How the Auto-scheduler handles capacity for Kanban teams

[#1]

This article focuses on how the Auto-scheduler in Kanban Plans handles capacity based on the capacity distribution algorithm. For a more holistic view of the Auto-scheduler, please see the following resources:

• How does the Auto-scheduler in my plan work? 
• Capacity in scrum vs. kanban teams in your plan 

---

Advanced Planning handles capacity based on the capacity distribution algorithm. For Kanban teams, this means distributing work evenly on a weekly basis. While it excels in preventing overcommitment and ensuring resources are not underutilized, its approach to task distribution can sometimes seem enigmatic.

This article demystifies the auto-scheduler's logic in distributing workload and the different scenarios users can expect based on their Issue estimation and Capacity weekly hours.

Scenario 1: High capacity & low workload

Let's imagine in this scenario, your Teams Capacity is set to 25 weekly hours and all of your Issues Estimates (h) are equally set to 2 hours. In these conditions. For an even distribution of work, you would have to calculate "25 weekly hours / 5 working days = 5 hours per day".

Under these conditions, since all Issues are estimated at 2 hours, using up all of the available capacity would not be an equal split for the week which is why you can expect the following example where 2 Issues per day are allocated:

Mon: 2 + 2 = 4h
Tue: 2 + 2 = 4h
Wed: 2 + 2 = 4h
Thu: 2 + 2 = 4h
Fri: 2 + 2 = 4h

Total: 20/25h

With the 5 hours remaining, if it were to be allocated, this means that there will be 1 day where more work is being done, hence no longer being an even split of workload for the week.

Scenario 2: Capacity & precise workload

Let's imagine in this scenario, your Teams Capacity is set to 25 weekly hours and all of your Issues Estimates (h) are equally set to 5 hours.

This is the best-case scenario and ideal conditions which doesn't always happen. This assumes that you've estimated your Issues with the perfect workload that aligns with your weekly capacity and allows you to fully utilize your Team's Capacity with an evenly distributed workload. With the ideal estimation of 5 hours per Issue, you can expect the following example:

Mon: 5h
Tue: 5h
Wed: 5h
Thu: 5h
Fri: 5h

Total: 25/25h

Because your Issues are "precisely" estimated in a way that allows for even distribution, you should able to utilize the full Teams Capacity for the week and still get an even workload per day.

Scenario 3: Low capacity & high workload

Let's imagine in this scenario, your Teams Capacity is set to 25 weekly hours and all of your Issues Estimates (h) are equally set to 7 hours.

Under these conditions, your Issues are estimated too high in comparison with your Team's Capacity weekly hours. The auto-scheduler will always try to avoid overbooking or going over your capacity and as a result, you can experience unused capacity due to the fact that no more work can be allocated to that week without going over the limit.

Pair that with trying to evenly split the workload, you might get a combination of remaining capacity + work carrying over to another day. In this example of overestimating Issues, we can expect the following:

Mon-Tue: 7h
Wed-Thu: 7h
Fri-Mon: 7h

Total: 21/25h

From Friday, it is not possible to allocate another Issue without going over capacity hence why the workload is carried over to Monday (weekend excluded). In addition to carrying over the workload, because the auto-scheduler always tries to distribute the workload evenly for the entire week, we can also see the following week being affected by the work carried over, resulting in a higher unused capacity.