I am sure you’ve heard of various Agile metrics – cumulative flow, velocity and burndown just to name the most popular ones. However, when we talk about predicting how long a certain item is going to take and giving such information to your clients, most of these metrics are of little use. To help solve this issue, Daniel S. Vacanti has introduced Actionable Agile metrics. Which were focused on measuring the duration of production and suggesting ways to improve it. Curious to know more? Here we go!
Actionable Agile metrics were first introduced in 2015 when Daniel S. Vacanti wrote a book titled ‘Actionable Agile Metrics for Predictability‘. The main focus here was improving predictability in Agile projects and introducing metrics that could help provide these numbers. This was something the author felt Agile projects lacked previously and available metrics could not provide alone. Teams used burndown chart to try and keep the number of tasks to do going down steadily and observed team velocity to make sure the team keeps working in a similar pace.
However, just these metrics alone could not accurately tell when a new item is expected to be finished. In fact, obsessing over team velocity has even caused damage to teams in a way that they were no longer thinking what should be done. Instead, they were simply working on more and more items, without a clear goal.
Mr. Vacanti saw the error in such ways and proposed a new set of metrics to help Agile teams have more predictability on their projects.
Actionable Agile metrics
To get a better understanding of when items are going to be done, Mr. Vacanti chose one Agile metric to follow. To put it in his words – ‘If nothing else, track cycle time’.
For those of you not familiar with cycle time, this is a metric used to describe the amount of time from when the team starts working on an item, until it is finished. If you are using a Kanban or a similar Agile board, this would be the whole time from the moment item was pulled out of the backlog, until it was pulled in to the Done section.
Cycle time is nothing new to those familiar with Agile. In fact, the average of it has been used in combination with lead time (total time from the moment item was created until it was completed) to give a better understanding of how long tasks will take in Kanban projects. But Daniel chose to take this metric a little further in order to get better estimations and ways to improve delivery times for the team’s benefit.
New take on Cycle time
Since most processes are random, he suggested to create a scatterplot of all tasks the team has finished where one axis showed the date when a task was done and the other axis showed its cycle time. By observing this data he then drew lines on this scatterplot to display the 50th, 70th, 85th and 95th percentile.
With this information at hand, he would immediately be able to see that, for example: 50% of tasks took 6 days to be finished, 70% of tasks took 9 days to be finished, 85% of tasks took 15 days to be finished and 95% of tasks took 22 days to be finished. Slicing the cycle time data in this way gave a lot more insight into what delivery dates could be expected versus just observing the average cycle time.
As listed above the average cycle time is 6 days, meaning 50% of items will be finished between 0-5 days. But 85% of items will be finished between 0-15 days. So if you need to predict how long a new task will take, will you be more likely to go for the average or the 85th percentile? I would think the later is the better bet. After all, I don’t see anyone getting angry about an early delivery. Thus, this way of slicing data does provide more insight and is more useful for project managers.
Areas for improvement
Actionable Agile metrics are great for getting more insight into how your team works and providing more accurate predictions, but this is not all it does. It is also a great opportunity to analyze what is happening in the team and make improvements. The ideal case in your scatterplot are not large gaps between the percentiles, but instead little ones. In this case, you optimize the process and have more accurate predictions for the delivery time.
Imagine if the difference between your 50th and 85th percentile is just 1 day. In this case you have a very clear understanding of when most tasks will be delivered and can communicate that to your clients. So what are the ways to improve your actionable Agile metrics?
First up, you should instate a WIP limit. Nothing affects the cycle time as multitasking does. While you may tell yourself you’re great at it, the reality is you will get distracted and take more time overall to complete your items.
If you are already in a middle of a project, start by picking one task you are currently working on and focus on it until its finished. Once that is is done, do not go for a new task, but instead, finish another task you have already started and continue until everything that has been started is done. If you have just started a new project, instate a WIP limit that works for your team and keep to it. Instead of being all over the place, stick to one task per team member at a time.
This way you will be able to better focus on tasks at hand and finish them in a more timely manner. Which in turn will lessen the gaps between 50th, 85th and 95th percentile and will allow you to have a more predictable process.
Another tip on decreasing cycle time is to identify and eliminate any blockers. What I mean by blockers are tasks that have been in progress for a very long time and are no closer to the finish line. Mr. Vacanti called those items aging in progress and in fact these tasks are the ones that push up your 95th and 85th percentile up. More importantly, sometimes this happens for no reason other than that a team member has so many things going on, he or she simply forgot.
To improve your cycle time and get more predictability it is important for you to identify those items and engage your team into completing them. This may be as easy as asking someone to do it, or it may be a true issue that will require attention from several team members. Either way, once tasks like these are complete, your team ca start focusing on new stuff.
Actionable Agile metrics are a great way of evaluating your teams efforts, identifying the pain points and most importantly predicting when work is going to be finished. By applying such metrics you can take most of the guessing out of the question and rely on facts instead. This is why we have just included actionable Agile metrics reports in Teamhood.
Actionable Agile metrics in Teamhood
To help you navigate and improve project predictability, we felt it was important you get automated metrics in Teamhood. You do not have to calculate or estimate anything, as all the data from your projects is right there. So here is our take on actionable Agile metrics.
The first graph you will see on this page is Total progress displaying all planned, started and completed tasks in three pie charts based on count, hour estimations and point estimations. This will give you a general idea of project status and if you use estimations in hours or points, you can compare item completion vs estimation completion.
The current status
Following them you will see two graphs. First will show the percentage of tasks that have been completed on time in the last month as well as in the previous 3 months. Which will tell you if the due dates and predictions set are accurate or maybe in need of improvement. You can always expect some items to be late, but if only half is delivered on time, there may be a need to rethink your estimation accuracy.
Immediately next to it you will be able to see the possible cause of late delivery, a chart displaying items aging in progress. Here, we decided to focus on three percentiles – 70th, 80th and 95th. Giving you a look into the longest times tasks have been in progress and blocking new work on being finished.
The third graph on this report is time to complete or cycle time. Instead of a messy scatterplot diagram, we wanted to give you a clear visual representation of your data. So here again you will see 3 percentiles – 70th, 80th and 95th as well as how much time each of them take. You can also review the data from past three moths and compare on how the team has improved or declined.
The last graph on this report is for new vs completed items. Here you will see the data from last three moths showing how many new items were added and how many items were completed as well as their averages. For new projects it is natural for new items to exceed completed ones, but for those working in a mature environment, this number should be close.
List of things to do
Lastly, all of these graphs are fun and all, but not particularly straightforward on what should be done next. To make sure you eliminate blockers and work on tasks that are already overdue, Teamhood provides you with an actionable items list.
Here you will see two lists – one of overdue items and the second of items aging in progress. Giving you a clear plan of action where to start in order to improve your cycle time and lessen the gaps. And once you are done with those, do not forget to limit WIP and work on one or at most two items at a time to take the full advantage of predicting more accurately with the actionable Agile metrics.