FogBugz uses an algorithm called Evidence-Based Scheduling (EBS) to compute the probability that your product will ship by any given date.
From the Reports menu, choose the project for which you want a report.
The report page has three sections.
In the top left, you'll see a table of all the upcoming releases for this project, sorted by their official date. Drag the slider up and down to control which releases should be included in the chart. For example, if you have a "beta" release, drag the slider to that release to show the beta ship dates.
In the top right, there is a slider listing priorities 1-7. Drag this slider up and down to control which priority cases should be included in the chart. You can use this to quickly see how the schedule is affected based on which priority features you implement. For example, if the ship dates EBS calculates are simply too late, you can quickly determine if you can finish on time by eliminating lower priority cases.
The chart at the bottom of the page has three tabs. It can show you three kinds of reports:
Ship Date Confidence Distribution
This chart shows the probability that you will finish all estimated cases for any given date, for the release and priorities you selected above. The x-axis shows a ship date, while the y-axis shows the probability. You can move your mouse around the chart to examine numbers in detail.
It answers questions like:
- Our customers expect the new release on September 1. What is the probability that we will ship by then?
- By what date can we be 90% confident that we will have shipped?
Even though EBS doesn't perfectly predict the future, this chart is extremely useful in planning. You might discover, for example, that you only have a 3% chance of hitting the official date. When you have that information early, you can decide to cut features or change the official date before it's too late.
Observe the steepness of the chart.
- A very steep chart reflects a great deal of confidence in the schedule. For example, if the 5% date and the 95% date are only a week apart, you can be pretty darn sure that you're going to ship that week.
- A gradual chart reflects a lot of uncertainty in the schedule. If the chart gradually rises over the course of several months, EBS cannot predict the ship date with much accuracy. This happens for two reasons:
- EBS doesn't have a lot historical data on how good the estimators are. This data is gleaned from timesheets. Ensure that developers are using the Working On menu to track their time. After the first six features per estimator have been completed which have estimates and which were tracked on time sheets, EBS will have much more realistic data about the quality of the estimators and the chart can steepen.
- Estimators have not been very good at estimating features in the past. If estimators have a poor track record predicting how long features will take, the EBS algorithm reflects this by showing a gradual chart. Encourage your estimators to check their estimation history regularly, and revise their estimates continually. As they get better and better at estimating, the steepness of the chart will reflect more and more predictable ship dates.
Ship Date Over Time
Every night, FogBugz stores a snapshot of the ship date confidence distribution for that day. This chart allows you to see how the estimated dates are changing from day to day. On the x-axis, it shows the date on which the calculation was done. On the y-axis, it shows the ship dates calculated on that date.
There are three lines. The lower line shows the 5% confidence ship date (there was only a 5% chance of hitting that date); the middle line shows the 50% confidence ship date, and the top line shows the 95% date.
It answers questions like:
- Are we staying on schedule or slipping?
- Are our estimates getting better or worse?
Observe whether the chart is climbing, sinking, or flat.
- A chart which is climbing means that the ship date that FogBugz estimates is moving out into the future. Yes, this means you're slipping, but it's not because things are taking longer than estimated. EBS should have already reflected things taking longer than expected in its original calculation of the date. If the date that EBS calculates keeps moving out into the future, it's because you're adding features. To avoid schedule risk, create a schedule item for each developer on the team called Buffer - New Features, and estimate how much time needs to be reserved for new features. Later, when you do add new features, either charge their time directly to this buffer schedule item on timesheets, or reduce the buffer schedule item estimate.
- A chart which is climbing faster than one day per day reflects a project that will never ship, because you are adding features faster than you are implementing them.
- A chart which is sinking means that the ship date FogBugz estimates is coming forward. This reflects features you had originally planned which you now think you won't do.
- A flat chart reflects a project that is under control.
Observe whether the three lines are converging (moving closer together).
- If the lines are moving closer together, the range of possible ship dates is getting narrower and narrower. This reflects increasing confidence in your ship date.
- If the lines are moving apart, the range of ship dates is widening. This reflects decreasing confidence in your ship date.
Developer Ship Dates
The developer ship date chart can be used to see which team members have the most work on their plates and are in the critical path.
It answers questions like:
- How can you redistribute work to ship earlier?
- Which developers have the most reliable estimates?
For each developer, you get a horizontal bar which shows five points: the developer's 5%, 25%, 50%, 75%, and 95% probability ship dates.
The further to the right a developer's bar appears, the later that developer will finish their work.
Observe whether the bar is wide or narrow. The widest bars reflect the developers who have the most uncertainty in their estimates.
In some cases, this may not matter. If Alice has a very wide bar, but she still has a 95% chance of finishing well before Bob, then Alice is not in the critical path, and her poor estimates cannot affect the schedule.
If it does matter, click on the developer's name to see that developer's estimate history, so you can see why EBS doesn't trust that developer's estimates.