# Leveraging Kanban Practices in Scrum for Sustainable Throughput
Written on
Chapter 1: Understanding Throughput
To aid our teams in achieving a sustainable work pace, it is essential to grasp what that pace entails. Throughput serves as an excellent starting point, as it is a straightforward metric to monitor. Let's delve into how to get started.
When a team consistently completes appropriately sized work items, we can evaluate how many of these tasks are finalized over a specified timeframe. This metric is known as throughput. Appropriately sized items typically refer to those that can be finished within a single sprint. If your team has established a Service Level Expectation (SLE) for each work item, then "right sized" can also mean "completed within your SLE."
It's important that work items are relatively similar in size, though not necessarily identical.
The eighth principle from the Agile Manifesto emphasizes:
Agile processes encourage sustainable development. Stakeholders, developers, and users should be able to maintain a steady pace indefinitely.
By monitoring our throughput and utilizing appropriately sized work items, we can foster sustainable development and uphold that pace continuously. This data-driven approach assists us in accurately predicting when a larger batch of tasks may be completed. Enhanced predictability fosters trust within our organization, allowing us to efficiently deliver what our customers desire, when they need it, while optimizing our resource usage.
Section 1.1: Clarifying Velocity vs. Throughput
For the context of this discussion:
Velocity = The total story points completed in a given timeframe.
The distinction between throughput and velocity lies in the fact that throughput focuses solely on the count of completed items, disregarding their sizes. In contrast, velocity aggregates the estimates of all work items.
I believe velocity isn't the most effective metric for establishing a sustainable pace for your team. It should primarily be used internally to facilitate meaningful discussions regarding the work involved in completing specific tasks.
When we estimate with story points, we typically do so when we possess the least knowledge—often during backlog refinement or when a stakeholder requests a timeline for project completion.
Subsection 1.1.1: The Challenge of Estimating
How often have you found that a task is more complex than initially anticipated once you begin? Had you known this beforehand, you could have provided a more accurate estimate. However, such knowledge often comes only after the work has commenced.
We usually make estimates when we have the least information available, and it’s common for our estimates to be incorrect despite our best intentions—they are essentially educated guesses.
Section 1.2: The Pressure to Adjust Estimates
Have you ever faced pressure to modify your estimates? Picture this: you and your team assess all necessary tasks, compile the totals, and report back to a stakeholder or product owner. If the total exceeds their expectations, they might question your estimates, suggesting they are too high.
Have you felt pressured by stakeholders or senior colleagues to lower your estimates simply because they seem excessive? This scenario is particularly common before work begins.
Estimating story points is inherently subjective. To cultivate a sustainable pace, it is advisable not to base decisions on such subjective data.
Chapter 2: The Importance of Actual Throughput
We require data that reflects completed work—what we call throughput. Once a work item is finalized, we gain complete insight into what it took to finish that task. This information is immutable and aids us in establishing a sustainable pace.
Throughput allows us to refine our forecasts. By counting the number of items accomplished over a specific period, we can estimate how long it might take to finish a larger project comprising multiple tasks. Data based on actual outcomes is difficult to contest.
For instance, if we finished ten tasks in three sprints, it is reasonable to expect similar completion rates in the following three sprints, provided all factors remain constant. This method eliminates the influence of subjective opinions on our estimates.
Collecting Throughput Data
Gathering throughput data is surprisingly simple. At the end of each sprint, total the work items in your "DONE" column. If you use a tool like JIRA, just check the completed items in your sprint report and note their count.
I maintain this data in a straightforward spreadsheet.
What to Do with Throughput Data
Throughput in Sprint Reviews:
If something is hindering our progress, transparency is crucial so we can seek assistance. Conversely, if improvements are evident, discussing the changes that contributed to this progress can benefit others.
Throughput in Sprint Retrospectives:
We can delve deeper into throughput during retrospectives by examining our data:
- What has changed?
- Is our current pace sustainable?
- What process enhancements could we implement to bolster sustainable throughput?
- Was this sprint an anomaly?
- What factors contributed to any slowdowns during this sprint?
Analyzing this data helps us determine our next steps.
Throughput in Sprint Planning:
Utilizing throughput data is vital during sprint planning. To uphold a predictable process, we should aim to align the number of work items entering our process with those exiting it. In simple terms, we should estimate completing a similar amount of work in each sprint as we did in the previous one.
Maintaining a steady pace in this manner will guide us towards a predictable workflow and enhance our ability to forecast completion dates with greater accuracy.
Summary
Collecting throughput data is straightforward and immensely valuable. As a lagging indicator, it gives us complete insight into actual performance, allowing us to make more accurate forecasts without adjusting figures based on subjective opinions.
This method is effective only if work items are similarly sized or "right sized." What constitutes "right sized" can differ among teams, but it promotes breaking down tasks into smaller units, leading to more opportunities for early customer feedback.
Such feedback enables us to adjust our approach based on informed insights, significantly increasing our chances of success.
Throughput is a critical metric to track for establishing a sustainable pace within your team. This pace becomes achievable when we start forecasting using data derived from actual performance.
We’re not aiming to boost throughput every sprint; rather, we will naturally improve as we learn to work more efficiently.
If you’ve begun exploring throughput and are seeking ways to enhance it, I recommend reading my article on work item aging next.
I hope this article aids you in your pursuit of the right metrics to support your teams, fostering a positive work environment and delivering genuine value to customers with each sprint.