BT

Facilitating the Spread of Knowledge and Innovation in Professional Software Development

Write for InfoQ

Topics

Choose your language

InfoQ Homepage Articles Retrospectives are Weak - Here is How to Make Them Stronger

Retrospectives are Weak - Here is How to Make Them Stronger

Leia em Português

Bookmarks

Key Takeaways

  • Understanding why organisations settle for mediocrity in retrospectives
  • The dangers and opportunities in challenging how retrospectives are normally run
  • Some straightforward techniques and processes that can be applied immediately in retrospectives to improve trust and build a interdependent team
  • How to recognize the key skills and behaviours of high performing teams and how to coach them
  • The four steps to transform everyday retrospectives in order to precipitate real change and improvement

The idea and practice of frequent retrospectives at every level of the organization is one of the most important ideas introduced by the agile movement into organizations. Instead of once-a-year annual performance reviews for people, organizations which are actively increasing their agility have team retrospectives, group retrospectives, and enterprise retrospectives on a weekly, monthly, or quarterly basis. We would suggest that organizations who want to evolve and change quickly should review their work systems on a regular basis.  

But to really reap the benefits retrospectives need to be conducted at depth and with a level of honesty and integrity that many organisations and teams find difficult to handle.   

While the move from infrequent to more frequent self examination has led to a significant shift in company behavior and increased business value, it is now time to go deeper in our practice of retrospectives. Let’s begin by acknowledging that the way that we practice retrospectives today leaves room for improvement:

  • Ron Jeffries and others have lamented the fact that Scrum teams do not appear to drive significant improvements in their engineering practices even though they hold frequent retrospectives. This has led to the charge that team-level Scrum retrospectives are often ‘shallow’ and focus only on surface issues.
  • Many common retrospective practices, such as the “What went well/What needs to be improved” sticky note exercise, are subject to significant cognitive biases and limitations which reduce their utility. They depend on the participants having a bias-free memory of events which they have observed or participated in, an assumption that very rarely holds true.
  • Too often retrospective practices do not clearly distinguish between changes in process and changes in outcome. Both authors of this article of have seen teams leave a retrospective with an outcome goal (e.g., “Improve quality”) with no agreement regarding the policy, process, and people changes that are going to be made in an attempt to reach that outcome.

The agile community has made a great contribution to organizational improvement by introducing the idea of retrospectives. It is now high time to refine our skills to make retrospectives more powerful.

Solution: Are you ready for Radical Retrospectives?

Perfection is the Goal

We would argue there is a paucity of ambition about what can really be achieved if retrospectives are run at depth with a culture of honesty and integrity (see the case study below of what can be achieved if effective team coaching is used in a situation where mediocrity has come to be tolerated).  

When we examine organizations which are genuinely extraordinary we see that they strive for perfection and that they want to do things that have never been done before. For example:

  • Morningstar’s vision is to “be an Olympic Gold Medal performer in the tomato products industry.” and for “Morning Star colleagues to develop a clear vision of perfect results” (italics added).

Note that Toyota’s vision is not to improve compared to its previous performance or to be better than its competitors: it is seeking perfection in the performance characteristics which it has decided matter most. Similarly, Morningstar wants “perfect results,” not better results.

None of this can be achieved if the skills, attitude and beliefs needed to deal with the inevitable obstacles are not in place.  Once the basic building blocks are in place teams can go beyond ‘surface agreement’ and dive into ‘what is really going on’.  Here is a short checklist for how to take a ‘deep dive’ and transform your retrospectives.

Building the Foundations

Many organisation accept superficial agreement and mediocre performance within teams for good reason - it is a lot less effort to do so than to challenge the status quo or face what is really going on.  To face up to dealing with conflict and tell the truth about underperformance is not easy but for organizations who want to transform their retrospectives and are willing to grasp the challenge the results can be spectacular.   

Here is a four step sequence for effective retrospectives: 

  1. Build trust.
    • If the leader wants to build trust with his/her team members the most impactful thing they can do is model the behaviors they expect from others.  This is easy to say but not often easy to do.  
    • If the leader expects team members to own their mistakes they have to be willing to own up to their mistakes.  If the leader wants to establish a culture where it is acceptable to make mistakes and learn from them the leader has to show that making mistakes will not lead to retribution. 
    • The agile community has developed many trust building exercises and approaches. One that we like is called the “Anti-Trust Exercise.” A coach works with a group to define all the activities which destroy trust such as:
      • Encouraging gossip and personal attacks about everybody but particularly the leader
      • Making sure important information is kept to as few people as possible. Copy inconsequential information to everyone
      • Creating confusion as to the purpose of the team.
      • Not telling the group how success will be judged.  
      • Rebuking team members frequently and in public

Once a team understand how detrimental these ‘anti-trust’ activities can be then it becomes far easier to understand how the opposite behaviors can build trust.  As part of the process the team has to agree on how it will behave if a team member is not keeping to those standards. 

  1. Build clarity
    • An important way to help an team increase clarity is to conduct a short survey which reflects the state of the team to itself. This short survey might include statements such as:
      - I can describe the purpose of my team in a meaningful way
      - I agree that team members tell the truth to each other
      - Team members are clear about how decisions are made.
      - I know how the success of my team will be measured
      - I agree that team members hold each other accountable
      - I believe that team members trust each other
      - I would choose to be part of this team if I had a free choice
    • Another great way to build clarity is to share ‘micro-communication’ techniques such as verbally recapping what has been agreed to and what the action items are. This simple technique ensures that everyone has shared understanding and increases accountability.
  2. Knowledge and skills injection
    • Once trust and clarity are at high enough levels, the organization and the people in the organization can learn new things: How does a software development team create a devops pipeline? How can value be evaluated quickly and at low cost?
    • As part of this step it is often necessary for teams to develop new skills which they can use to deal with the conflict and confusion that often arise when the reality of how the team has been operating is uncovered.  
    • While this step may appear trivial, it is usually not because it requires an organization to budget time for improvement instead of being solely dedicated to deliver.
  3. Genuine experimentation
    • New knowledge triggers changes in behavior which in turn changes results. Having a strong experimentation framework which captures changes in knowledge, behavior, and results is critical.
    • One simple way to start is to create a simple experimentation worksheet which captures a baseline metric, a hypothesis, and a date by when the experiment will be evaluated.

Once these building blocks are in place, then retrospectives can become transformative instead of merely corrective. 

Team case study

In this section we describe a team case study in which we applied the concept of ‘strong’ retrospectives in a way that helped a team in the food industry.

Context

This intervention happened in 2016. A food manufacturing company had created an international remote team of 11 who had responsibilities for product and process implementation. Each team member was based at a different site in Europe. Because of the distance issues they did not see each other often but used phone and Skype to keep in contact.

Each member of the team was dealing with the same type of issues and facing similar problems with stakeholders but in different locations. The team was set up so that they could learn and help from each other and support the wider company.

Problems: Why the coach became involved.

On the surface the team was working well enough. However underneath the superficial cohesion the sense of ‘belonging’ to their local site was stronger than the sense of belonging to the team. This meant they often prioritised the local issues over regional or global issues and were not giving honest feedback to each other. They also did not share resources with the wider team. In addition some team members were promoting their own local staff who did not have the necessary skills. The work the team had to do together was usually late and not to a great standard. Three of the team members had left in the last six months.

A new boss was appointed and he took a fresh approach. He identified the big issues and decided to bring in a coach to work with the team for an extended period of time to deal with the underlying issues.

Solution: What the coach did

After confirming the new boss’s diagnosis the new coach developed a program to address the key issues. She used the four step sequence that we describe above.

Step 1: Build trust -  Individual interviews with team members and stakeholders 

During the individual interviews the coach took time to build trust with individual team members demonstrating she would not punish them for telling the truth but she was not going to accept the ‘elephant in the room’ not being named.

Step 2: Build clarity - Gallup Q12 engagement survey

This revealed the system to itself and provided a baseline metric for experimentation (step 4).

Step 3: Knowledge and skills injection - Three workshops with the whole team and their boss over eight months

In the first workshop she worked on:

  • Helping the team define and record their values.
  • Using an appreciative enquiry approach to explore how the team could perform if they were being their best selves.
  • Drawing up a team charter about how they would treat each other.

She worked on a number of specific skills to help the team face up to the reality of the situation and become truly high performing.  How to:

  • Give positive and then developmental feedback.
  • Build trust between team members.
  • Build interdependency between members.
  • Deal with conflict constructively.

As a result of the workshop the team came up with specific commitments such as:

  • Gaining clarity on the functional priorities.
  • Committing towards the key collaborative platforms.
  • Creating an environment where everybody was valued and everybody was heard.
  • Ensuring that collaboration was visible at the level of behaviours.
  • Moving from a team of managers to a team of leaders.
  • Creating tangible strategic objectives that they could only achieve when they collaborated together.
  • Developing a shared understanding where they were going together as a team.
  • Daring to have crucial conversations and agree actions and not being afraid to ‘put the elephants in the room’.

The process was not always smooth. There were a number of occasions where team members were intensively challenged by the process, especially giving feedback when they were more used to ‘letting things go’. As part of the programme the coach conducted a number of one to one sessions between the workshops to deal with individual needs and specific situations. One team member left 3 months into the programme.

Step 4: Genuine Experimentation - Gallup Survey

After an eight month programme the score on the Gallup Q12 survey had gone from only 30% to team members rated as engaged to 75% rated as engaged. Anecdotal evidence from stakeholders suggested that the team was subsequently valued much more highly in the organisation. Apart from the one team member who left the rest of the team stayed stable for 12 months.

It takes skill and determination to improve a team but there are good models to follow. This case study illustrates how our four step sequence helped one team and the same principles can be applied when a truly transformative approach is needed.

Conclusion

Retrospectives are weak because people and organizations have not made a genuine commitment to greatness. In this article, we describe how to work with an organization to determine whether there is a desire for transformation and, if there is, how to then create retrospectives that support greatness.

About the Authors

Michael de la Maza is a Scrum Alliance Certified Enterprise Coach (CEC). As an Agile consultant, his major engagements have been with Paypal, State Street, edX, Carbonite, Unum, and Symantec. Previously, he was VP of Corporate Strategy at Softricity (acquired by Microsoft in 2006) and co-founder of Inquira (acquired by Oracle in 2011). He is the co-author of Professional Scrum with TFS 2010, Why Agile Works: The Values Behind The Results, and Agile Coaching: Wisdom from Practitioners. He holds a PhD in Computer Science from MIT. He is on email at michael.delamaza@gmail.com and on Twitter @hearthealthyscr.

Bob Griffiths has been working as a team and individual coach for over 30 years.  He has worked extensively in manufacturing, IT and financial services and is the author of ‘Grow your own Carrot’ - the definitive guide to the GROW coaching model as well as many articles and videos on virtual coaching.  He runs the CoachMaster Network which provides coach training and has developed CoachMaster® software to build coaching capability in organisations.  He firmly believes the best learning happens when people are focused, challenged and having fun.  He is on email at robert.griffiths@thecoachmasternetwork.com  

Rate this Article

Adoption
Style

BT