10. Experiments
It is a truth universally acknowledged that a zombie
in possession of brains must be in want of more brains.
—Seth Grahame-Smith, Pride and Prejudice and Zombies
In This Chapter
• Explore ten experiments to improve continuously.
• Learn what impact the experiments have on surviving Zombie Scrum.
• Discover how to perform each experiment and what to look for.
This chapter presents experiments to help teams advance their ability to improve. Some experiments offer inspiration on how to run a Sprint Retrospective differently. Others take continuous improvement to the organizational level.
Experiments for Encouraging Deep Learning
Double-loop learning is a form of deep learning where existing rules, procedures, roles, and structures are challenged (see Chapter 9). Because this doesn’t come naturally to most people, we’re sharing our favorite experiments to get started.
Share an Impediment Newsletter throughout the Organization
The impediments that make it hard for Scrum Teams to work empirically often involve people across the organization. Helping these people understand the impediments and the problems they cause creates awareness that enables double-loop learning, which can lead to systemic improvements.
Effort/Impact Ratio
Steps
To try this experiment, do the following:
1. With your Scrum Team, ask everyone to silently write down impediments they see that are making it hard for them to build what stakeholders need or ship fast(er), or both. What skills are missing? Where is protocol getting in the way? Which people do they need, but don’t have access to? After a few minutes, invite people to pair up to share and build on their individual ideas. Together, share all impediments and pick the three to five impediments that are most impactful (e.g., with dot-voting).
2. For the biggest impediments, ask “What is lost because of this? What would we and our stakeholders gain when this impediment is removed?” Capture the consequences for the various impediments.
3. For the biggest impediments, ask “Where do we need help? What would help look like?” Collect the requests for help for the various impediments.
4. Compile the biggest impediments, including their consequences and requests for help, in a format that you can easily distribute to everyone who has a stake in your work. It could be a mailing, a paper newsletter, a blog post on your intranet, or a poster that you put in a heavy-traffic corridor. Include the purpose of your team and how to contact you. You can also include the accomplishments of your team, of course.
Our Findings
• Make sure to include (higher) management and consider informing them up front. Also, they will probably appreciate a shorter, more concise version of the newsletter.
• Transparency can be painful. Be honest but tactful in your messaging, and don’t blame others or be negative. State what is happening and make clear requests for help.
• If you are planning to do this experiment frequently, make sure to include the accomplishments of your team as well. What is going well? What has changed since the previous newsletter? And most important: from whom did you receive (unexpected) help?
Ask Powerful Questions during Sprint Retrospectives
As we explored in the previous chapter, people’s deeply held beliefs, assumptions, and values influence how successful they will be in changing. For example, when Development Teams feel that talking to customers is the responsibility of Product Owners, they limit their opportunities for collaboration. When people assume that their feedback is only useful when the entire Product Backlog has been implemented, they will struggle to embrace empiricism. Many of these assumptions are subconscious and need to be surfaced in order to be challenged. This experiment is designed to help teams by asking powerful questions that reveal hidden assumptions.
Effort/Impact Ratio
Steps
To try this experiment, listen for statements about whether or not something is possible. The Sprint Retrospective is a good opportunity, but so are other moments when teams are together. Ask “What do you believe to be true that makes you say that?” Work together to reframe the answer into a statement that starts with “I believe that . . . .” See Table 10.1 for examples.
When you have identified a belief, use the following powerful questions to gently challenge it. We took inspiration from work done by people in the Liberating Structures community (primarily Fisher Qua and Anja Ebers) on “Myth Turning”:1
• What would need to happen for you to let go of this belief?
• Who else believes this to be true?
• Where does this belief benefit you?
• Where do you see this belief confirmed?
• What are the signs that others are starting to question this belief?
• What is irrevocably lost when we don’t do this?
• What happens when this belief turns out to be wrong?
1 Lipmanowicz, H., and K. McCandless. 2014. The Surprising Power of Liberating Structures: Simple Rules to Unleash a Culture of Innovation. Liberating Structures Press. ASN: 978-0615975306.
Asking these questions won’t convince people to change their beliefs, but it may help them learn and reflect on why they hold those beliefs. In doing so, they may discover that changing a belief benefits them, but that decision is up to them.
Our Findings
• People may be overwhelmed and frustrated if they are not used to deep questions such as these. Ask permission from your team to help them reflect and learn by asking a deep question every now and then.
• Don’t tell people what their belief should be. Don’t share your own, unless people specifically ask you to. Invite people to challenge yours too. Make the identification of underlying beliefs a team effort or something for people to reflect on themselves.
Dig Deeper into Problems and Potential Solutions, Together
Effectively analyzing and removing impediments is important for deep learning and continuous improvement. Teams have to learn how to ask or write questions that include different perspectives, and to identify specific and actionable solutions. The Liberating Structure “Discovery and Action Dialogue”2 is ideal for this exploration. It contains a sequence of questions that teams can ask to understand the problem, come up with solutions, and specify steps that need to be taken.
2 Lipmanowicz and McCandless, The Surprising Power of Liberating Structures.
Effort/Impact Ratio
Steps
In Discovery and Action Dialogue, groups answer the following sequence of questions together:
1. How do you know when the problem is present?
2. How do you contribute effectively to solving the problem?
3. What prevents you from doing this or taking these actions all the time?
4. Do you know anybody who is able to frequently solve this problem and overcome barriers? What behaviors or practices made their success possible?
5. Do you have any ideas?
6. What needs to be done to make it happen? Any volunteers?
7. Who else needs to be involved?
Follow these steps for Discovery and Action Dialogue:
1. As input for Discovery and Action Dialogue, help your team or teams identify their biggest impediments. Many of the other experiments in this book are helpful here. Either choose the most important topic with one team or have participants from multiple teams form groups around different topics.
2. Give the group(s) enough time (at least 30 minutes) to answer the sequence of questions. If it makes sense, groups can deviate from the order or revisit earlier questions when new insights emerge.
3. When you run a Discovery and Action Dialogue with multiple teams, add opportunities for teams to share their findings with the whole group and gather feedback. A Liberating Structure such as Shift & Share3 is ideal for this.
3 Lipmanowicz and McCandless, The Surprising Power of Liberating Structures.
Our Findings
• Encourage the team(s) to spend enough time on the first question. Asking additional questions such as “What’s so challenging about this problem?”, “Is there a deeper problem we’re not seeing?” or “What happens when we don’t solve this problem?” will help dig deeper (see Figure 10.1).
Figure 10.1 Dig deeper into problems and potential solutions with Discovery and Action Dialogue.
• Keep in mind the concept of 15% Solutions—described next in this chapter—when asking what needs to be done to make the solution happen.
• Use a host when the team is struggling to maintain a good pace and flow for answering the questions. The host asks the questions in order and gives everyone the opportunity to speak to each question, while the host keeps track of time.
Experiments for Making Improvements Tangible
It’s easy for teams to remain stuck in vague and promising improvements such as “Communicate more” and “Involve stakeholders.” But when improvements are this imprecise, it is hard to know where to start and hard to validate when you’re done. The experiments in this category all revolve around making your improvements as tangible and small as possible.
Create 15% Solutions
Continuous improvement works best when changes are small and start from what people can change on their own. To help with this focus, the organizational theorist Gareth Morgan proposed the concept of “15% Solutions.”4 Working from the assumption that people don’t have control over 85% of their work situation, the focus shifts to the 15% they do have control over. Not only is this more motivating, it also keeps improvements small and free of the barriers that make the 85% so difficult to control, such as organizational culture, existing hierarchies, and rigid procedures. If everyone starts with where they have autonomy and opportunity to change, all those 15% changes together easily snowball into significant change across the organization.
4 Morgan, M. 2006. Images of Organization. Sage Publications. ISBN: 1412939798.
This experiment helps your Scrum Team define 15% Solutions and create change even in environments where little seems possible. It is based on the Liberating Structure “15% Solutions.”5
5 Lipmanowicz and McCandless, The Surprising Power of Liberating Structures.
Effort/Impact Ratio
Steps
To try this experiment, do the following:
1. Use 15% Solutions at the end of every meeting. This helps people turn what they’ve learned into actionable steps. Preferably, use a shared impediment or challenge to give focus to the 15% Solutions.
2. Ask everyone to generate a list of 15% Solutions for themselves. Ask “What is your 15 percent? Where do you have discretion and freedom to act? What can you do without more resources or authority?”
3. Invite people to share their ideas in pairs for five minutes. Encourage them to help each other make their 15% Solutions as tangible as possible. Questions that help are: “What is the first step to do this?” or “Where would you start?”
4. For maximum transparency, collect the 15% Solutions in the team room, for example around a Scrum Board if your team uses one.
Our Findings
• Don’t restrict the use of 15% Solutions to Sprint Retrospectives. Use them to identify where to start refactoring large and complex code bases, to identify next steps after a Sprint Review, or to conduct multiteam retrospectives.
• Help people resist the temptation to define actions for others or for the group as a whole, effectively moving away from what they control themselves. 15% Solutions work when people focus on their own contribution. It’s okay if solutions overlap or when they are not clearly related.
Focus on What to Stop Doing
Continuous improvement can easily devolve into adding more things to an already overcrowded to-do list: one more check to the Definition of Done, one more workshop to an overcrowded agenda, or one other technology to research. But as you add more, you are less likely actually to get any of it done.
Instead of adding more things to do, find things that you are doing that are unproductive and eliminate them. The Liberating Structure “TRIZ”6 is a huge help, as it invites creative destruction of activities that limit innovation and productivity by engaging everyone in a playful way. The name TRIZ is an abbreviation of the Russian version for “theory of the resolution of invention-related tasks.”
6 Lipmanowicz and McCandless, The Surprising Power of Liberating Structures.
Effort/Impact Ratio
Steps
To try this experiment, do the following:
1. Create three rows on a flip chart, leaving space open in each of them. Don’t add labels to the rows as this might spoil the twist of the second round.
2. Give everyone ten minutes to make a first list of everything they can do to guarantee the worst possible outcome. Ask: “How can you personally contribute to making our team so zombified in its ability to ship fast and collaborate with stakeholders that it becomes the prime example of ‘Zombie Scrum’ on Wikipedia?” Do this step first individually and in silence for a few minutes, then in pairs for a few minutes more. Encourage people to be creative and practical while keeping things within the boundaries of the law. Then invite people to share and build on their ideas for a few more minutes in pairs. Take five minutes to collect the most salient examples on stickies and place them into the top row.
3. Give the participants ten minutes to make a second list of activities the team is already doing that resemble or are closely related to items on the first list. Ask: “If you’re brutally honest, which items represent things we’re already doing or are moving in that direction?” First, give people a few minutes of individual reflection before they pair up to share their thoughts and notice patterns. Capture the most salient patterns of Zombie Scrum by moving the items from the top row to the middle one.
4. Give the team ten minutes to make a third list of all the activities or behaviors from the second list that they want to stop from now on. Start this step individually, then do it in pairs, and then repeat with the whole group. In the third row, capture the items that the team is going to stop doing from now on. Resist the temptation to add actions to stop something.
Our Findings
• Invite participants to have serious fun, go a bit over the top, and have a laugh while they’re doing it. This helps create a safe environment where people feel comfortable being honest.
• For deeper reflection, replace the activities and behaviors in TRIZ with beliefs and norms. Which beliefs should we have about each other, our work, and our stakeholders to guarantee the worst possible outcome? Which beliefs are already present or similar? Which should we let go of?
Create Improvement Recipes
Vague improvement ideas such as “More collaboration” or “Use Sprint Goals,” or ideas that have no clear start and end, won’t propel teams forward. This experiment is about translating vague improvements into something specific by building on each other’s intelligence and creativity. Just as a cookbook gives you detailed instructions on how to cook a dish with local ingredients, improvement recipes clarify the ingredients, the steps, and the expected outcomes. This experiment is based on the Liberating Structure “Shift & Share.”7
7 Lipmanowicz and McCandless, The Surprising Power of Liberating Structures.
Effort/Impact Ratio
Steps
To try this experiment, do the following:
1. During a Sprint Retrospective or a multiteam Retrospective, identify a handful of areas for improvement. Ask everyone to self-organize into small groups (three to five people) by picking the improvement they care about the most. Provide each group with an empty whiteboard or flip that designates their “station.”
2. First individually and in silence, invite everyone to think about what the recipe might look like by asking: “If we want to achieve this, what would help us do that? What practices come to mind? What have you tried elsewhere that might work here?” (two minutes). Then invite people to share their ideas in small groups and pick one (five minutes).
3. Explain the Definition of Done for a recipe. Each recipe needs to clarify: What is it trying to achieve (“purpose”)? Who needs to be involved (“people”)? What steps need to happen and in what order (“steps”)? And how do you know that the recipe is working (“success”)? If you want, you can prepare a canvas for each of the recipes.
4. Give the groups ten minutes to create a first increment of their recipe. Encourage groups to use their full creativity by writing, drawing, and using symbols.
5. Invite each group to pick one station owner. That person remains with their station for the remaining rounds. The others move clockwise to the next station. The station owner updates the new group on the progress and works together with them to build on the increment, adding improvements and clarifications as needed (five minutes).
6. Repeat as many times as necessary for each group to visit each station.
7. Ask the groups to return to their original station and take a look at the final version of their recipe that was created incrementally by the various groups.
8. Ask people to write their name on a sticky note and put it on one recipe they’re willing to commit to making possible. Give people a few minutes to synchronize how and where to start for the recipe they’ve picked.
Our Findings
• Improvement recipes often capture recurring patterns or local strategies to resolve impediments. Sharing useful recipes with other teams—both inside and outside the organization—is a great way to learn.
• If you notice that the recipes are superficial and vague, encourage groups to keep asking “How will we do this?” when they visit a new station.
• For initiatives that span more Sprints, encourage groups to frequently synchronize their work and progress until the purpose has been fulfilled.
Experiments for Gathering New Information
Sometimes we tell teams that “it’s hard to squeeze juice from dry oranges.” It is our—admittedly blunt—way of telling them that their toolbox or the flow of new ideas has dried up to the point that their continuous improvement has stalled. In this category, we share experiments that help bring in new ideas or people to see possibilities where you couldn’t see them before.
Use Formal and Informal Networks to Drive Change
Changing the environment of Scrum Teams can be difficult when you’re trying to do it alone, especially in large organizations where the people with influence are difficult to get in touch with. When working to remove an impediment, you should first find people in your organization who are facing similar impediments and work together to remove them. This experiment is about leveraging the informal and formal networks in organizations to create change. It is based on the Liberating Structures “Social Network Webbing” (for an example, see Figure 10.2) and “1-2-4-All.”8
8 Lipmanowicz and McCandless, The Surprising Power of Liberating Structures.
Figure 10.2 An example of a social network web9
9 Source: Lipmanowicz and McCandless, The Surprising Power of Liberating Structures.
Effort/Impact Ratio
Steps
To try this experiment, do the following:
1. Start by inviting Scrum Masters, Product Owners, and members from Development Teams who are eager to start removing organizational impediments to Scrum.
2. First individually (one minute), then in pairs (two minutes) and in groups of four (four minutes), ask people to answer “What are the biggest impediments we face? What in this organization is making it difficult for us to work empirically?” Collect the biggest impediments.
3. Create a wall or use the floor for the social map. Bring stickies with different colors.
4. Start creating the social network map by asking the participants to write down their name on stickies. Put them in the center of the social network. These people are the “core group.”
5. First individually (one minute), then in pairs (two minutes) and then with another pair (four minutes), ask people to identify the key groups or departments you need support from to remove impediments. Limit to a maximum of ten groups and create a legend where each has a different color or symbol (ten minutes).
6. Invite everyone to write down the names of people they know in the organization on separate stickies, making use of the legend. Ask people to place the stickies on the map based on how close or distant they are to the people currently present (ten minutes).
7. First individually (one minute), then in pairs (two minutes), and then with another pair (four minutes), ask people to answer “Whom would you like to include to remove the impediments we’re facing? Who has influence, a fresh perspective, or the skills we need?” Write down the names on separate stickies, making use of the legend. Map the stickies on the network based on their current and desired involvement. Update the legend as new groups are identified (fifteen minutes).
8. Ask everyone to take a look at the map that is emerging and ask: “Who knows whom? Who has influence and expertise? Who can block or boost progress?” Draw lines between people and groups to connect them based on the answers (fifteen minutes).
9. Use the experiment “Create 15% Solutions” elsewhere in this chapter to generate strategies for involving influential people who are distant or to work around blockages. How can you leverage your network to involve the right people? This can be as simple as making a phone call, sending an email, or asking someone closer to you to make the connection. You can use the experiment “Share an Impediment Newsletter throughout the Organization” to inform people in the network.
Our Findings
• Pay close attention to black holes in the map. These are the departments or groups you need support from, but where you don’t know anyone (directly or indirectly).
• This experiment works best when you repeat it every so often. Try to expand your “core group” with people who are willing to help. As your network grows, it becomes progressively easier to remove blockages or boost progress.
Create a Low-Tech Metrics Dashboard to Track Outcomes
How well is your team performing? Do you know what outcomes you’re delivering? More often than not, teams try to answer these questions by tracking velocity or number of items completed per Sprint. Although these metrics tell you how busy you are, they don’t tell you how useful that work actually is. Even worse, organizations often tell teams what to measure and then compare them with other teams. In this experiment, we outline the steps for helping teams to select their own metrics.
Effort/Impact Ratio
Steps
To try this experiment, do the following:
1. Before starting this experiment, clarify the difference between output- and outcome-oriented metrics. Refer back to Chapter 9 for examples.
2. First individually (one minute), then in pairs (two minutes), then in groups of four (four minutes), ask people to consider how they would know that their team is doing better. Ask: “How do we know that we’re responsive to our stakeholders? What metrics would go up when we do a good job and down if we don’t?” Together, collect relevant metrics with the team (five minutes).
3. Repeat for quality. Ask: “How do we know that our work is of high quality? What metrics would go up when we do a good job and down if we don’t?”
4. Repeat for value. Ask: “How do we know that we’re delivering value through our work? What metrics would go up when we do a good job and down if we don’t?”
5. Repeat for improvement. Ask: “How do we know that we’re finding time to improve and learn? What metrics would go up when we do a good job and down if we don’t?”
6. Together, look at the selected metrics and remove obvious duplicates. First individually (one minute), then in small groups (four minutes), ask people to remove metrics that the team can do without, while still being able to measure their progress on responsiveness, quality, value, and improvement. Together, keep the most minimal set that covers these areas (five minutes).
7. For each of the metrics left, explore how to quantify them well and where to get the data from. If additional research or setup is needed, you can add this work to the Product or Sprint Backlog.
8. Set up a dashboard—preferably just a whiteboard or flip—that the team updates (at least) once every Sprint. Create graphs for the various metrics to track trends. Resist the temptation to set up overwhelming dashboards in digital tools. First build the discipline to track a handful of metrics and inspect them every so often. Low-tech dashboards, such as whiteboards, promote experimentation because they’re easier to change in terms of presentation, content, and format.
9. Inspect the dashboard together during Sprint Reviews or Sprint Retrospectives. What trends are obvious? When you run an experiment, what would you expect to see change? A Liberating Structure such as “What, So What, Now What?”10 is well-suited for this.
10 Lipmanowicz and McCandless, The Surprising Power of Liberating Structures.
Our Findings
• When it comes to metrics, it is easy to try to measure too much. Be purposefully minimalistic by starting with the essentials: for example, stakeholder happiness and cycle time. Add more metrics when it helps your learning and when teams develop a rhythm in maintaining and inspecting them.
• Don’t turn metrics into key performance indicators (or KPIs) and work hard to prevent others from doing so. When metrics are used to appraise the performance of teams, it incentivizes them to “game” the numbers. Instead, use metrics purely for learning what works and what doesn’t.
• Don’t hide your dashboard from stakeholders. Instead, engage them in making sense of the data and finding opportunities for improvement. They benefit from the data just as much as your team does.
Experiments to Create a Learning Environment
Continuous improvement requires trying new things, some of which will lead to improvements and some will not. People who have to worry about the repercussions of making mistakes or being criticized will avoid trying new things, which means they can’t learn and improve. In this part of the chapter, we share experiments that make it easier to promote a culture of learning.
Share Success Stories and Build on What Made Them Possible
Instead of focusing on the things that don’t go well—as is all too easy in Zombie Scrum—you can help teams focus on what is already working and improve from there. Sharing successful experiences, stories, and strategies from the past is a good way to both build safety and uncover unseen paths forward. This experiment is based on the Liberating Structure “Appreciative Interviews.”11
11 Lipmanowicz and McCandless, The Surprising Power of Liberating Structures.
Effort/Impact Ratio
Steps
To try this experiment, do the following:
1. This experiment can be done anytime. The Sprint Retrospective is a natural opportunity, but so is the start of a Sprint Planning or Sprint Review. You can do it with a single team or with multiple teams to spread stories and learnings.
2. Ask everyone to form pairs and sit face to face. Make sure that everyone has something to write on and something to write with.
3. In turn, invite people to interview each other for five minutes per person. Ask: “Share a story of a time when we worked together to overcome a small or large challenge and you are proud of what we accomplished. What made the success possible?” Interviewers primarily listen, asking a clarifying question now and then. Make sure they also take notes, as they’ll need those in the next step.
4. Invite the pairs to find another pair. In ten minutes, everyone retells the story of their partner (about two minutes each). When one person is retelling the story, the others listen closely for patterns in what made the stories possible. Afterwards, gather key insights from the whole group and collect them on a flip (ten minutes).
5. First individually and in silence, ask people to think about what they can do to have more of these stories in the future (two minutes). Ask: “How can we build on the root causes of success? How can we be successful more often?” Then invite people to share their ideas in small groups (four minutes). Collect the most salient ideas with the whole group (ten minutes).
6. Use the experiment “Create 15% Solutions” or “Create Improvement Recipes” elsewhere in this chapter to translate potential improvements into specific actions.
Our Findings
When you do this experiment, watch out for the following:
• When you have an uneven number of participants, you’ll end up with one group of three. Let this group creatively work within the same time box as the other groups.
• Pay close attention to group dynamics and posture when people share their stories or retell those of others. Not only is it nice for groups to reminisce about their successes, it is also a positive experience to hear someone else retell your story in their own words.
Bake a Release Pie
A great way to build team spirit is to acknowledge small successes as they happen. For example, teams can celebrate every time they release to production, or when they automate another part that would require manual work otherwise. We’ve found it helpful to celebrate these successes in a simple and playful way while also giving everyone on the team the opportunity to contribute.
Effort/Impact Ratio
Steps
To try this experiment, do the following:
1. With your team, identify a specific achievement that is worthy of celebration when it happens during a Sprint. Pick an action that helps you work more empirically and that is challenging or frequently postponed. For example, a release to production. Or verifying an assumption with a real user. Or pairing up with others instead of adding more work to “In Progress.”
2. Find a big sheet of paper or a whiteboard and draw a big circle on it. Divide the circle into six or eight slices so that it represents a “release pie” (see Figure 10.3). Put the diagram in a visible spot in the team room.
Figure 10.3 Celebrate releases by creating a release pie!
3. Every time your team completes the action it identified, mark one of the slices. You can add the initials of the person who completed it, but only when everyone on your team has the opportunity to actually complete or contribute to the action.
4. When all the slices of the pie are marked, go out and get a real pie. Or something else that the team enjoys and gets them together.
Our Findings
• You want to set goals that are challenging but possible to achieve multiple times a Sprint. Adjust the number of slices and the difficulty to the capabilities of a team.
• Pick actions where achieving them is visible to others in the team. Otherwise, the decision to mark a slice becomes too subjective and based on individual motivations.
Now What?
In this chapter, we explored a set of experiments that are designed to help your team, and the entire organization, improve continuously. In part this involves double-loop learning. But it also requires a safe environment, new inspiration from external sources, and tangible improvements. Use these experiments, or draw inspiration from them, to start improving continuously now.
Looking for more experiments, recruit? There is an extensive arsenal available at zombiescrum.org. You can also help expand our arsenal by suggesting what worked well for you.”
