By Mia Horrigan and Andrew Auld co-authors
We were working with a team that was a shared services team providing components to help other dependency areas deliver larger capabilities end to end. As a critical integration point, a number of other delivery areas were dependent on them for their delivery and the often they inherited the work from teams outside their functional area. They were always needing to “spill” work over as it wasn’t getting completed within the timeframes needed by the other dependency teams, lots of handovers and some items had been sitting in progress for over 3 months.
The leadership had wanted the team to start to work in a more agile way. There was a lot of resistance from within the teams as the nature of the shared services work was seen by some in the team, as “not suitable for agile”. We adopted an interactive approach and started to implement Kanban with Scrum to bring transparency to the work that was being done. With transparency came improved decision-making about the amount of work and where should the team focus at any given point in time.
Evidence Based Metrics is key to understanding success
We turned to Evidence Based Management (EBM) metrics to try and get insight into how we could improve their time to market key value area. Specially, we focused on reducing batch size as a leading indicator and our assumption was that this would result in improvements to lead time and cycle time. Smaller batch sizes would help getter better predictability and accuracy in their estimations and ensure work can be completed within a Sprint (two weeks) rather than having a lot of spill over work as was previously the case).
Reduce batch size to improve time to market
Over the course of last year, the average lead time for the team was 52 days. Since implementing Kanban with Scrum over a 3-month period, the average lead time dropped to 17 days! This represented a 67% improvement. As highlighted in the graph below, the cycle time reflects the time the PBI (product backlog item) was first committed to when it was done.
The total average cycle has almost halved since the team started Kanban with Scrum (Red Line). The overall line is trending down, there are more “dots” meaning the work is being broken down and evaluated and our forecast predictability has improved in that if a PBI is brought in to committed likely to be completed within 10 days cycle time.
What did teams do to make cycle time go down?
Using Kanban required the team to actively and continuously address batch size, transparency, the rules they used for progressing work, and collecting EBM metrics, particularly around lead time.
So what’s next?
The goal is to continue to improve and get the lead time down as low as possible through small continuous improvements. By focusing on these indicators, we are starting to see results and value delivered earlier.
Conclusions
Reducing batch size and moving to a Kanban approach , has given the team a greater focus on work age and getting things to ”Done” as well as limiting work in progress to ensure work flows better through the Sprint. The team also have much better transparency and visualisation of the prioritised Product Backlog and on what they are actively working on. The combination of these factors – using EBM, Kanban with Scrum – have helped them see this great improvement and will ultimately help them getting better predictability and accuracy in their estimations.
