The Cycle Time Scatterplot

Content:
1. Introduction to the Cycle Time Scatterplot
2. How does this chart work?
3. What's that telling us?

You can find more information about the Analytics module in the dedicated article.

1. Introduction to the Cycle Time Scatterplot

The Cycle Time Scatterplot chart is a representation of how long it takes to get things done for individual items on your Kanban board. It visualizes the work items that have been completed on a given board (inside single or multiple workflows). 

The chart visualizes the cycle time of tasks within a predetermined time frame. It also gives information on when the cards' completion in the selected workflow can be expected. The certainty of this is measured with percentiles.
To learn more about the scatterplot concept, please visit our blog.

2. How does this chart work?

Across the "x" axis is a representation of time, i.e. Completion date. You can see the dates across the button of the chart. 
On the "y" axis is the Cycle time in days.

Every time when an item is completed, the system plots a dot on the vertical against the date of completion, and where it goes depends on the cycle time*. For example: on April 24th, an item was completed, and it took 9 days to get done.  
A lot of the dots have small numbers inside. It means that the same number of items have been finished on that day. On hover, you can see the card IDs. 

* - If a card has spent time in a column that does not participate in the cycle time calculation for the board, this time will be deducted from the card's metrics and its Done date (on the Scatterplot) might differ from the card's Done date recorded in its History tab.

You can therefore use the 'ignore cycle time calculation' option on the chart so the analytics engine would take into account the full time the card has spent on each stage of the process and the Done dates (on the scatterplot and the card's history tab) should then be aligned.
 
Important: 
The Cycle Time Scatterplot allows for recalculation based on working days.

Note: the X-axis (where applicable) would still display all calendar days/weeks.

The selection of working days recalculates the cycle time for all cards shown in the graph, but they do not change their position in time. Let's say we have a card that was moved to ‘In Progress’ on a Wednesday e.g., 06.01.2021, and was completed on Saturday 09.01.2021, and the team is working from Monday to Friday.

If no working days option is selected, the card's cycle time would be calculated as 4 days.

If the working days (Mon-Fri) option is selected, that card's cycle time will be recalculated from 4 days to 3, but the point for it on the Cycle Time Scatterplot would still remain on the line corresponding to the date it was completed on (09.01.2021).

In case that you need precise cycle and log times for internal reporting, use the Advanced search capabilities with the different chart views and widgets.

Hover on the Card ID to open the respective "Work Item Details" box. It contains specific information relevant to the card. In addition, you can check the time this card has spent in every stage of the workflow. 

Besides showing you some visual representation of how work actually happens inside the board, this chart is important for one other reason.
It provides a probabilistic view that is activated via the percentiles (50%, 70%, 85%, 90%), which you control from the Chart Controls. Select the percentiles for the corresponding control.  

  
If you go all the way to the left following the horizontal lines, you will see a number for each of those lines.

• 95% -> 11 days
• 85% -> 6 days
• 70% -> 3 days
• 50% -> 2 days 
  
   

3. What's that telling us?

Let's start with the 50% percentile. If you draw a horizontal line all across the chart, it will cut the dots in half - 50% of the dots will be below the line, and 50% of them will be above the line. 
The chart is telling us that 50 % of those items took 2 days or less to be completed. An alternative assumption might be that when an item enters our process, it has a 50% chance to be completed in 2 days or even less. 

Imagine that you get a new request coming into this workflow. You naturally ask the question: How much time do I need to get this done? Then you can instantly say that there is a 95 % probability that your team will complete the work items within 11 days as long as it's something like what you used to do here.

If it is not something radically new, then 11 days will be more than certain to be sufficient.
This chart allows you to forecast work on an individual work item level.  

You can calculate the rest of the percentiles in the same way. These percentiles are important because they show us how much risk we are willing to take. If we are fine being wrong 50% of the time, we will say that 2 days are needed to complete any card on this workflow. 

If we prefer to stay on the safe side and be more accurate in our forecast, then we will interpret the process using the 85% or 95% percentile, which gives a higher level of certainty. 

Practical Tips: 

• Keep the dots as close to the "x" axis as possible because that's how you decrease cycle time and deliver faster. 
• Pay attention if the dots tend to grow on the right.  If they do, it means that your cycle time is going up, and you are getting slower over time. Investigate the causes behind delayed tasks and take some actions to introduce process improvements. 
  

• You can mark the cards that have been blocked using the "Blocked items" control to the right of the chart. You can toggle on and off which cards have been blocked. It is a good visual indicator, which gives you an idea of how many cards have been blocked.