Vose SoftwareModelRisk example · Projects and construction
Cost Estimate with Contingency
How much contingency does the warehouse budget need? Three-point estimates turned into a P50 and a P80 budget.
From the model’s saved run
The budget
P80 = the cost with an 80% chance of not being exceeded.
| Measure | Unit | Value |
|---|---|---|
| Base estimate (sum of the most likely costs) | US$m | 36.80 |
| P80 budget | US$m | 38.60 |
| Contingency for a P80 budget | US$m | 1.80 |
| Contingency as a share of the base estimate | 4.9% | |
| Chance that the base estimate is enough | probability | 12.5% |
| Mean total cost | US$m | 37.83 |
S-curve (cumulative probability) of Total cost from the saved simulation.
The business question
The estimator's cost plan for a new 40,000 m² distribution warehouse has twelve items. For each one the team has agreed a lowest plausible cost, a most likely cost and a highest plausible cost.
Adding up the most likely costs gives the base estimate of US$36.8 million — but the ranges are not symmetric: most items can overrun by more than they can underrun, so the base estimate is likely to be too low.
The model replaces each item by a PERT distribution built from its three estimates, adds them up by simulation, and reads off the budget at the confidence level the sponsor wants. The difference between that budget and the base estimate is the contingency, and the ranking of the items shows where better estimates would narrow the range most.
- Who it is for
- Cost estimators, quantity surveyors, project managers and the sponsors who approve budgets.
- Main outputs
- Total cost: P50, P80 and P90, the chance that the base estimate is enough, the contingency needed for a P80 budget, and which items drive the range.
How the model works
Three sheets after this one, read from left to right. Each one feeds the next.
Inputs sheet
For each item, the lowest, most likely and highest plausible cost, with the basis of the estimate. Each becomes a PERT distribution: most likely values carry the most weight, extremes are possible but rare. One correlation says how strongly the items' costs move together: 0, independent, as shipped.
Estimate sheet
The base estimate is the sum of the most likely costs. In each iteration a Normal copula draws twelve U values with that correlation between every pair, each item's PERT turns its U into a cost, and the total is their sum.
Results sheet
The total cost's percentiles: P80 is the budget with an 80% chance of not being exceeded. The contingency is that budget minus the base estimate. The tornado ranks the items by how much they drive the total.
ModelRisk techniques
Find similar models
- Product
- Type
- Application
- Uses
The results in detail
Budget at other confidence levels
Choose the confidence level, then read the budget.
| Measure | Unit | P50 | P80 | P90 | Mean |
|---|---|---|---|---|---|
| Total cost | US$m | 37.81 | 38.60 | 38.99 | 37.83 |
| Expected cost, calculated exactly (check) | US$m | 37.83 |
Which items drive the range
Lowest and highest mean total over ten equal groups of each item (VoseSimSensitivityMean): estimate the top items more carefully first.
Tornado chart: the inputs ranked by how far the mean output moves across each input’s range.
The ranking as a table 12 inputs
| Input (ranked by influence) | Lowest mean | Highest mean | Rank |
|---|---|---|---|
| Mechanical and electrical services | 37.12 | 38.66 | 1 |
| Steel frame | 37.23 | 38.60 | 2 |
| Foundations and ground-floor slab | 37.31 | 38.48 | 3 |
| Cladding and roofing | 37.39 | 38.32 | 4 |
| Site preparation and earthworks | 37.42 | 38.32 | 5 |
| Preliminaries and site management | 37.51 | 38.25 | 6 |
| Racking and warehouse fit-out | 37.55 | 38.20 | 7 |
| Drainage and external works | 37.58 | 38.16 | 8 |
| Offices and staff amenities | 37.64 | 38.01 | 9 |
| Loading docks and doors | 37.65 | 38.01 | 10 |
| Design and professional fees | 37.66 | 38.01 | 11 |
| Sprinklers and fire protection | 37.67 | 37.97 | 12 |
How to read these results
- The base estimate adds up most likely costs. Because most items can overrun by more than they can underrun, the sum of the most likely values sits below the middle of the range, and the chance that it is enough is low.
- The contingency is not a guess or a fixed percentage: it is the difference between the budget at the chosen confidence level and the base estimate. Sponsors often fund the P50 and hold the step to the P80 as a management reserve.
As shipped the items are independent (correlation 0 on the Inputs sheet), which lets overruns on some items cancel underruns on others.
When costs move together — a longer programme, a tight labour market — the range is wider: at a correlation of 0.3 the P80 contingency rises to about 6.7% of the base estimate (US$2.5 million), and the tornado's bars even out, because each item then carries the others that move with it. PC-02 adds risk events such as a delayed power connection.
- The items at the top of the tornado are where the range comes from: a firmer price or more design there narrows the budget more than anywhere else.
See it move. Open the workbook, change an input on the Inputs sheet and run the simulation: every table and chart on this page updates in Excel.
Inputs and assumptions
Every assumption in the model, with its uncertainty and where it comes from. The figures are illustrative.
Site and substructure 3 inputs
Below ground and around the building: where ground conditions bring the most uncertainty.
| Assumption | Unit | Uncertainty | Low | Most likely | High |
|---|---|---|---|---|---|
| Site preparation and earthworksGround investigation found variable fill on the north side. | US$m | PERT · correlated | 1.80 | 2.20 | 3.20 |
| Foundations and ground-floor slabPiling may be needed under the high-bay area. | US$m | PERT · correlated | 4.00 | 4.60 | 6.00 |
| Drainage and external worksAttenuation tank sized by the local authority. | US$m | PERT · correlated | 1.20 | 1.50 | 2.10 |
Building 3 inputs
The frame and envelope: mostly quoted prices.
| Assumption | Unit | Uncertainty | Low | Most likely | High |
|---|---|---|---|---|---|
| Steel frameSteel price quoted, valid for 60 days. | US$m | PERT · correlated | 6.50 | 7.20 | 8.80 |
| Cladding and roofingTwo subcontract quotes; the higher includes rooflights. | US$m | PERT · correlated | 3.80 | 4.30 | 5.40 |
| Loading docks and doors24 dock levellers, specification agreed. | US$m | PERT · correlated | 1.10 | 1.30 | 1.70 |
Services and fit-out 4 inputs
Services design is least advanced, so its ranges are widest.
| Assumption | Unit | Uncertainty | Low | Most likely | High |
|---|---|---|---|---|---|
| Mechanical and electrical servicesDesign at stage 2; the power connection cost is not yet confirmed. | US$m | PERT · correlated | 4.50 | 5.20 | 7.00 |
| Sprinklers and fire protectionInsurer's requirements still to be confirmed. | US$m | PERT · correlated | 1.00 | 1.20 | 1.60 |
| Racking and warehouse fit-outClient's layout, supplier's budget price. | US$m | PERT · correlated | 2.80 | 3.10 | 3.90 |
| Offices and staff amenitiesTwo-storey office block, standard finish. | US$m | PERT · correlated | 1.40 | 1.60 | 2.10 |
Preliminaries and fees 2 inputs
Time-related site costs and fees.
| Assumption | Unit | Uncertainty | Low | Most likely | High |
|---|---|---|---|---|---|
| Preliminaries and site managementTime-related: a longer build costs more. The programme is not modelled here, so this is a range of its own. | US$m | PERT · correlated | 2.50 | 2.80 | 3.60 |
| Design and professional feesPercentage fees on the construction cost, entered here as a range of their own. | US$m | PERT · correlated | 1.60 | 1.80 | 2.20 |
How the costs move together 1 input
One correlation for every pair of items.
| Assumption | Unit | Value |
|---|---|---|
| Correlation between any two items' costs0 = independent, as most cost plans assume. A shared cause, such as a longer programme or a tight labour market, makes costs move together: try 0.3 and simulate again. | 0 – 1 | 0.00 |
Units and conventions
- Money
- US$ million (US$m), at today's prices, excluding inflation and VAT.
- P50 / P80 / P90
- The cost with a 50%, 80% and 90% chance of not being exceeded.
- Contingency
- Budget at the chosen confidence level minus the base estimate (the sum of the most likely costs).
- Base estimate
- The single-number estimate that adds up most likely values.
Key assumptions and limitations
- Independence
- Items are independent as shipped (correlation 0). In reality some move together (labour rates, a longer programme). Raise the correlation on the Inputs sheet to see how much that widens the range.
- Scope
- Only estimating uncertainty within the agreed scope: no risk events (see PC-02) and no scope change.
- Distributions
- PERT from the three estimates; the low and high are treated as true minimum and maximum.
- Data
- The project and all costs are invented for this example.
How to use it
- Replace the twelve items and their three estimates with your own on the Inputs sheet (the model holds exactly twelve. Ask for the high estimate first, then the low, then the most likely: it reduces anchoring).
- Run a simulation from the ModelRisk ribbon (10,000 samples with a fixed seed are set).
- Read the P50, P80 and P90 on the Results sheet and choose the confidence level for the budget.
- Work on the items at the top of the tornado first: better information there narrows the range most.
The tornado chart fills in after a simulation with ModelRisk’s classic engine.
ModelRisk functions used
- VosePERT
- Each item's cost from its low, most likely and high estimates, and its U value.
- VoseCopulaMultiNormal
- Twelve U values in each iteration, one per item, from the correlation matrix on the Estimate sheet.
- VoseInput, VoseOutput
- Name the twelve items and the total for the Results Viewer and the sensitivity analysis.
- VoseSimMean, VoseSimPercentile, VoseSimProbability
- Statistics of the simulation for the tables and charts on the Results sheet.
- VoseSimSensitivityMean
- The tornado chart: every VoseInput ranked by how far the mean total cost moves across its range.
Further reading
- How to calculate cost contingency (Vose guide)
- Sensitivity analysis in Excel: a tornado chart step by step (Vose guide)
- AACE International (2008). Recommended Practice 41R-08: Risk Analysis and Contingency Determination Using Range Estimating.
- Vose, D. (2008). Risk Analysis: A Quantitative Guide, 3rd edition. Wiley — chapter on eliciting expert opinion.
Download the workbook and run it yourself. It opens in Excel with ModelRisk. The free trial runs every example model.
Related example models
- ModelRisk★★★★★Rework and Change OrdersIs a 5% change allowance enough?Three classes of change, each a random number of events with random costs, added up with VoseAggregateMC.
- ModelRisk★★★★★Risk Register: Threats and OpportunitiesHow much risk contingency does the hospital extension need?A risk register turned into a simulation: risk events with linked cost and schedule impacts.
- ModelRisk★★★★★Material Price Escalation and Procurement StrategyBuy the steel now, as needed, or with a price cap?Monthly steel and concrete price paths over 18 months and four ways of buying the steel, on the same paths.
An illustrative example: the figures are invented for teaching and do not describe a real company, project or asset. Last updated 9 October 2026.