Sensitivity Analysis Tables: Presenting Scenarios Clearly to Executives

A sensitivity analysis table is among the most analytically valuable—and most frequently misused—visual elements in consulting deliverables. Done well, it shows executives exactly how a key outcome changes as critical assumptions vary, giving them the tools to make a decision under uncertainty. Done poorly, it produces a grid of numbers that obscures rather than illuminates the finding.

The difference between a useful sensitivity table and a confusing one is almost entirely about design: how the table is structured, how scenarios are labeled, which assumption dimensions are highlighted, and how the table is connected to the recommendation.

---

What a Sensitivity Analysis Actually Communicates

Before the design, understand what the table is communicating.

A sensitivity analysis answers the question: "How much does the outcome change if our key assumptions are wrong, and by how much?"

It serves two executive audiences:

The skeptic: The executive who doubts the base case assumption and wants to know "what happens if you're wrong about growth rate?" The sensitivity table gives them a defensible answer without requiring a debate about the single-point estimate.

The decision-maker under uncertainty: The executive who knows the assumptions are uncertain and needs to understand the range of outcomes before committing. The sensitivity table shows whether the decision is robust (good outcome across a wide range of assumptions) or fragile (good outcome only under favorable assumptions).

A well-designed sensitivity table turns both audiences from resistors into informed decision-makers.

---

The One-Dimensional Sensitivity Table

The simplest sensitivity table shows how a single output changes as a single input varies.

Structure:

• One column for the input variable (the assumption being varied)
• One column for the output variable (the outcome being measured)
• 5–7 rows representing a range of assumption values
• The base case row highlighted

Example:

| Revenue Growth Rate | NPV (€M) | |---|---| | 5% | 42 | | 8% | 61 | | 10% (base case) | 75 | | 12% | 89 | | 15% | 110 |

Design principles:

• The base case row should be clearly differentiated: bold text, shaded background, or a horizontal line above and below
• Include a "base case" label in the row rather than relying on visual formatting alone
• The input variable range should span plausible scenarios, not extreme outliers (a 50% revenue growth scenario is neither plausible nor useful)
• The output range should tell the decision-making story: does the NPV stay positive across all scenarios? At what point does it turn negative?

---

The Two-Dimensional Sensitivity Table

The two-dimensional table (often called a "data table" or "tornado table") shows how an output changes across two input variables simultaneously.

Structure:

• One input variable drives the rows
• A second input variable drives the columns
• Each cell shows the output value for that combination of inputs
• The base case cell highlighted

Example (NPV in €M):

| Revenue Growth → | 5% | 8% | 10% | 12% | 15% | |---|---|---|---|---|---| | Margin (↓) | | | | | | | 12% | 28 | 44 | 55 | 66 | 84 | | 15% (base) | 42 | 61 | 75 | 89 | 110 | | 18% | 57 | 79 | 96 | 113 | 138 | | 20% | 65 | 90 | 110 | 130 | 159 |

Design principles:

• The base case row and column should be visually differentiated (different background, bold borders, or both)
• Add a color gradient to cells: green for favorable outcomes, white for neutral, red for unfavorable. This immediately shows the executive which combinations produce acceptable vs. unacceptable results.
• Include a "Decision threshold" annotation—a line or border showing where the NPV is zero, or where the metric passes the client's internal hurdle rate

The color gradient approach in practice: Setting conditional formatting to display green (favorable) through red (unfavorable) allows an executive to visually identify the range of assumption combinations that support the recommendation versus those that don't. This is significantly more useful than requiring the reader to parse 20 individual numbers.

---

The Scenario Table: Named Scenarios

Sensitivity tables show continuous variation in individual assumptions. Scenario tables are different: they show a small number of defined scenarios, each representing a coherent set of assumption combinations.

When to use scenarios instead of sensitivity tables:

• When the key uncertainty is qualitative, not quantitative (e.g., "does the competitor respond aggressively?")
• When multiple assumptions move together (in a downside scenario, both revenue and margin are lower; treating them independently would overstate the range)
• When you want to name the scenarios to make them more memorable and decision-relevant ("Base," "Upside," "Downside," "Stress")

Scenario table structure:

| | Downside | Base | Upside | |---|---|---|---| | Revenue growth | 5% | 10% | 18% | | EBITDA margin | 12% | 15% | 20% | | Market entry cost | €30M | €25M | €20M | | NPV | €42M | €75M | €132M | | IRR | 12% | 18% | 28% | | Payback period | 4.2 years | 3.1 years | 2.3 years |

Design principles:

• Scenario names should be descriptive and non-judgmental. "Base," "Upside," "Downside" are standard. "Management" vs. "Conservative" names can create awkward implications about whose assumptions are being accepted.
• Column order: Downside, Base, Upside from left to right (most readers process severity left-to-right)
• The Base column should be visually differentiated as the primary reference
• Financial outputs should be grouped together and separated from assumption inputs by a horizontal line

---

Tornado Charts: Visualizing Input Sensitivity

A tornado chart is a horizontal bar chart that shows how sensitive an output is to each of several input assumptions, ranked by impact. The most sensitive assumption is at the top; the least sensitive is at the bottom—creating the "tornado" visual shape.

When to use a tornado chart:

• When you have five or more input assumptions and want to show which matter most
• When the primary message is "sensitivity is concentrated in a small number of assumptions"

Structure:

• Each row represents one input assumption
• The bar extends left (unfavorable impact) and right (favorable impact) from the base case value
• Rows sorted by total bar width (most sensitive assumptions at top)

Example message a tornado chart communicates: "Of the seven assumptions in our model, 80% of the NPV variation is driven by revenue growth rate and margin—the other five assumptions matter much less."

This framing helps executives focus their challenge and due diligence on the assumptions that actually change the recommendation, rather than questioning every assumption uniformly.

---

Connecting Sensitivity Analysis to the Recommendation

A common consulting mistake: presenting a sensitivity analysis as a standalone "here are all the scenarios" exhibit without connecting it to the recommendation.

Executives looking at a sensitivity table need guidance on what it means for their decision. The slide containing the sensitivity table should answer three questions:

1. What is the base case assumption? (Label the base case clearly)
2. What is the outcome range? ("Under all plausible assumptions, NPV ranges from €42M to €132M—positive in all scenarios")
3. What does this mean for the decision? ("The recommendation is robust: even in our downside scenario, NPV exceeds the investment threshold by €17M")

The sensitivity table is evidence for a conclusion, not a standalone exhibit. Write the slide title as the conclusion: "The Recommendation Is Positive Across All Plausible Scenarios" rather than "Sensitivity Analysis."

---

The Downside Case Discipline

In consulting, there is a professional standard around downside case construction that matters: downside cases should be genuinely realistic, not strawmen.

A downside case built around implausible low values ("what if revenue is 40% below base?") undermines the analysis. If the executive knows the downside is constructed to make the recommendation look good even under bad conditions, the sensitivity analysis doesn't build confidence—it destroys it.

The credibility test: Can the engagement manager describe a specific, plausible set of circumstances that would produce the downside scenario? If yes, the downside is credible. If the downside requires a combination of simultaneously adverse events that almost never occur together, it's not a useful downside.

This is a judgment call, and senior partners apply it rigorously in review. A sensitivity analysis that the partner thinks is "set up to win" will receive as much skepticism as one that shows the recommendation breaks under moderate adverse conditions.

---

Formatting Checklist for Sensitivity Tables

Before finalizing any sensitivity or scenario table:

• [ ] Base case clearly identified with visual differentiation (bold, shading, border)
• [ ] All cells correctly calculated (verify the math)
• [ ] Input variable range is plausible, not extreme
• [ ] Output values formatted consistently (same decimal places, same currency notation)
• [ ] Decision threshold visible if applicable (zero line, hurdle rate)
• [ ] Slide title states the conclusion of the sensitivity analysis, not just its topic
• [ ] Color coding (if used) follows firm's positive/negative convention

---

Related Resources

• How to Show Complex Financial Models Simply in Client Presentations
• The 10 Chart Types Every Management Consultant Needs to Master
• How to Build a Consulting-Grade Waterfall Chart in PowerPoint
• The Executive Summary Slide: How Top Consultants Lead With the Answer