Evaluate Process Performance: Throughput Histogram
In the Kanban Method, throughput is the key measure of workflow efficiency. No matter how many tasks are being worked on, throughput only measures what has been completed.
While kanban throughput is commonly displayed as a bar chart, you can gain much better insights into your process by using a throughput histogram. Histograms display the frequency distribution of your data. Learning to read these charts enables you to track your workflow performance over time, determine your team capacity and assess your process efficiency.
The Kanban throughput histogram
Average throughput is one of the most accurate approaches to track workflow performance, and the preferred method of visualizing this metric over a period of time is using a throughput histogram.
Histograms show the shape and distribution of your data. The greatest benefit that comes with their use is the ability to monitor basic information about the dataset, such as the median value, the width of spread and the overall distribution of the assignments of your team.
Throughput histogram allows you to visualize how consistently your team is delivering results. On the horizontal axis is the throughput (number of tasks completed), while the vertical axis shows the frequency (number of days) that had a certain throughput. The vertical percentile lines indicate the probability that throughput will be repeated in the future – for example, using the graph above, there is a 50% chance that the future throughput of this team will be 5 items/day or fewer. There is an 85% chance that the throughput will be fewer than 7 items a day. These values can be used to define service level agreements.
Plotting a throughput histogram
Let’s look at plotting a histogram to show the throughput of a team over a month. Here is how the team performed in table form:
It is difficult to make sense of the numbers using such a simple table. One thing that does stand out is the two zero values cropping up every seven days – weekends. Let’s try displaying this data on a scatterplot:
This makes more sense – it’s easy to see how the team’s throughput has varied from day to day over the month. However, we can’t easily see the average throughput of our team or work out trends over time.
Let’s make a histogram of this data. First, count the number of days that have a certain throughput.
Let’s plot this data as a histogram diagram.
Now that we have a histogram, can you work out what useful conclusions and insights can be drawn from it?
How to read throughput histograms
Average throughput measures how much work has been done on average during a week or month, tracking how trends develop over time. We can also use this value to perform rough estimation and commit to how many items a team can deliver for a certain time period.
Our team’s average throughput over the month is the most important metric we want to discover. There are three ways to calculate an average – they can give the same result, but frequently have some differences and mean slightly different things.
One important thing to note about a typical throughput data set is the appearance of zero values due to non-work days – weekends, public holidays, vacations etc. This can make it seem like your team is less productive than they are!
However, it is useful to be aware of the average throughput calculated both with and without weekend values. Including the weekend lows gives you an accurate throughput value to calculate over longer periods of time, while leaving them out will give you a more accurate throughput average for the typical business day.
- Mode is the easiest average to calculate – it is the number that appears most often. In this case, a throughput of 0 is the mode. As we discussed above, these zero values are weekend days that don’t reflect a typical day of work. Removing the weekend values gives a mode of 3.
- Median shows the middle number of a data set. For example, the median of the following set of values (1, 2, 4, 7, 8) is 4. The median of this next set (3, 3, 4, 5, 5) is also 4. If the data set is an even number, the value between the two central numbers is taken. The median in the histogram above is 6 tasks per day.
- Mean is the average calculation that you are most likely to be familiar with. This involves adding up all of the values and dividing them by the number of instances in the data set. The total throughput of the entire month is 111 work items, with 30 instances (days) in the set – giving a mean of 3.7 tasks completed per day.
What is the point of calculating the different types of average? Depending on the range of the data set, different ones can be more likely to predict the central tendency of the set. This central tendency or central value reflects the actual average throughput of the team.
Average throughput is a great way to perform rough estimations. For more accurate long-term forecasts though we recommend using the Monte Carlo method.
Have you found histogram diagrams useful to visualize your throughput data? Which type of average best reflects the performance of your workflow? Tell us about your experience in the comments!
THE COMPLETE GUIDE TO
This guide will equip you with practical, proven methods of making reliable delivery commitments based on your past performance data.
Meet the Author
Sonya Siderova is a passionate product manager and a driving force behind Nave, a Kanban analytics suite that helps teams improve their delivery speed through data-driven decision making. When she's not catering to her two little ones, you might find Sonya absorbed in a good heavyweight boxing match or behind a screen crafting a new blog post.