Teams seeking increased productivity and transparency are no strangers to Kanban. With the help of Kanban boards, they are visualizing and optimizing the workflows to get better results. When it comes to measuring the effort of a Kanban team, however, there is some doubt about just how this should be achieved.
To help you out, we have decided to discuss the 4 main Kanban metrics. And give you some examples of how they can be applied and utilized in real-world scenarios.
The 4 Kanban metrics are:
Teams or companies may use different names to mark these values. As long as the calculations are kept the same, this is no issue. It is not the name, but the calculation itself that brings value for the Kanban teams.
So, to better understand why these metrics are tracked, let’s take a closer look at each of them.
Definition: WIP represents the number of items that are currently ‘In Progress’.
While hearing WIP your first instinct may be to refer to WIP limits, Work In Progress alone stands not for the limit of the items, but the actual number of ‘In Progress’ items on your Kanban board. Having said that, the goal of any team should be to keep the WIP number lower or equal to that of their WIP limit.
💡 If done right – lower work in progress can result in higher troughput. As you see, Kanban metrics are related to each other.
Continuously tracking the WIP of your team can help with:
Let’s say you manage a team of 7 engineers and want to ensure they are moving tasks across the workflow as quickly as possible. For this, you may want to look into their WIP. Since there are 7 people on the team, the most effective way of working would be to have 7 tasks in the WIP – one for each of the team members. If you see a slightly higher WIP, this may be a cause of dependencies holding some tasks back. However, if the WIP is double or triple the number of team members, this is cause for concern as the team members may be multitasking.
Definition: Throughput provides us with an understanding of how many items a team has finished during a defined period of time.
In other words, it allows us to understand the pace of work as well as how it compares to the previous periods. As a result, we get a better grasp of the team velocity metrics and how much work can the team handle going forward.
When evaluating throughput, you should be wary of the work item size in the compared periods. This metric will be easiest to follow if your work items are relatively the same size or if you are comparing longer time periods. If this is not the case, take that information into account and apply probabilistic forecasting methods.
The period for which to calculate throughput depends on the process of each team. To achieve the best results, you should aim to compare the results for periods of equal value. For example, each month.
Tracking Throughput can help you with:
Let’s say you run a manufacturing company where you accept client orders. A new order comes in and you have to forecast how long it will take for you to deliver the request. By looking at throughput, you know how many requests your team can deliver in a certain amount of time. With that, and the length of the waiting list, you can accurately forecast how long the order will take to be delivered.
Definition: Work item age provides us with the information for how long the work item has been worked on already.
As such, work item age is a powerful Kanban metric that can help us identify bottlenecks as well as largely complicated items on the Kanban board. Allowing the whole team to pull together to help move it along the workflow.
Another term commonly used with work item age is items aging in progress. As you can probably tell, this term refers to those items that have been stuck ‘In progress’ longer than the average cycle time of the team. Having a list of such items across your projects can help with identifying the next course of action and priority steps.
Tracking work item age will help with:
Your software development team seems to be working well, they deliver work items each week, but you notice some items get stuck ‘In Progress’ for longer than others. Due to this, some of the feature development gets delayed and starts lagging behind schedule. By tracking the work item age, you will quickly be able to see which of the items seem to be getting stuck and identify the possible issues. They may be too big, too complicated, or maybe there is a lack of experience in the team to handle them. This way, you will be able to identify and deal with issues quickly.
Definition: Cycle time refers to how long each item has spent ‘In progress’ before it was completed.
Cycle time is similar to work item age, however, this calculation is only done after the item has been completed. Just like the previous metric, it is also calculated separatelly for each work item. However, it is rarely used as such for analysis purposes.
When analyzing performance, you may want to calculate the average cycle time for the team and then compare it against other time periods or teams. This is one way to use cycle time. The second and more preferable option is to calculate cycle time percentiles. What this means is taking the cycle time data, sorting it from shortest to longest period and then selecting X% amount for accuracy, where the most lengthy cycle time defines the final number. So if the team has finished 3 items in 3, 5 and 9 days. It’s cycle time for the 70th percentile will be 5 days.
For example, you could calculate how much time it took your team to complete 50% of all items, how much time it took to complete 75% of all items, and how much time it took for 95% of all items. This will give you cycle times for the 50th, 75th, and 95th percentiles. Then you can use the calculations to create more accurate forecasts of your work going forward. You will select percentile based on how accurate you need to be.
Tracking cycle time allows the team:
Let’s say you work with client projects and an urgent request comes in. They also want to know how fast you can deliver on it. Cycle time gives us the calculation of the touch time the team has before they can deliver a result. Thus it is a perfect reference when you want to give an accurate forecast. By knowing that the complexity of the request is very low, that 95% of your items have a cycle time of 10 days, and that 80% of items – 5 days. You can safely forecast that the request will take around 5 days to be completed once the team starts working.
All of the 4 Kanban metrics above are useful on their own, however, they are truly powerful when combined. The team can benefit from combining metrics from the WIP, throughput, and cycle time for example to forecast when each item on your Kanban board will be completed.
In fact, we at Teamhood took a stab at implementing such calculations and providing our users with a forecasting layer. We also named it ‘When will it be done’ to give you more clarity on what it does.
By activating this layer on your Kanban board, you will be able to see the forecasts for each item based on your previous performance. To give you more control over the forecasting, you can choose the period of days to be used for the forecast and the accuracy you wish for.
This will give you one more tool to understand and analyze the team’s effort with Kanban metrics.
Want to give it a try? Explore the Teamhood Kanban solution in more detail.