The five-step Sprint Retrospective is a well-recognised and frequently used approach to performing a successful Sprint Retrospective. The Retromat website automates this pattern helping teams, coaches, and scrum masters to choose appropriate activities for each of the five steps. Other tools, including tools that facilitate remote retrospectives, embody the five-step pattern (or elements of it) as the basis of their automation.

The five steps are:

• Setting the stage - warming the audience up and gaining their engagement

• Gathering data - understanding the team's priorities and concerns for discussion

• Generating insights - discussing the items raised, using root cause analysis and other techniques to create deeper understanding

• Decide what to do - action planning so the team can be even better next Sprint

• Closing - ensuring that the Sprint Retrospective has a definite end

As with any commonly used technique or pattern, there is the opportunity to make assumptions about the quality of the performance of the retrospective. A pattern does not ensure a high quality retrospective it just helps make it more likely. To paraphrase George Box

All patterns are incomplete, but some are useful…

No Silver Bullet…

The five-step pattern does not address two common problems in Sprint Retrospectives. First, there is no mechanism to ensure that objective data is brought into the retrospective and reviewed as part of gathering data and generating insights. Teams should be encouraged to always review objective data about their sprint performance as a way of identifying topics that have not otherwise been raised, or of altering priorities. We should always strive for balance between data driven analysis and more qualitative analysis in our retrospectives.

Too many teams ignore metrical analysis and, consequently, fail to detect problems related to workflow. Arguably, this problem is exacerbated by tools such as Retromat, where the majority (possibly, all) the techniques presented for generating data and insights are qualitative rather than quantitative in nature.

Second, it does not ensure an equal emphasis on reinforcing positive outcomes as well as eliminating negative outcomes. Sprint Retrospectives are not just about resolving problems. They are also about sharing patterns of success across the team, so that these positive outcomes can be propagated and reinforced within the team.

In reality, these two problems are linked. When we make a change for the better as a result of our retrospection, how do we confirm that the change is real and not just perceived. The answer is through the use of data. To confirm a positive change in performance we need to know both the data before the change and the data afterwards.

… But A Firm Foundation

The five step pattern for retrospectives provides a firm foundation to help us achieve the right change in the team at the right time for the team. Its emphasis on gaining insight, then deciding what to do, based on that insight, is exactly right. The problem is that many of the tools that support the five step pattern don’t talk about the use of metrical data.

Even though we are using a powerful pattern, we still need to apply our full rigour and discipline. Only in this way will we achieve consistently effective retrospectives over an extended period of time. Our rigour and discipline tells us that data driven analysis is as important as qualitative analysis. Let’s make sure we use both approaches with equal enthusiasm.