Throughput is the total number of items completed per unit of time. You might have a throughput of 2 per day, 10 per week, or even 17 per sprint. Whatever your preferred time time unit, this flow metric helps you understand how quickly you finish work. That understanding is critical for forecasting how long it will take to complete a collection of work items.
How do you calculate Throughput?
There are two things you need to define to calculate throughput:
Your Finish Line – or the point in your process that items are considered complete
Your Time Unit – day, week, month, etc.
Defining the Finish Line
In order to define a finish line, you have to understand your process. A Kanban board is a great way to do visualize your process and make sure that it is clearly understood by all. Take a look at the board below:
Kanban board with clearly marked “Finish Line”
This team has defined the Done column as their finish line. So, any items that move into the Done column are counted as Throughput.
Choosing a Time Unit
You can use any time unit you desire to measure throughput. You can measure per day, per week, per month, per Sprint – you get the idea. If you aren't sure, use day as that is easy to measure and other types of time units are made up of days (unless you need to get smaller than that).
Why should I care about Throughput?
Looking at your Throughput allows us to analyze how consistently you deliver value. Consistency of throughput, and how it compares to the rate at which you start work, is one indicator of how stable your process is.
Perhaps the most common use for the Throughput metric is providing forecasts for completing multiple work items. You can use Cycle Time to forecast for single items, but you need a rate metric like Throughput to provide forecasts for groups of work items.
How do you use Throughput to forecast?
Traditionally, people use their Throughput to determine an average rate at which work is finished and then divide the total work by that average. However, forecasting based on averages will produce average results. Obviously, we don't suggest you do that.
Fortunately, you can use a Monte Carlo simulations that can use your Throughput data to simulate probable outcomes based on the variation found there. It’s a much more accurate, not to mention risk-aware, way to deliver forecasts. Read more about Monte Carlo simulations and forecasting.
Getting started
Teams often start looking at Throughput around the same time they begin looking at Cycle Time, WIP, and WIP Age. Focusing on building stability in these key flow metrics is a good start. The more stable your basic metrics are, the fewer outliers your forecasts have to account for and the more your forecasts are perceived as acceptable and, most importantly, accurate.
Interested in tracking flow metrics like this one? Try out ActionableAgile for free and reach out if you’re interested in joining our customer success program!