How late will the release really be?
Eight tasks, three-point estimates and a commitment date at eighty percent confidence. Then run the same table on the other duration distribution and watch the answer move six days, which is worth knowing before you quote either one.
Words on this sheet
- Predecessors: The tasks that have to finish before this one can start. Put their IDs here, separated by commas, and leave it blank when nothing has to come first.
- F statistic: The variation between the groups divided by the variation inside them.
SaaS Intermediate Schedule Risk Pro engine
After you install, this is the model to open.
Can We Ship on That Date?
- In your spreadsheet, click the Sortia icon in the strip of icons down the right-hand edge. No strip? Click the arrow at the bottom-right to open it. You can also use Extensions, then Sortia, then Open Sortia.
- Click Start from a template and put that name in the search box.
- Pick the card with that name and click Load this template. It arrives on a new tab with real numbers already in it.
This one runs on a Pro engine, and every free install includes five full-quality runs on your own numbers, shared across all five Pro engines rather than five for each. After that, Pro is $199/year.
The answer
- The plan on paper
- 39 working days at the likely estimates
- A coin flip
- 44.1 days at P50: the plan is already five days optimistic
- Commit to
- 47.8 days, the P80 across 20,000 schedules
- What drives it
- 91% criticality of the schema migration; the front end just 9%
Eight tasks and one release. Adding the likely estimates down the longest chain gives 39 days, which is the date on the roadmap. Click Run on the twenty thousand trials the template loads with, PERT selected, and the picture is fuller: mean 44.3 days, P50 44.1, P80 47.8, P95 51.5. Commit to 48 working days and you are right about 80 times in 100.
The criticality table says the spine is the spec freeze, the integration test, the hardening and the rollout at 100 percent, with the schema migration at 91 and the backend at 76. The billing integration drives the finish in only 15 percent of runs and the front-end build in 9, so the widest estimate on the sheet is not the one that decides the date, which is a useful correction to the instinct that says worry about the third party.
Now the second run, and it is the point of this template. Switch the duration distribution from PERT to Triangular and run again on the same seed and the same table. Nothing about the plan changed. The P80 moves from 47.8 days to 53.7 and the P50 from 44.2 to 49.3. Nearly six days of commitment, out of a choice in a segmented control. The reason is straightforward once you see it.
PERT weights the most likely value four times against the two ends, which is a claim that your likely estimate carries most of the information. Triangular takes the three points literally and gives the extremes their full share, so a table with long pessimistic legs, like this one, produces a much fatter right tail. Neither is right in general.
PERT is the better model when the three numbers came from somebody describing a normal week and the pessimistic case is a genuine bad week. Triangular is the better model when the pessimistic number is a hard limit somebody has actually hit, or when the estimates came out of a contract rather than out of a memory. The honest practice is to know which of those your estimates are, and to say which distribution you used when you quote a date.
If you cannot tell, run both and quote the wider one, because the cost of being late is almost always larger than the cost of quoting five days more. What this cannot tell you is whether the tasks are independent. If the same two engineers are on the backend and the billing integration they cannot really run in parallel, and neither distribution will find that; run Resource Load on the same table with an Owner column added first.
To adapt it, put your own tasks in, and set the pessimistic estimates from the worst version of that task your team has actually lived through rather than from a round multiple of the likely one.
The model
It arrives on a tab called Template: Release Date, carrying these columns:
- ID
- Task
- Predecessors
- Optimistic (days)
- Likely (days)
- Pessimistic (days)
Once it is in your sheet
- The model arrives with real numbers in it and runs as it stands, so you can press the button first and understand it second.
- Change the numbers to yours. The sheet marks which cells are inputs and which hold formulas, and most labels carry a note explaining the row.
- Press the run button at the bottom of the panel. It is labeled for the tool you are in, and the result lands on its own tab, with a written reading of it beside the figures.
Never used Google Sheets? Start here goes the whole way, in seven steps, and assumes nothing.
Next question
- Does your roadmap leave revenue on the floor?Which Features Fit the Quarter?
- What will your ARR be in 12 months?SaaS ARR Forecast with Uncertainty
- What is staying put actually costing you?Switch Cloud Providers, or Stay?
- Will this market tip to a single winner, or stay split?Format War Tipping Simulator
- Should you sell the hardware below cost to win the platform?Platform Launch Pricing
- If you raise prices, do you come out ahead?If We Raise Prices, Do We Come Out Ahead?
Every model like this one, and the method behind them: Schedule risk analysis.