Vose SoftwareModelRisk example · Projects and construction
Rework and Change Orders
Is a 5% change allowance enough? Change orders and rework as a random number of events of random size.
From the model’s saved run
Is the allowance enough?
The total cost of changes and rework against the allowance in the cost plan.
| Measure | Unit | Value |
|---|---|---|
| Change allowance (Inputs sheet) | US$m | 3.00 |
| Chance of exceeding the allowance | probability | 66.4% |
| P80 cost of changes and rework | US$m | 4.18 |
| Allowance needed for 80% confidence | of contract | 7.0% |
| Expected cost of changes and rework | US$m | 3.46 |
| Expected cost, calculated exactly (check) | US$m | 3.46 |
S-curve (cumulative probability) of Total change cost from the saved simulation.
The business question
A US$60 million school building will take 24 months. The cost plan carries a 5% allowance for changes, the usual figure. Changes come from three sources. The client will instruct changes — a few a month, but nobody knows the rate for this client, so the rate itself is uncertain.
The design will have errors and omissions that the contractor is entitled to be paid for. And defective work will have to be redone. Each change is small on its own, but a few are expensive, and their number is uncertain.
The model counts the events in each class, gives each one a cost, adds them up, and asks whether the 5% allowance is enough — and what allowance would be.
- Who it is for
- Cost managers, quantity surveyors, contract administrators and clients setting change allowances.
- Main outputs
- Total cost of changes and rework: mean, P50 / P80 / P90 · the chance of exceeding the allowance · the allowance needed for 80% confidence · the number of change orders and its range · the contribution of each class.
How the model works
Three sheets after this one, read from left to right. Each one feeds the next.
Inputs sheet
The contract sum, its duration and the allowance. For each class, the expected number of events per month and the mean and spread of an event's cost. The client's change rate has its own uncertainty.
Changes sheet
The number of events in each class is Poisson with mean = rate × months (for client changes, the rate is first drawn from a Gamma). VoseAggregateMC then adds up that many Lognormal costs.
Results sheet
The total change cost against the allowance, the allowance that would cover it with 80% confidence, the number of client change orders, and how each class contributes.
ModelRisk techniques
Find similar models
- Product
- Type
- Application
- Uses
The results in detail
Cost by class
P50 / P80 / P90: a 50%, 80% and 90% chance of not being exceeded.
| Measure | Unit | P50 | P80 | P90 | Mean |
|---|---|---|---|---|---|
| Client-instructed changes | US$m | 1.62 | 2.31 | 2.74 | 1.73 |
| Design errors and omissions | US$m | 1.06 | 1.55 | 1.84 | 1.15 |
| Rework of defects | US$m | 0.56 | 0.71 | 0.81 | 0.58 |
| Total | US$m | 3.36 | 4.18 | 4.70 | 3.46 |
Number of events
10th percentile, median and 90th percentile.
| Measure | Unit | 10th pct | Median | 90th pct | Mean |
|---|---|---|---|---|---|
| Client change orders | orders | 23 | 37 | 56 | 38 |
| Design errors and omissions | events | 10 | 14 | 20 | 14 |
| Rework items | events | 22 | 29 | 36 | 29 |
What drives the total
Lowest and highest mean total over ten equal groups of each input (VoseSimSensitivityMean), drawn from the overall mean.
Tornado chart: the inputs ranked by how far the mean output moves across each input’s range.
The ranking as a table 4 inputs
| Input (ranked by influence) | Lowest mean | Highest mean | Rank |
|---|---|---|---|
| Client-instructed changes: cost | 2.39 | 5.01 | 1 |
| Design errors and omissions: cost | 2.75 | 4.58 | 2 |
| Client change rate | 2.68 | 4.43 | 3 |
| Rework of defects: cost | 3.18 | 3.78 | 4 |
How to read these results
- The expected cost of changes is above the 5% allowance, so the allowance is more likely than not to be exceeded. The P80 shows what allowance would give 80% confidence.
The client's change rate is uncertain, so the number of change orders varies much more than a Poisson count with a known rate would: a demanding client can double the count. The client's changes lead the tornado chart: their total cost has the longest bar, and the uncertain rate behind it one of the next.
The tornado ranks only the rate and the class totals: the counts in between are part of the same chain, not separate causes.
- Rework items are frequent but small and similar in size, so their total is much more predictable than the cost of client changes, where a few large changes dominate.
- Change control that reduces the client's rate, or a more complete design at tender, attacks the drivers at the top of the ranking.
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.
Contract 3 inputs
The contract the changes are measured against.
| Assumption | Unit | Value |
|---|---|---|
| Contract sumLump-sum building contract. | US$m | 60.0 |
| Contract durationStart on site to practical completion. | months | 24 |
| Change allowanceThe usual allowance in the client's cost plan. | of the contract sum | 5.0% |
Client-instructed changes 4 inputs
Variations the client asks for. The client's rate is itself uncertain.
| Assumption | Unit | Value |
|---|---|---|
| Expected changes per monthSimilar school projects: 30 to 45 changes over two years. | per month | 1.60 |
| Uncertainty of that rateSome clients change far more than others: the rate is drawn from a Gamma with this spread. | coefficient of variation | 30% |
| Mean cost of a changeChange logs: US$45,000 on average. | US$m | 0.045 |
| Standard deviation of the costMost are small, a few are large. | US$m | 0.080 |
Design errors and omissions 3 inputs
Changes the design team causes and the client pays for.
| Assumption | Unit | Value |
|---|---|---|
| Expected events per monthDesign 85% complete at tender. | per month | 0.60 |
| Mean cost of an event | US$m | 0.080 |
| Standard deviation of the cost | US$m | 0.120 |
Rework of defects 3 inputs
Work that has to be redone; recovered from nobody.
| Assumption | Unit | Value |
|---|---|---|
| Expected events per monthQuality records of the main contractor. | per month | 1.20 |
| Mean cost of an event | US$m | 0.020 |
| Standard deviation of the cost | US$m | 0.025 |
Units and conventions
- Money
- US$ million (US$m); event costs entered in US$m (0.045 = US$45,000).
- Rates
- Expected number of events per month of the contract.
- Allowance
- A percentage of the contract sum reserved for changes.
- P80
- The cost with an 80% chance of not being exceeded.
Key assumptions and limitations
- Events
- Events occur at random through the contract (Poisson); costs are independent of each other and of their number.
- Client rate
- The client's average rate is uncertain (Gamma, 30% coefficient of variation): some clients change much more.
- Data
- The project, rates and costs are invented for this example.
How to use it
- Calibrate the rates and costs from the change logs of similar projects (number per month, average and spread of cost).
- Set the contract sum, duration and allowance on the Inputs sheet.
- Run a simulation from the ModelRisk ribbon (10,000 samples with a fixed seed are set).
- Compare the allowance with the P80 change cost; the drivers show which class to manage.
The tornado chart fills in after a simulation with ModelRisk’s classic engine.
ModelRisk functions used
- VosePoisson, VoseGamma
- The number of events per class; the client's uncertain change rate.
- VoseAggregateMC, VoseLognormalObject
- The total cost of a random number of Lognormal events.
- VoseInput, VoseOutput
- Name the uncertain quantities and the outputs.
- 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 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)
- Hanna, A. S. et al. (2002). Impact of change orders on labor efficiency for mechanical construction. Journal of Construction Engineering and Management 128(1), 57–64.
- Vose, D. (2008). Risk Analysis: A Quantitative Guide, 3rd edition. Wiley — chapter on aggregate distributions.
Download the workbook and run it yourself. It opens in Excel with ModelRisk. The free trial runs every example model.
Related example models
- 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★★★★★Cost Estimate with ContingencyHow much contingency does the warehouse budget need?Twelve cost items with three-point ranges added up by simulation: the simplest useful risk model.
- 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.