it doubles as an exam answer.

Because a real Six Sigma team is already rooted in business terms, the problem statement — which combines statistical concepts with business goals and needs — is a good reference point for anyone presenting throughout and after a DMAIC project. So:

If you are not sure whether something should be included, ask: does it relate to the problem statement, or to a solution for the problem statement? If not, it probably does not need to be communicated in a general business environment.

The manual pairs this with a warning about a specific professional failure mode: as you work with data, you may be surprised how much your interest in the numbers grows. That interest leads to valuable discoveries — but you must avoid the temptation to include data or conclusions in a presentation simply because you find them interesting or "cool."

---

Tailoring to the Audience

The invested, statistically literate audience

Consider a Black Belt presenting at an Improve tollgate to a leadership team that includes the sponsor and two other Black Belts, perhaps a Master Black Belt. This audience is both invested and capable of following the statistical reasoning.

That does not mean including all your data — your audience is also busy. But it does mean you should anticipate a few methodology questions and can incorporate technical terms and results without stopping to explain them.

The appendix technique. For this audience, CSSC's specific recommendation is to build a presentation appendix holding your analysis, data tables, and even specific Minitab output, where it can be displayed quickly if a question calls for it. The main deck stays business-facing; the evidence stays one click away. This matters most in high-cost or high-risk situations, where leadership is more likely to want the background analysis because they need to fully understand all assumptions and risks before deciding on the recommended improvements.

The uninvested or non-technical audience

Presenting to general management, or explaining changes to frontline staff in a department, calls for a different build: heavier on what changes and why it helps, far lighter on method.

---

The Five Presentation Tips

CSSC lists five techniques for creating viable business presentations from statistical work.

1. Understand the target audience. Tailor the information to the audience's level of concern and knowledge, as above.
2. Tell a story with text and images. A narrative structure — the problem, what the data showed, what the team did, what changed — carries a business audience through material that a table of results cannot.
3. Be clear and concise. Keep things short and simple; this is the direct countermeasure to Obstacle 4.
4. Do not misuse your tools. Avoid misuse of technology, presentation, or statistical tools — including chart types that distort the data and slide effects that distract from it.
5. Do not let the presentation drive the project. The project's purpose is improving a business process. If work is being done, or evidence framed, to make a slide land rather than to answer the business question, the tail is wagging the dog.

---

Realistic Exam Scenario

A Green Belt has finished the Control phase of a warehouse picking project. Cpk rose from 0.87 to 1.42, a two-sample proportion test confirmed the mis-pick reduction at p = 0.003, and the I-MR chart has been stable for eleven weeks. She must present twice: first to an executive steering committee approving a rollout to four more sites, then to the warehouse floor staff.

Executive deck. Lead with business impact — mis-picks down 61%, an annualized $340,000 in avoided rework and expedited freight, sustained over eleven weeks. Keep the control chart, because it visually proves sustainment, and keep one sentence establishing that the improvement was statistically confirmed rather than luck. Push the proportion test output, the Gage R&R study, and the capability calculations into the appendix — this is a high-cost decision, so leadership may well ask for the assumptions, and the appendix has them ready.

Floor deck. No p-values, no Cpk. The story is: here is the error that used to happen, here is what changed about the pick sequence, here is how we will know if it drifts back, and here is who to tell. Standard deviation might earn a sentence if the monitoring chart is going up on the wall; the hypothesis test will not.

Same project, same evidence, two builds. Applying the problem-statement test to each is what produces the difference.

---

Common Exam Traps

• Trap 1: Reading "most people don't care" as "leaders don't care about results." They care intensely about conclusions, results, and project outcomes. They do not care which test produced them.
• Trap 2: Removing statistics to simplify. Statistics are what give conclusions credibility. Translate, do not delete.
• Trap 3: Answering every methodology question in the main session. Some are valid — sample size and accepted risk may be worth covering. Debating whether power should have been 0.9 after decisions are already made is moot. The published fix is keeping a knowledgeable sponsor engaged throughout.
• Trap 4: Showing every chart you built. Multiple views of the same data read as repetition to a business audience and waste meeting time.
• Trap 5: Skipping the appendix for a high-risk decision. High-cost, high-risk proposals are exactly where leadership wants the background analysis available on demand.
• Trap 6: Including a finding because it is interesting. The test is relevance to the problem statement or its solution — not analyst enthusiasm.