Are your retrospectives actually moving the needle, or are you just having fun?
Summary
- Agile delivery teams aspire to continuously improve their performance in terms of delivery cycle-time, throughput and predictability. Achieving these goals contributes ultimately to business outcomes like reduced operating costs, increased revenues, and improved margins.
- A team’s delivery performance is a function of both their delivery capabilities and team dynamics. Thus both types of data (team health survey data and delivery performance metrics) are essential inputs for improvement.
Retrospectives play an essential role in steering teams towards improved delivery performance. They operate in a continuous feedback loop where teams review their challenges and achievements, and seek opportunities for improvement. Retrospectives follow a fairly consistent sequence of steps:
- Gather the team and the data. Before the actual retrospective event, ask the team to complete a brief survey on how they feel they are functioning as a team. Data from the survey can help pinpoint issues with team dynamics that may be adversely impacting their performance. This data can be tracked over time to help identify trends in team sentiment. The survey should be sent out with the retrospective invitation, but ask the team to complete it before the event. Results from the survey should be presented with the other data collected at the beginning of the retrospective.
The ideal data set to be reviewed as the first agenda item in the retrospective should be along the lines:- Actions from the last retrospective
- Data from most recent team survey
- Throughput Chart
- Done vs. Planned Chart
- Generate insights. Armed with this data as context, conduct the retrospective: The team creates cards on the retrospective board for what went well, and what could be improved. (5-10 minutes). The facilitator then leads the team through a discussion of the cards created, allowing further elaboration and context. Once this is done give the team a few minutes to vote on the cards. This will help establish the relative importance of the issues in the minds of the team.
- Decide what to do. The team brainstorms on actions needed to address the most important issues. Vote on the actions – see what everyone feels most strongly about. Decide which actions should be worked on in the next sprint.
- Close the retrospective. Get any feedback from the team on the retrospective itself. Save the retrospective board as a record for future reference.
Quantitative vs. Qualitative Data
Qualitative data is typically collected through surveys. Qualitative data can be subjective, and can be difficult to translate into improvements or may lead a team to pursue improvements based on personal opinion. Hence qualitative data should always be accompanied by objective data (for example Kanban Flow Metrics) on a team’s operational performance, which will be more objective and more actionable. Metrics to consider reviewing as part of a retrospective might include:
- Cycle Time (Time in days to complete backlog items – visualized using a Cycle Time Scatterplot or Cycle Time Histogram).
- Throughput (How many backlog items did we complete in the last sprint – visualized using a Cumulative Flow Diagram).
- Related to throughput is predictability – percent items completed (done) vs. planned. Or, did we meet our SLE (Service Level Expectation): percent items completed within a target cycle time, e.g. 85% sprint backlog items completed in 10 days or less.
Survey Data
Delivery Performance Data
We should have all of the relevant data to set the stage for the retrospective. This team has a fairly stable velocity and their predictability (SLE) has improved to around 85% on average – pretty solid. The questions might be:
- What can we do to make further improvements on our delivery throughput
- How are people feeling about the pace of work, and
- Do we need some team-building activities
Good tools can enable and support this model of improvement. iRetro supports both team surveys and agile retrospectives, and consolidates data from both into a single dashboard to help focus improvement initiatives.
