What can we learn from 3 charts?
Notice that this team doesn't care about intermediate columns like Development, Review, QA, ... Once the team starts a piece of work, it is considered in progress. When it is deployed into production, it is considered done.
The backlog isn't growing, and it's being replenished at regular intervals. 49 items arrived in the last three months, 50 finished, and the To Do count sits right where it stood 90 days ago. Work in progress is stable, work isn't piling up faster than it's being cleared. This looks like a healthy pace.
The cycle time scatter plot and the histogram are two different views on the same information.
The scatter plot focuses on the individual pieces of work. It helps us identify the outliers. Now that the work has been done, the latest outliers are worth discussing during the next retrospective. Could we have split this work in a way to get feedback sooner?
85% of the work was done in 17 days or less over the last three months. The items that take longer pull the tail out to 30+ days. Those are the ones worth examining.
One thing worth a conversation: of the items currently in progress, one is already older than 85% of everything this team finished in the last three months. It is worth asking whether we can do something about this.
This is the fan from the Monte Carlo page, built from this team's own history. Every forecast learns from the 90 days before the dotted line, and nothing else.
Drag the handle back to any day since January. The chart then shows the forecast as the team would have made it that day, knowing only what it knew then. The thick line to the right of it is what really happened next. Press play and watch the two side by side.
That line changes colour with where it lands in the fan. Teal is above the 50% line, yellow between the 50% and the 85% line, orange between the 85% and the 95% line, and red is below all three. Pointing at the chart colours the actual number the same way.
Don't expect reality to follow the 50% line. It wanders around it, sometimes above, sometimes below. That's what 50% means.
All three lines are "at least" lines. The 85% line says the team finished at least that many items in 85 out of 100 futures. So reality running above the fan isn't a miss. It's a good stretch. A miss is reality dropping below the 85% line.
You'll find a few of those, and each one has a story. Early in the year, the forecast learns from 90 days that include the holidays, so it expects less than the team then delivers. In late April the team hit a slow spell, and reality sinks under the fan for a while before it catches up.
Over the whole year the fan holds up. Take a forecast made on any day since January and the team did at least as well as the 85% line about 9 times in 10, whether you look two weeks ahead or three months. Forecasts made in April are the exception, closer to 6 in 10. If anything the fan is a little cautious. Don't read those numbers as 250 separate tests, though. Two neighbouring days learn from almost the same 90 days and get checked against almost the same future, so this year's three-month forecasts add up to only about two independent ones.
At "now" there's nothing to check against yet, so read the fan the other way round. Point at a day ahead and read the 85% line. Four weeks from now, this team finishes at least 11 more items in 85 out of 100 futures. Twelve weeks out, at least 40. That's the number I'd put in front of a stakeholder, not the 15 or 47 the 50% line offers.
Want to see these charts for your own team? Try them on your own data. Your data is safe and won't leave your computer. Nothing is sent anywhere, not even to me.