Making Release Retrospectives Strategic and Effective: Part 4

In Part 1 of this multi-part blog series, I presented the case for tying release retrospectives with strategic objectives and strategic metric, and outlined a 5-step process for defining strategic objectives to drive agile transition in an enterprise:

  • Step 1: Define strategic objectives and associated strategic metric. See Part 2 of this blog.
  • Step 2: Conduct periodic measurements to collect data to support the strategic metric. See Part 3 of this blog.
  • Step 3: Use release retrospectives to analyze the strategic metric data, and discuss likely causes for the issues revealed by the metric
  • Step 4: Develop appropriate action plan to address the issues revealed by the strategic metric
  • Step 5: Implement the action plan developed in step 4.

In this part, I will elaborate on Step 3.

Step 3: Use release retrospectives to analyze the strategic metric and likely causes for the issues revealed by the metric:  A release retrospective should be held at the end of a release cycle to review and analyze the strategic metric measurements, and likely causes for the issues revealed by the metric.  This is how you tie release retrospectives with strategic objectives and metric, and make release retrospective of strategic importance.    The release retrospective meeting or session is attended by all team members, ScrumMasters and Product Owners for the release cycle, and appropriate managers.   It is important that the strategic metric data are collected (as explained in Step 2 above – see Part 3 of this blog series) ahead of a release retrospective.  The data are reviewed and analyzed at the Release Retrospective meeting.

Table 3 shows examples of issues revealed by the strategic metric.  In a release retrospective, the participants should identify the issues revealed by the strategic metric, and start discussing the root causes that may be responsible for them.   This analysis of root causes may continue even after the release retrospective is over, as it may require more in-depth analysis.   If the root cause is not identified, at best only symptoms may be treated, and the issues will continue to recur.

Table 3: Analysis of the strategic metric and likely causes for the issues identified with the metric

Strategic Metric

Identify Issues and
perform root cause analysis

A. Field data on product feature usage by customers
  • Identify rarely used or least used features
  • Identify important missing features
  • Identify features that give poor user experience or are difficult to use
B. Concept to Customer value realization cycle time
  • Identify bottlenecks and causes of delays, and waiting periods and buffers in the existing end-to-end process from concept to customer value realization
C. Release cost data in financial terms
  • Break down the release cost for development and delivery into its component to identify the contribution of different cost components
    • Cost and opportunity cost represented by rarely used features
    • Cost represented by poor release productivity
    • Cost represented by poor level of automation along the end-to-end value chain (automated unit tests, automated QA tests, automated regression test, automated acceptance test, automated build process, lack of continuous integration, etc.)
    • Cost represented by personnel
    • Cost represented by equipment (Dev and QA environment, build machines
    • Cost represented by software licenses: Development tools, QA tools
D. Release Productivity = (Release velocity / Release cost) Example factors that may have reduced release productivity:

  • Poor performance of cross-functional teams
  • Low degree of automation
  • Lack of skills and training issues
  • Personnel issues
  • Multiplexing and multitasking
E. Number of customer-reported new issues in a given period of time
  • Identify issues that could have been prevented with rigorous design review
  • Identify issues that could have been prevented with rigorous code review
  • Identify issues that could have been prevented with rigorous unit testing  by developers
  • Identify issues that could have been prevented with rigorous QA testing  by QA testers
  • Identify issues that could have been prevented from more effective customer training and support
F. Productivity-Quality composite measure = Release Productivity / Number of customer-reported new issues
  • Does the metric suggest that the team is treating productivity and quality as two competing factors that must be traded off; therefore, any attempt to increase productivity invariably reduces quality, and vice versa?
  • Identify lack of any skills or training issues

For all issues identified and their root causes analyzed (as exemplified in Table 3), senior management should ask these questions for each strategic objective and the associated target goal:

  • Did we meet our target goal or did we fall short of it?
    • If we fall short on a target goal, by what amount?
    • If we fall short, what might be the root cause?
  • What trend are we seeing? Is the trend improving in the right direction or going in a wrong direction, i.e., deteriorating? What might be the root cause for the deterioration?

In Part 5 of this multi-part blog series, I will present the details of Step 4: Develop appropriate action plan to address the issues identified by the strategic metric, and Step 5: Implement the action plan developed in step 4.

Part 1: Overview of Release Retrospective
Part 2: Define strategic objectives and associated strategic metric
Part 3: Conduct periodic measurements to collect data to support the strategic metric
Part 5: Developing and Implementing the Action Plan

This entry was posted in Agile Adoption, Agile Benefits, Agile Development, Agile Management, Agile Metrics. Bookmark the permalink.

2 Responses to Making Release Retrospectives Strategic and Effective: Part 4

  1. Pingback: Making Release Retrospectives Strategic and Effective: Part 3 | The Agile Management Blog

  2. Pingback: Making Release Retrospectives Strategic and Eff...

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>