Vose SoftwareModelRisk example · Manufacturing
Warranty Cost Forecast
What will this year's sales cost in warranty claims, and when? Failures run off through a two-year warranty.
From the model’s saved run
What will warranty cost?
This year's sales, over the whole warranty.
| Measure | Unit | 5th pct | Median | 95th pct | Mean |
|---|---|---|---|---|---|
| Warranty cost | US$k | 51.3 | 68.7 | 88.9 | 69.3 |
| Warranty cost per tool | US$ | 3.94 | 5.25 | 6.74 | 5.29 |
| Warranty claims | claims | 288 | 382 | 493 | 385 |
| Measure | Unit | Value |
|---|---|---|
| Expected cost, calculated directly | US$k | 69.3 |
Histogram of Warranty cost from the saved simulation.
The business question
A power-tool maker launches a redesigned drill with a two-year warranty. Sales should grow through the year to about 13,000 tools. On the previous model, 448 of 14,000 tools came back under warranty — most within the first months, a typical early-failure pattern — at about US$180 a claim for parts, labour and shipping.
The redesign should be better, but nobody knows by how much; it could even be worse. Finance must provide for the warranty cost of the year's sales, and at the year end hold a reserve for the claims that will arrive over the following two years.
- Who it is for
- Product, quality and finance teams; anyone setting a warranty provision.
- Main outputs
- The warranty cost of this year's sales, per tool and in total · the claims by calendar quarter · the reserve needed at the year end, with a prudent percentile.
How the model works
Three sheets after this one, read from left to right. Each one feeds the next.
Inputs sheet
Sales in each quarter of the year (PERT). The previous model's units and warranty claims; how much better or worse the redesign could be. The Weibull shape and scale of the time to failure; the average and spread of the cost of a claim.
Warranty sheet
The claim rate is a Beta distribution from the field data, times the redesign factor. Each quarter's sales bring a Binomial number of claims, and VoseRunoff spreads them over the eight quarters of the warranty — each at a Weibull time capped at two years, with a Lognormal cost. The four quarters' runs are added up by calendar quarter.
Results sheet
The total warranty cost and the cost per tool; the claims by calendar quarter; the reserve for claims after the year end.
ModelRisk techniques
Find similar models
- Product
- Type
- Application
- Uses
The results in detail
The year-end reserve
Claims still to come, after the year end, on tools already sold (US$ thousand).
Line chart: 95th percentile, Median, Mean, 5th percentile.
| Measure | Unit | 5th pct | Median | 95th pct | Mean |
|---|---|---|---|---|---|
| Reserve needed at the year end | US$k | 28.0 | 38.5 | 50.4 | 38.8 |
| Measure | Unit | Value |
|---|---|---|
| Average reserve | US$k | 38.8 |
| Average warranty cost | US$k | 69.3 |
| Reserve as a share of the warranty cost | 56% | |
| Prudent reserve (75th percentile) | US$k | 43.4 |
How to read these results
- This year's sales should cost about US$69,000 in warranty claims — about US$5.30 a tool — but anything from about US$51,000 to US$89,000 is plausible (5th to 95th percentile). The redesign factor drives most of that range: the new design is the biggest unknown.
- Most claims come early — a Weibull shape below 1 means early failures — so the warranty cost peaks in the fourth quarter, when the year's growing sales meet their first months of use, and then fades over two years.
- At the year end about 56% of the warranty cost is still to come on tools already sold: a reserve of about US$39,000 on average, and about US$43,400 at the 75th percentile. The reserve, not the year's claims, is what the accounts must carry.
- The direct calculation — expected tools × expected claim rate × average claim — gives the same mean. The simulation adds the range, the timing and the reserve.
- Watch the claims in the first months after launch: early failures reveal the redesign factor quickly. Updating the claim rate with that data will narrow the range.
See it move. Open the workbook, change an input on the Inputs sheet and run the simulation: every table and chart on this page updates in Excel.
Inputs and assumptions
Every assumption in the model, with its uncertainty and where it comes from. The figures are illustrative.
Sales this year 4 inputs
Tools sold in each quarter.
| Assumption | Unit | Uncertainty | Low | Most likely | High |
|---|---|---|---|---|---|
| Quarter 1Rising as the launch builds. | tools | PERT | 2,200 | 2,500 | 2,900 |
| Quarter 2 | tools | PERT | 2,600 | 3,000 | 3,500 |
| Quarter 3 | tools | PERT | 3,000 | 3,500 | 4,200 |
| Quarter 4 | tools | PERT | 3,300 | 4,000 | 4,900 |
Field data and the redesign 3 inputs
| Assumption | Unit | Uncertainty | Low | Most likely | High |
|---|---|---|---|---|---|
| Previous model: tools in the fieldTwo full years of warranty. | tools | Fixed | 14,000 | ||
| Previous model: warranty claims3.2% of tools. | claims | Fixed | 448 | ||
| Redesign factor on the claim rateEngineering expects fewer failures, but a new design can surprise. | × previous | PERT | 0.60 | 0.90 | 1.30 |
Failures and their cost 4 inputs
| Assumption | Unit | Value |
|---|---|---|
| Time to failure: Weibull shapeBelow 1: most failures come early. | 0.80 | |
| Time to failure: Weibull scale | years | 3.0 |
| Average cost of a claimParts, labour, shipping. | US$ | 180 |
| Standard deviation of the cost of a claim | US$ | 120 |
Units and conventions
- Money
- US$; totals in US$ thousand.
- Quarters
- Calendar quarter 1 is the first quarter of this year; quarters 5 to 11 come after the year end.
- Claim rate
- The share of tools sold that come back under warranty within two years.
- Reserve
- The expected cost of claims after the year end on tools already sold.
Key assumptions and limitations
- Sales
- Tools are sold at the start of their quarter; the warranty runs eight quarters from the sale.
- Claims
- At most one claim per tool; tools fail independently; claim costs are independent of timing.
- Field data
- The previous model's experience applies, scaled by the redesign factor.
- Data
- The product and all figures are invented for this example.
How to use it
- Enter your sales forecast by quarter.
- Enter your previous product's field data and how different the new design could be.
- Run a simulation from the ModelRisk ribbon (10,000 samples with a fixed seed are set).
- Book the expected warranty cost per tool, and hold a prudent percentile of the year-end reserve.
Simulate this model with ModelRisk’s classic engine: it does not run in Turbo mode yet.
ModelRisk functions used
- VoseBeta
- The claim rate the field data support.
- VoseBinomial
- The number of claims on a quarter's sales.
- VoseRunoff
- Spreads a number of claims over periods, each at a random time with a random cost (an array).
- VoseWeibullObject, VoseXBounds, VoseLognormalObject
- Time to failure, capped at the warranty, and the cost of a claim.
- VoseSimMean, VoseSimPercentile
- Statistics of the simulation on the Results sheet.
Further reading
- Monte Carlo simulation explained (Vose guide)
- Blischke, W. R. & Murthy, D. N. P. (1994). Warranty Cost Analysis. Marcel Dekker.
- Vose, D. (2008). Risk Analysis: A Quantitative Guide, 3rd edition. Wiley.
Download the workbook and run it yourself. It opens in Excel with ModelRisk. The free trial runs every example model.
Related example models
- ModelChoice★★★★★Supply Chain Resilience DecisionA nearshore second source, a better offshore supplier, vertical integration, or the status quo?Three decisions and ten chance events. Best first decision: Single Offshore Supplier, expected value $12.1m.
- ModelRiskOptimizer★★★★★Inventory Policy Chosen by the OptimizerWhich reorder point and order quantity cost least while keeping the service the customers were promised?Demand over a random lead time as an aggregate; the reorder point and order quantity chosen together by the Optimizer for a service target.
- ModelRisk★★★★★Inventory Policy under Uncertain Demand and Lead TimeWhen should we reorder?Demand during a random lead time as an aggregate, the service level of each reorder point, and its cost.
An illustrative example: the figures are invented for teaching and do not describe a real company, project or asset. Last updated 9 October 2026.