What-If Analysis Results

What-If Analysis results are derived from two types of calculations: One-Way and Multi-Way analyses. Each approach is designed to assess the influence of input changes on output values in distinct ways, offering different insights into model behavior.

These two analysis types complement each other: One-Way analysis pinpoints key drivers, while Multi-Way analysis explores synergies and interactions between variables.

Results from a One-Way What-If Analysis

One-Way What-If Analysis evaluates the impact of changing each input individually while keeping all other inputs constant at their base values. This method is quick and effective for identifying the most influential variables in a model. Each input, defined using functions like RiskVary or RiskAutoVary, is varied across a defined range (e.g., -10% to +10%), and the resulting changes in selected outputs are recorded. What-If Analysis then ranks the inputs by their impact on each output, helping users understand which variables matter most. Graphs such as Tornado, Spider, and Sensitivity charts visually communicate these effects.

 

Data Window

The Input section on the left side of the One-Way What-If Analysis window lists all the input variables that have been varied during the analysis. For each input, the table displays key information about how the variable was defined, its location in the spreadsheet, and the specific values tested across a series of steps.

Each row in this section corresponds to a specific step in the variation of an input. It shows:

  • The function type used (e.g., RiskAutoVary, RiskPert) to define how the input varies,

  • The cell location of the input in the worksheet,

  • The exact value evaluated at that step,

  • And the absolute and percentage change from the base case.

This detailed breakdown helps users trace how inputs are structured and monitor the direction and magnitude of the variations being assessed.

 

The Output section on the right side of the One-Way What-If Analysis window displays the results of recalculating the model outputs for each step of every input variation. This section allows you to observe how changes to a single input affect the selected outputs in your model.

 

Each output is represented with the following three columns:

  • Value: This is the actual output value calculated when the input variable is set to the specific value shown in the Step. It reflects how the output responds to a single change in that input while all others remain at their base values.

  • Change: Displays the absolute difference between the recalculated output value and the base case output value (i.e., when all inputs are at their original settings). This shows the dollar or unit impact caused by the variation in the input.

  • Change (%): Shows the percent change from the base case value for the output. This is helpful for comparing the relative sensitivity of outputs to various inputs, especially when outputs have different units or scales.

Together, these columns provide a clear, step-by-step view of how sensitive each output is to each input, helping prioritize variables for model refinement or decision-making.

Reports

  • Summary Report: The What-If Analysis Summary Report provides a comprehensive overview of the analysis settings and results, making it easy to understand the scope and impact of the sensitivity analysis. At the top of the report, a summary table presents key configuration details: the number of analyses runs, total recalculations, the number of inputs and outputs evaluated, and how many inputs were identified manually versus automatically. It also includes constraints such as the maximum number of inputs and the threshold for reporting significant changes, offering a quick reference for understanding the context and limits of the analysis.

     

    Beneath the summary table, the report includes a detailed results section formatted similarly to the What-If Data window. For each input, the table shows all test values across the defined step range and the corresponding effects on each output, capturing the absolute and percentage changes in cost, revenue, and profit. This allows users to trace precisely how each input influences model behavior and supports decisions about which variables are most critical to performance.

     

  • Output Detail Report: Provides an in-depth view of how each input variable affects a single output in a One-Way What-If Analysis. A separate worksheet is generated for each output selected in the analysis, allowing for focused evaluation of that output's sensitivity to changes in the model’s inputs. Within each worksheet, inputs are ranked by percent change in their impact on the output. The report table includes the input's name, the workbook and cell it belongs to, the type of function used (e.g., RiskVary, RiskAutoVary), and detailed statistics for the step that caused the most significant output change. This includes the input value at that step, the resulting output value, and their respective percentage changes.

     

  • Tornado Report: Offers a visual and numerical summary of how input variations affect a specific output, helping users quickly identify which variables have the greatest impact. Each output selected in the What-If Analysis generates its own worksheet, beginning with a Tornado Graph at the top. This chart displays inputs on the Y-axis and the range of resulting output values on the X-axis. Each bar represents the full range of impact caused by varying a single input, from its lowest tested value to its highest. The inputs are automatically sorted, with the longest bars (i.e., those causing the most meaningful change in the output) appearing at the top. This makes it easy to spot the most influential drivers in the model visually.

     

    Below the chart, a supporting data table presents the exact numeric values used to build the tornado graph. This table includes columns for each input’s name, workbook, cell reference, function type, and the values and percent changes for both the minimum and maximum steps of the input variation. This allows users not only to interpret the visual graph but also to see the underlying data that justifies the input ranking.

  • Spider Report: Offers a detailed view of how changes in multiple inputs affect a single output, using a line graph to illustrate trends and potential nonlinear behavior. A separate worksheet is created for each output, starting with the Spider Graph, which plots the input variation (as % of base case) on the X-axis and the corresponding output variation (as % change) on the Y-axis. Each line represents an input variable. The slope of a line reflects how sensitive the output is to that input: steeper slopes indicate greater sensitivity. Unlike the tornado graph, which focuses on extremes, the spider graph shows how the output behaves across the full range of each input, helping detect nonlinear relationships or threshold effects.

    Below the graph, a data table provides the precise values used to construct each line. For each input, it records the cell and step used, the run number (for multiple runs), the input value, its change, and % change from base, and the corresponding output value, its change, and % change from base.

    This table enables users to analyze exact numeric relationships and confirm patterns observed in the chart. Because spider graphs can become cluttered, it is best to use them with a limited number of high-impact inputs.

     

  • Sensitivity Report: Provides a focused view of the relationship between a single input and a single output. For each output selected, a separate worksheet is generated containing a series of sensitivity graphs, one per input. Each graph plots the input values on the X-axis against the resulting output values on the Y-axis. This straightforward visualization lets users see how output values change in response to variations in a specific input, helping identify patterns such as linear trends, thresholds, or inflection points.

    Below the charts is a data table. This table details the input value, absolute change, and percent change from its base value, as well as the corresponding output value, with its absolute and percent change from the base value.

 

Results from a Multi-Way What-If Analysis

Multi-Way What-If Analysis, on the other hand, explores combinations of input values by varying multiple inputs simultaneously. Inputs designated with RiskVaryMulti, RiskVaryMultiTable or any @RISK function with the Multi parameter are tested in groups (defined by the Group Size setting), What-If Analysis evaluates every possible combination of values across the specified ranges. This approach provides insight into interaction effects and the compounded impact of variables on outputs. Results include ranked combinations, tables of input-output results, and Multi-Way tornado graphs—offering a deeper analysis of when input interdependence is relevant.

 

Data Multi-Way Window

The Multi-Way Data Window displays the full set of results from a Multi-Way What-If Analysis, ranking every tested combination of input values based on their impact on a selected output. At the top of the window, users can choose the output of interest, and the table below dynamically updates to show all corresponding input combinations. Each row represents a unique scenario in which inputs labeled with Multi (like RiskVaryMulti, RiskVaryMultiTable, or @RISK like RiskPertMulti) are varied simultaneously.

 

For each scenario, the table includes the Rank and resulting Output value, along with its absolute and percentage change from the base. The values for each input included in the Multi-Way analysis and the percentage change from each input’s base value are also included.

 

This ranked list helps identify the most and least favorable input combinations, making it especially useful for scenario comparison and stress testing.

 

Reports

  • Output Detail Multi-Way Report:

    This report presents ranked results for Multi-Way What-If Analyses in a clear tabular format, with a dedicated worksheet for each output. It uses the same logic and ranking structure as the Multi-Way Data Window, displaying all tested combinations of inputs labeled with a multi-function (e.g., RiskVaryMulti) and the resulting output value. It allows users to analyze how multiple input variables acting together affect a single output. The table is divided into two main sections:

    • The Output Variation section (left side) shows the output value, rank, and percent change from the base value.

    • The Input Variation section (right side) details the input values used in each scenario and their corresponding percentage changes from their base values.

    This report format is handy for identifying high-impact input combinations to understand which variables interact most critically to affect the selected output.

     

  • Tornado Multi Report:

    This report provides a ranked sensitivity analysis of all inputs labeled with a multi-function based on their collective impact on a selected output. It is generated as a separate worksheet for each output involved in the Multi-Way What-If Analysis. It helps identify which individual multi-input drives the greatest variation in that output when changing across their defined ranges. A multi-input is a combination of inputs that are varied simultaneously.

    At the top of each worksheet, a tornado graph visually summarizes the impact of each multi-input. Each bar represents the minimum and maximum changes caused by a combination of inputs. The bars are sorted from top to bottom in descending order of output impact, so the most influential combination of input values appears at the chart's top. The X-axis shows the change in output (either in absolute terms or as a percentage), making it easy to compare the relative importance of each input.

    Below the tornado graph, the report includes a data table that presents the numerical details used to generate the chart. Each row corresponds to one multi-labeled input and includes its rank (based on output impact), input name, cell location in the spreadsheet, and the range of variation applied during the analysis. The table shows the output value and the percent change for each input.