Mastering Predictable Feature Delivery on Live Service Games- Part 3
Improving workflow predictability with service level expectations by setting clear, data-driven timelines.
Deep Drive podcast generated with Notebook LM.
Learn how Service Level Expectations improve workflow predictability by setting clear, data-driven timeframes for task completion.
Discover how SLEs integrate with Scrum and Kanban to boost planning accuracy, optimise WIP limits, and support hybrid project management approaches.
Understand the challenges of SLEs, from managing stakeholder expectations to avoiding oversimplification and how to use them as tools for continuous improvement.
Welcome back to part three!
In parts one and two, we looked at building a structured workflow to improve flow, predictability, and metrics to quantify team performance. Based on the work by Daniel S. Vacanti, author of Actionable Agile Metrics for Predictability: An Introduction, Actionable Agile Metrics for Predictability Volume 2 and co-author of Flow Metrics for Scrum Teams, this guide will walk you through designing and managing an effective workflow to enhance your team’s ability to deliver value consistently.
Part three discusses the next step: using Service Level Expectations to enhance transparency and focus on high-priority issues. Let’s get started!
We will break this down into four parts:
Using Service Level Expectations (this edition)
Continuous Improvement
Using Service Level Expectations
Service Level Expectations (SLEs) are essential for setting performance benchmarks and defining clear team expectations. An SLE estimates how long a work item should take from start to finish, typically expressed with a confidence level and a time interval, such as “75% of our features are completed within 45 days.”
When teams receive a new work item, they should right-size it appropriately and strive to meet or surpass the SLE timeframe. This method helps balance speed and quality while providing clear performance metrics for stakeholders.
The term “expectation” is crucial in understanding SLEs. They aren’t commitments or guarantees. While it might seem that SLEs primarily benefit stakeholders, they are fundamentally meant for internal use. Development teams employ these expectations to enhance transparency, conduct inspections, and facilitate adaptations. They allow teams to evaluate ongoing work against their SLEs and pinpoint work items that may not meet the expected timeframe.
As work items linger unfinished, the likelihood of not meeting the SLE increases. For example, if the age of a work item surpasses the median completion time of the team’s tasks, it suggests a greater risk of delay. This risk becomes more evident as more time passes without completion.
Visualising these at-risk items allows teams to focus on high-priority issues during the Daily Scrum, aiding in tactical inspections and adaptations. This proactive management helps maintain consistent team performance.
Integrating SLEs with Project Management Frameworks
Cross-Methodology Synergies
Service Level Expectations (SLEs) can be seamlessly integrated with popular project management frameworks such as Scrum and Kanban, significantly enhancing the predictability and efficiency of project outcomes. Here’s how SLEs can bolster these frameworks:
Scrum Integration
In a Scrum setting, SLEs can be leveraged to enhance sprint planning and review processes by providing a data-driven foundation for estimating and adapting task durations:
Sprint Planning: Use SLE data to set realistic sprint goals. For example, suppose your SLE suggests that 85% of tasks are completed within ten days. This insight can guide the team in selecting and committing to work items achievable within the sprint’s timeframe, enhancing predictability in delivery.
Sprint Retrospective: Assess the accuracy of SLEs during retrospectives by comparing the planned task durations against actual completion times. This analysis aids in identifying trends or anomalies, fostering discussions on process enhancements and strategic adjustments for future sprints to improve predictability.
Kanban Integration
Kanban teams can incorporate SLEs to manage workflow more effectively and Work Progress (WIP) limits, ensuring a consistent flow and timely completion of tasks:
WIP Limits: Align WIP limits with SLE data to optimise the load across the team. If SLE data indicates that tasks are taking longer than expected, adjusting WIP limits downward can help mitigate bottlenecks and maintain focus, thus stabilising workflow and enhancing predictability.
Flow Metrics: Integrate SLEs with Kanban metrics such as cycle time and work item age. This combination offers a detailed view of task durations from inception to completion, assisting in refining processes and setting more precise expectations for task completion times.
Cross-Methodology Synergies
Teams that employ a hybrid approach mixing elements of both Scrum and Kanban can use SLEs as a unifying metric to:
Harmonise Expectations: By establishing standard benchmarks for task completion times, SLEs ensure that all team members have aligned expectations, irrespective of the specific methodologies in use, enhancing predictability across the board.
Performance Tracking: Integrating SLEs with Scrum’s velocity and Kanban’s throughput enables teams to monitor performance against qualitative and quantitative targets, providing a comprehensive perspective for evaluating project health and team efficiency.
Integrating SLEs with these methodologies optimises workflows and significantly enhances the team’s ability to forecast and meet deadlines, delivering high-quality work with greater consistency. This approach ensures that project management is disciplined, transparent, and thoroughly data-informed, directly contributing to improved project predictability.
Calculate Your SLE
Determine Reasonable Performance Benchmarks
Let’s say you want to predict your team’s time to develop a LiveOps event.
For this, you can utilise a tool known as a “Cycle Time Scatterplot,” which offers a snapshot of your team’s previous performance by showing the completion times of past work items. This chart also lets you see different percentiles (50%, 70%, 85%, 95%) representing various potential delivery times.
From a typical Scatterplot, you might infer that there’s an 85% chance your team will complete future requests within 15 days or less. There’s also a 50% chance that tasks will be finished in 7 days or less.
To set up a practical Service Level Expectation (SLE), it’s helpful first to analyse historical cycle time data to gauge your team’s performance trends:
Cycle Time Analysis: Use a Cycle Time Scatterplot to review how long work items typically take to complete.
Percentile Selection: Select a percentile that reflects an acceptable risk level, such as the 85th percentile. This will be your SLE, indicating that you expect 85% of work items to be completed within this timeframe.
Calculating your SLE from past data helps ensure it’s realistic and attainable, laying the foundation for predictable workflow management.
Working with SLEs
Using SLEs in day-to-day operations
SLEs are a critical component of effective project management. They offer a structured approach to forecasting task completion times and improving delivery accuracy. By integrating SLEs into your project management processes, you can enhance predictability, align team efforts, and ensure high quality and efficiency. Below are some key strategies for working effectively with SLEs throughout your project lifecycle.
Right-sizing Work Items in Sprint Planning and Managing Delays
During sprint planning, aligning the size and complexity of work items with your Service Level Expectations (SLE) is crucial. This involves breaking down large or complex tasks into smaller, more manageable units more likely to be completed within the SLE timeframe. Continuously refining and adjusting work items that may exceed the SLE helps improve predictability and ensures the team can deliver features on time without compromising quality.
Using Percentiles as Intervention Triggers
To proactively address potential delays, integrate your SLE percentiles into your workflow to identify when work items are at risk of exceeding expected durations. You can visualise when a task is slipping beyond its scheduled completion time by overlaying the same percentile lines from your Cycle Time Scatterplot onto your Work In Progress (WIP) Ageing Chart. When a work item’s age reaches a specific percentile line, trigger interventions such as pairing (two developers working together), swarming (the entire team focusing on one task), or removing blockers. This proactive approach ensures that potential delays are addressed before they escalate, maintaining the flow and predictability of your workflow.
Regular Review and Adjustment of Your SLE
Service Level Expectations are not static; they should evolve based on the latest data and team performance to remain effective. Use sprint retrospectives to review flow data and assess the accuracy of your current SLEs. Based on recent performance trends and insights from your Cycle Time Scatterplot, make necessary adjustments to the SLEs. Regularly reviewing and adjusting your SLE ensures that it remains relevant and aligned with your team’s capabilities and requirements, enhancing overall project management and delivery outcomes.
Challenges with SLEs
Often criticised for creating unrealistic expectations and oversimplifying tasks.
While SLEs are commonly employed in project management frameworks, they have faced various criticisms. One major critique is the potential for misuse. For instance, SLEs can engender a false sense of certainty in the inherently uncertain realm of knowledge work, possibly leading to unrealistic stakeholder expectations. Additionally, there’s a risk that teams might prioritise adhering to SLE metrics over the fundamental goal of delivering value, an issue akin to the criticisms directed at Story Points in Scrum.
Another set of challenges relates to implementation. Setting effective and meaningful SLEs can prove difficult, especially for teams that are either new to project management methodologies, engaged in projects with diverse characteristics or do not have the appropriate software. There’s also the risk of oversimplification, where SLEs might reduce complex work processes to overly simplistic metrics, potentially misrepresenting the actual work involved.
Cultural and organisational issues also play a role. Some organisational cultures may resist increased transparency and accountability, which come hand-in-hand with SLE implementation. Additionally, if the focus is improperly managed, not meeting SLEs could foster a blame culture rather than encouraging continuous improvement.
Moreover, SLEs may not be suitable for all types of work. They tend to be less effective for highly creative or unpredictable tasks and might even clash with Agile principles, particularly the emphasis on customer collaboration over contract negotiation.
To mitigate these issues, it is crucial to view SLEs as tools for improvement and learning rather than rigid obligations. Teams should embrace SLEs’ probabilistic nature and integrate them with project management practices to promote continuous improvement and deliver real value.
Conclusion of Part Three
As we wrap up our discussion on Service Level Expectations, it’s evident that while SLEs can streamline processes and set clear performance metrics, they also present challenges that need strategic oversight to avoid unrealistic expectations.
Our focus on enhancing predictability continues in the next edition, “Continuous Improvement,” where we will explore how iterative refinements and proactive adjustments can improve predictability in live operations.
I just listened to this "podcast" episode. Damn, that's a good thing to have :D The only thing I'm worried about is that it omitted something.