June 25th, 2024

Enhancing Efficiency with Agile and Kanban in the Energy Sector

In early 2023, an energy company’s initiative faced significant challenges in improving delivery timelines and team productivity. The team had difficulties with inefficiencies in their workflow, resulting in missed deadlines and a lack of trust between business and IT departments. This was particularly problematic due to the high stakes involved in the energy sector where precise coordination and timely delivery are crucial.

Engagement Overview

Starting in March 2023, I worked closely with one of the company’s teams and leadership to adopt and implement Agile, with a strong focus on the Kanban approach. The primary goal was to streamline their processes, enhance predictability, and foster a culture of continuous improvement.

Initial Challenges

1. Lack of Trust: There was a noticeable lack of trust between the business and IT teams, which hindered effective collaboration and timely project delivery.

2. Inefficient Workflows: The existing workflows were not well-structured, leading to bottlenecks and delays.

3. Limited Visibility: There was insufficient visibility into the progress and status of work items, making it difficult to manage and predict project timelines. This lack of visibility extended to stakeholders, who did not have a clear understanding of the sources of demand being placed on the team, nor the team's overall capability to deliver work.

Approach and Solutions

1. Building Trust and Psychological Safety

Made adjustments to regular, transparent communication channels between business and IT.
Leveraged an agreed upon concept of ‘handshakes’ at the feature level to ensure clearer agreements on deliverables.
Initially, made adjustments to the Retrospectives being held to incorporate metrics and discuss using a flow-based approach. This made it less about what is ‘not going well’ and the format eliminated the potential for negative feedback that would further erode Psychological Safety.
Changed the stand up format to make it board centric, instead of focusing on the individual team members.

2. Workflow Optimization:

Mapped out the current workflows and identified key bottlenecks. Presented this to team leadership and once agreement was achieved, it was presented to the team and approved as the new way of working.
Implemented this new workflow in a new Kanban board in Azure DevOps to visualize work stages and enhance transparency.
Added a new work item type, “Technical Enablement,” to track backend implementations and maintenance tasks separately.

3. Metrics and Reporting:

Introduced lean metrics with the intent of using them as a better indicator of continuous improvement.
Focused on improving system lead times and throughput by emphasizing the completion of tasks over starting new ones.

4. Team Autonomy and Coaching:

Worked with development team lead ensure outcome-driven stand-ups and weekly meetings.
General coaching of individual team members on Kanban practices and the importance of flow efficiency.
Encouraged meetings that focused on finishing work on the board before starting new tasks.

Flow Data Analysis

Before Engagement (June 2022 - March 2023):

• Average System Lead Time (Cycle Time): 27.65 days

• 50th Percentile Lead Time: 21 days

• 85th Percentile Lead Time: 50 days

• Throughput: 19.40 work items/month

• The Cycle Time variability as shown on this scatter plot is very volatile

A natural question we can ask as we look at these graphs and numbers is: does this look like the metrics of a team that is - or isn’t - performing ‘well’?

It’s important to note that these numbers are neither ‘good’ nor ‘bad’, it’s just data. But it is always important to baseline the team’s capability to deliver, so as to have a visual of what the team was doing prior to the engagement and before introducing any changes.

There is another question we should at the very least ask: has this team already been performing at their peak? While usually as consultants we operate under the assumption that improvements can be made, one cannot rule it out entirely. It is important with any engagement to respect the context one is entering enough to at the very least entertain this possibility.

Trust-Building Phase (3/16/2023 to 9/30/2023):

As mentioned earlier, it took some time to earn at least enough trust from team leaders, stakeholders and executives to truly see results. Even so, we started to see some positives emerge from the Kanban Implementation six months into it. This was a period of time where an important development was taking place that would positively impact the team’s ability to deliver: a deployment pipeline.

• Average System Lead Time (Cycle Time): 27.25 days - a 1.4% improvement

• 50th Percentile Lead Time: 12 days - 43% improvement

• 85th Percentile Lead Time: 53 days - (-6%) change

• Throughput: 25.14 work items/month - a 30% improvement

• The Cycle Time variability as shown on this scatter plot continued being very volatile

Summary of results by October 2023

Improved Predictability:
- Enhanced ability to predict delivery timelines, with a focus on reducing variability and eliminating outliers.
- Better conversations between business and IT regarding the upstream process and sources of demand led to improved planning and execution.
Increased Efficiency:
- 1.4% reduction in average system lead times (cycle time).
- 30% increase in overall team throughput.
Enhanced Collaboration
- Improved communication between business and IT departments.
- Began to see a more collaborative environment.

It must be acknowledged that even though the overall numbers were better when looking at the averages, one could see there were a good amount of outliers that needed to finish making their way through the workflow and be resolved. The major improvement in the 50th percentile, while simultaneously seeing a negative trend in the 85th percentile proves this out.

Major improvements still needed to be seen in order to say that the implementation was a success and to even entertain the idea that this team could deliver predictable results, much less provide accurate estimates.

After Improvements Phase (10/1/2023 to 6/23/2024):

After some major improvements to the system and gaining trust with several key individuals inside and outside the team, we began to see some evidence of the changes that had been made to the workflow, the feedback loops and the communication between business and IT had begun to improve overall. There was also a better understanding from the business stakeholders about the importance of prioritization and using the capacity of the team carefully, as it is finite. Here are are the final metrics:

• Average System Lead Time (Cycle Time): 21.81 days - a 21% improvement from before the engagement

• 50th Percentile Lead Time: 13 days - 38% improvement from before the engagement

• 85th Percentile Lead Time: 42 days - 16% improvement from before the engagement

• Throughput: 27.56 work items/month - a 42% improvement from before the engagement

• The Cycle Time variability as shown on this scatter plot is now much tighter and exhibits less outliers, which lead to more a predictable system

Final Results by June 2024:

Sustained Efficiency Gains:
- Continued improvement in system lead times with an average of 21.81 days (21% improvement from before engagement).
- Sustained high throughput with 27.56 work items/month (42% improvement from before engagement).
Enhanced Predictability and Delivery:
- Consistent ability to meet 85th percentile delivery timelines at 42 days (16% improvement from before engagement).
- Improved 50th percentile system lead time to 13 days (38% improvement from before engagement), further increasing predictability.
Business Outcomes:
- Consistently using Monte Carlo simulations to predict timelines and commitments to customers, enhancing trust in the team’s system of delivery.
- Cost savings from improved workflow efficiency and better resource management.

Opportunities for further improvement:

Address High Outliers:
- Continue to focus on reducing variability in system lead times, especially in the 98th percentile range.
Sustain and Build on Gains:
- Maintain and further enhance transparency and collaboration between business and IT teams.
Expand Agile Practices:
- Explore additional Agile development practices that focus on a stable application delivery pipeline with automated quality gates.
- Further refine Kanban principles to continue optimizing the workflow, possibly leveraging the systematic implementation of the Kanban Maturity Model (KMM).
- There is still a need to understand the sources of demand, compared to team’s ability to deliver at higher level of the organization. This can be accomplished through a systematic implementation of Kanban at scale.

Conclusion:

This case study highlights how adopting Agile values and principles through the Kanban method, even in environments with limited trust, can lead to substantial improvements in workflow efficiency, predictability, and team collaboration. By focusing on transparency, trust-building, and continuous improvement, the team achieved desirable outcomes and enhanced their operational efficiency. Continued monitoring and process refinement will ensure sustained success and ongoing improvements in team performance.