Free calculator · Finance
Turn a three-point estimate into a budget
Enter optimistic, most likely and pessimistic delivery weeks and the team you plan to use, with your own weekly rates. The calculator gives the PERT mean, an 80% estimate and the budget for each, plus the range McConnell's cone of uncertainty puts around an estimate at your project's stage.
How this is calculated
The calculator applies the PERT three-point formulas to your optimistic, most likely and pessimistic delivery weeks, prices the result at the weekly cost of the team you enter, and puts McConnell's cone of uncertainty range around the mean for your project's stage.
Step by step
- Sort the three estimates so optimistic ≤ most likely ≤ pessimistic, and say so if they were entered out of order.
- PERT mean = (optimistic + 4 × most likely + pessimistic) ÷ 6 weeks.
- Standard deviation σ = (pessimistic − optimistic) ÷ 6 weeks.
- P80 ≈ mean + 0.84σ, the 80th percentile of a normal distribution.
- Weekly team cost: for each role, people × weekly cost per person, added together.
- Budget: weeks × weekly team cost, for the optimistic, mean, P80 and pessimistic estimates.
- Cone range: the mean × the low and high multipliers for the stage you choose (initial concept 0.25–4×, approved product definition 0.5–2×, requirements complete 0.67–1.5×, user interface design complete 0.8–1.25×).
Default assumptions
Assumptions marked adjustable can be changed in the calculator; the others are fixed parts of the model.
| Assumption | Default | Sources |
|---|---|---|
| Optimistic estimateadjustable | 8 weeks | |
| Most likely estimateadjustable | 12 weeks | |
| Pessimistic estimateadjustable | 20 weeks | |
| Engineersadjustable | 2.5 | None |
| Engineer cost per personadjustable | $2,500 a week | None |
| Designersadjustable | 0.5 | None |
| Designer cost per personadjustable | $2,500 a week | None |
| QA engineersadjustable | 0.5 | None |
| QA engineer cost per personadjustable | $2,500 a week | None |
| Product or project managersadjustable | 0.5 | None |
| Manager cost per personadjustable | $2,500 a week | None |
| DevOps engineersadjustable | 0 | None |
| DevOps engineer cost per personadjustable | $2,500 a week | None |
| How defined the project isadjustable | Approved product definition | |
| Standard deviations above the mean for the 80th percentile | 0.84 | |
| Initial concept: low multiplier | 0.25× | |
| Initial concept: high multiplier | 4× | |
| Approved product definition: low multiplier | 0.50× | |
| Approved product definition: high multiplier | 2× | |
| Requirements complete: low multiplier | 0.67× | |
| Requirements complete: high multiplier | 1.50× | |
| User interface design complete: low multiplier | 0.80× | |
| User interface design complete: high multiplier | 1.25× |
What this doesn’t model
- The P80 uses a normal approximation. It is closest for a total built from many separately estimated tasks; for one estimate of the whole project it is a rough guide.
- The cone ranges are McConnell's best case for skilled estimators. Projects that are not actively narrowing their uncertainty can miss by more.
- The team is assumed to work at the same size for the whole project.
- The rates are placeholders, the same for every role, not market rates or quotes.
- Sizing the work itself: the three estimates come from you, so the result is only as good as they are.
- Hosting, third-party services, licences and tools.
- Support, maintenance and changes after launch.
- Changes to the team during the project, hiring time and onboarding.
- Scope added after the estimate; the cone assumes the definition stays put.
- Taxes, payment fees and currency movements.
Sources
- Special Projects Office, Bureau of Naval Weapons, US Navy (DTIC AD0735902), Program Evaluation Research Task (PERT), Summary Report, Phase 1 (Jul 1958). Accessed . The original PERT report. Exhibit C gives the estimating equations for an interval from optimistic (a), most likely (m) and pessimistic (b) times: mean = (a + 4m + b) ÷ 6 and standard deviation = (b − a) ÷ 6. Read in the Internet Archive's copy of the DTIC scan.
- Steve McConnell, Construx Software, Software Development's Cone of Uncertainty (white paper) (Jan 2010). Accessed . Adapted from McConnell's Software Estimation (2006), building on Boehm's work. Estimate variability of 0.25x–4x at initial concept, 0.5x–2x at approved product definition, 0.67x–1.5x at requirements complete and 0.8x–1.25x at user interface design complete. The cone is the best case for skilled estimators; worse is easily possible.
- National Institute of Standards and Technology, Cumulative distribution function of the standard normal distribution (NIST/SEMATECH e-Handbook of Statistical Methods, 1.3.6.7.1). Accessed . The table gives 0.29955 between the mean and z = 0.84, so 0.5 + 0.29955 ≈ 80% of a normal distribution lies below mean + 0.84σ.
Last reviewed by the QuantmHill engineering team. Found an error?
Link to or cite this tool
Writing about this topic? Link to the calculator or cite it. Its method, defaults and sources are all on this page, so readers can check the numbers.
Embed this calculator
You can put this calculator on your own site for free. Paste the code below where it should appear. It loads the same calculator in a frame, with a link back to this page for the full method and sources.
The credit line links to this page with the anchor text “QuantmHill”. You may edit it, add rel="nofollow" or remove it — the calculator works the same either way. Add ?theme=light or ?theme=dark to the iframe address to fix its colour scheme; otherwise it follows the visitor's system setting.
Add this once per page, after the iframe, if you want the frame to grow and shrink with the calculator instead of using the fixed height above. It accepts messages from quantmhill.com only and resizes only the frame that sent them.
Frequently asked questions
A budget from your own three-point estimate of delivery time and your own team and rates. It gives the PERT mean and standard deviation in weeks, an 80th-percentile estimate, the cost of each, the weekly cost by role and the cone of uncertainty range for how far your project's definition has got. It does not size the work for you.
PERT weights the most likely estimate four times as heavily as the optimistic and pessimistic ones: mean = (optimistic + 4 × most likely + pessimistic) ÷ 6, and standard deviation = (pessimistic − optimistic) ÷ 6. These are the estimating equations in the US Navy's 1958 PERT summary report.
An estimate you would expect to meet about 80% of the time if the total were normally distributed: mean + 0.84 standard deviations. The 0.84 comes from the standard normal table in NIST's e-Handbook of Statistical Methods. It is closest when the total is built from many separately estimated tasks.
Steve McConnell's model, building on Barry Boehm's work, of how far estimates can be off at each stage. At initial concept they can be 0.25 to 4 times the eventual outcome; at approved product definition 0.5 to 2 times; at requirements complete 0.67 to 1.5 times; once the user interface is designed 0.8 to 1.25 times. McConnell describes these as the best case for skilled estimators.
Rates depend on country, seniority and whether the team is employed or contracted, so no single default fits. The calculator opens with the same round placeholder for every role, which is not a market rate or a quote; replace it with what each person costs you per week.
Hosting, licences and third-party services, maintenance after launch, scope added after the estimate, hiring and onboarding time, and taxes. The result covers the team's time for the scope you estimated.
Want an engineer to check your numbers?
Send us your inputs and the decision you're weighing. We'll reply within one business day with an honest read on whether we can help.