‑ Continuously Improve
Continuous improvement, also known as Kaizen, is one of the tenets of Agile and Lean.
Continuously trying to improve the process forces you to actually think about process. People and teams who do not have process thinking will not keep their effectiveness and efficiency status quo. Their process will degrade. A not-so-great process improvement method might keep you on the status quo line and a good one will actually make you continuously improve.
The continuous improvement efforts seek both incremental improvement over time but also changing the mindset to a better one, and thus help quicker closing the gap between the current state and the preferred state.
To continuously evaluate your current process and do small incremental improvements on it is often sufficient to get better effectiveness (make sure we are doing the right thing) and efficiency (make sure we are doing that thing right, both cases with as little waste as possible).
But you might be reaching the top of a smaller hill than you wanted. Then you need to go down that hill and up a steeper slope, and this is all about changing the mindset. More on this below.
A good method for incremental improvement is to have frequent retrospectives within the team. A 45-60 minutes meeting every week where you come up with three or four actions to do during the week is the recommended setup. Less frequent retrospectives will quickly make you lose focus on improvement.
There are three essential steps for a retrospective in its simplest form. First, you need to list and understand the problems the team have. Then, you need to suggest ways of improving these, often in the format of action points. And lastly, in the following retrospective, you need to follow up the action points.
Esther Derby and Diana Larsen wrote a great book about retrospectives where they proposed the following structure for a retrospective meeting:
- Set the stage
- Gather data
- Generate insights
- Decide what to do
- Close the retro
There are a lot of ways of doing these five steps, you can for instance consult the Retromat, a retrospective games suggester. But, here’s how I would do it:
1. Set the stage
This part of the retrospective should set the tone and the direction. Create a safe environment where the team participants can share their inner thoughts and feelings. Some teams start by reading the Prime Directive out loud.
Retrospective Prime Directive
Regardless of what we discover, we understand and truly believe that everyone did the best job they could, given what they knew at the time, their skills and abilities, the resources available, and the situation at hand.
Here are some other alternatives:
Take turns in giving the person next to you a back slap for something good they’ve done since the last retro. This sets a happy mood for the retro.
Who said it
Look through email threads, chat logs, story card comments etc. before the retro for fun, funny or great quotes. Read out the quotes and ask the team to guess who said it and in what situation. This shows how connected the team members are and sets a happy mood for the retro.
Know thy neighbour
Take turns in describing how the person next to you felt during the iteration. The person in question will confirm or correct the guess. This shows how connected the team members are.
You and me
Write one post-it to a person thanking them for something specific since the last retro and read it out loud before giving it to her/him. Write another post-it with something personal for yourself to improve upon.
Come up with a superhero name for yourself that should incorporate a trait of yours, such as commit-and-run-boy or delayed-reaction-man. Vote for the best one. This sets a happy mood for the retro.
If the time since the last retrospective could be summarized with a film title, which would it be? This simple exercise can be awkward but always brings out a good laugh. It’s a good opener that make people think about the things that have happened.
2. Gather data
In this part of the retrospective you try to elicit as much information from the team participants as possible. This includes following up the action points from the previous session - did you manage to do what needed to be done, was it good, shall we do it again?
The standard format for gathering data is to use the closed brainstorming method. Everyone starts by listing things (based on the different areas of the chosen retrospective game, like the ones below) on post-its (one thing per post-it). Then each take turn putting the post-its up on the board, briefly explaining them.
To inspire each other, this can also be done in an open brainstorming way - after a person has written one post-it, he/she reads it out loud and puts it up on the board.
This part is preferably timeboxed, for instance maximum 4 minutes, both to limit the number of post-its written and to make sure only important things show up.
The Sailboat exercise works well in many different situation, so this is my go-to-game when I don’t need to focus on a particular outcome of the retrospective.
The board is divided into four areas. One is things that makes us go faster and be happy (wind, sun). Another is things that makes us go slower (anchor). A third is risks that we know about that we need to steer clear of (land). The last area is the unknown risks, the gut feelings, a place to voice opinions and fears (mine).
The Spiderweb exercise works best for teams with a strong focus on doing things. I usually run this every fourth retro or similar for most teams.
In this retro game, the board is divided into five areas - things we should keep doing, things we should stop doing, things we should do less off, things we should do more of, things we should start doing.
The Team radar can be done every month or every second month, or for some team this is a quick exercise at the end of every retro.
The exercise requires the team to have agreed upon team values that they hold dear and want to focus on. For every value, the participants individually rates the teams achievement in that area. Then, a radar plot is drawn from the participants mean values.
The Timeline retro usually fits at the end of a larger iteration, release or project.
You list positive and negative things that have happened during this time on a timeline. The timeline makes it easier to remember approximately when a thing happened and also remember things because they happened during a specific period of time. You also list possible deltas, changes, that you come to think of right away.
The Glad-Sad-Mad game is good to run every third or fourth retro to get a feeling of the feelings in the team. The board is divided into three areas - things that make you glad, things that make you sad and things that make you mad.
The 4 L’s is a general purpose exercise. The board is divided into four areas - things you have liked since the last retro, things you have learned, things you have lacked and things you have longed for.
Remember the future
This game is something you can use now and then to make the team think a little bit deeper on their situation. It can also be used as an intro to Improvement kata. The team lists things that worked well in an utopic future iteration.
The Undercover boss game works well from time to time to make the team reflect. The team lists things that their boss would complain about if he/she had been undercover in the team since the last retro.
This is a little more lightweight retro game, to get a break from the tougher ones. The team nominates user stories for the Oscars in three different categories:
- Best story
- Most annoying story
- Most complex story
Traditional American Wedding
This retrospective game can be used from time to time to get a new viewpoint on the process. I do not feel it should be used too often, since the format is fun to look at but fairly boring when it comes to implementation. The board is divided into four areas - something old (things that we’ve been stuck with for ages), something new (things that showed up recently), something borrowed (things that other people do that we perhaps should copy), something blue (things that makes us sad).
This retrospective is the speedy version, named after an old team mate expressing his confusion or concerns. It can be used whenever, but is probably best used when you have to shorten the retrospective meeting significantly. Everyone lists their WHUUUT?!-moment of the week.
3. Generate insights
In this stage you structure the data to see trends, analyze and understand the problems.
The standard format is simply to remove duplicates and prioritize which post-its to discuss. If there is a lot of data, you can ask “5 why’s” for every post-it and group them by root cause, and then discuss the different root causes.
In the simplest version, set a short timebox and make sure that everyone dot-votes on the post-its they want to discuss.
During the discussion phase, try to timebox each topic. The Lean Coffee format is useful. You give 5 minutes per topic, and that’s it for that topic. Or, you give 3 minutes per topic and then vote if you should give it another 3 minutes.
4. Decide what to do
In this stage you list all the possible actions you could take. These might be solutions or just the first step towards a solution. You prioritize them and you bring just a few into the action point list. Too many action points makes it hard to focus. A good number for a smaller team is around 4. T
Each action point should have a person assigned to it who makes sure it gets done. This is not necessarily the person actually doing the action.
As an extra, for each action point, draw a scale from 0 to 100 percent and let each participant rate the probability of the point getting implemented. This could help when prioritizing.
5. Close the retro
The last part of the retro is where you wrap things up and thank everyone for the retro. You would want some feedback on the process, a retro of the retro, and these are some ways of getting that:
Nail it to the door
When the participant’s leaves the retro, ask them to put two stickies on the door, one with a positive comment and one with a possible improvement of the retro.
Draw a scale from 1 to 10 and let each participant rate (by putting a mark on the scale) how satisfied they are with today’s retro.
Create a poster together with the action points to improve visibility and follow-through.
Changing the mindset
Changing the mindset requires the person or the team to go through a sort of trauma. To go from one mindset, such as “I work best alone” to “we work best as a team” is not an easy step to make.
One of the most well-known change management models is the Kurt Lewin model with its three stages. Even if reality is not quite so simple, it serves as a good tool for understanding change.
- The thawing stage is probably one of the more important stages to understand in the world of change we live in today. This stage is about getting ready to change. It involves getting to a point of understanding that change is necessary, and getting ready to move away from our current comfort zone.
- The second stage occurs as we make the changes that are needed. Changing something is not instantaneous, it’s more of a process, a transition, bridging the gap between mindsets.
- The freezing stage is when you set out to make the transition/change permanent. A well-known method is the 30-day-challenge, when you do something everyday for 30 days. This makes you learn the change and it will stick better.
Hence, teams that are having a continuous and frequent improvement process tend to adapt to change faster.
The Marshall Model
Bob Marshall has a model concerning organizational maturity, that can and should be applied to teams as well. He naturally calls it The Marshall Model.
He lists four mindsets (or rather seven stages), ad-hoc, (novice/competent) analytical, (early/mature) synergistic and (early/proficient) chaordic.
- Ad-hoc is when the organisation is making everything up as they go along and repeatedly solving the same problems, often in different ways. My take on this is that this is most likely the most common mindset of a team, simply because we’re not mature enough in general … yet.
- Analytical is when there is an organisation structure that on paper looks great but every part of the structure work in silos and focusing on local goals. This is also a common state of a team, at least from time to time. The workflow is structured, but when it comes to the details, it’s still ad hoc. Every team to local optimizations of their work.
- Synergistic is when the organisation has a holistic view of the situation with common systemic goals. The decision-making is often evidenced based and everyone knows what’s really valuable for the company and the end users. The synergies you get is often higher motivation, faster learning and great adaptability. This is where Agile is aiming.
- Chaordic is when you leave the rules and guidelines behind. The organisation is always in process thinking mode and can intuitively grasp complex situations. This is what is required for true self-organization.
Between the four mindsets are tough transition zones requiring good change management to get through. This is why we see so many teams doing something that looks like Scrum in a Waterfall. What’s important to note in this model is that a team that does mature Waterfall may deliver more and better than an immature Agile team.
Moving from the left to the right in this diagram is something Bob Marshall calls Rightshifting.
The Collaboration Model
Another model for changing mindsets is the Collaboration model with its 5 steps that partially overlaps the Marshall model.
Since this model focus on collaboration and a tenet of Agile is exactly that, it can easily be applied to all Agile practices.
A kata is a pattern you practice to learn a skill, to form a “muscle memory”. The Improvement Kata helps you build a culture of kaizen by making continuous improvement a habit. It is part of the Toyota Kata. At Toyota, where this kata were invented, it is said that the employees spend about 40 percent of their working day to process improvement, and they do it quite effortlessly.
The Improvement Kata consists of three steps in a planning phase and one step in the execution phase.
Understand the direction and the challenge. Set the goal.
Grasp the current condition. What’s the current situation. Where are we?
Establish the next target condition. Set a subgoal, a step on the way towards the goal.
Work towards the target condition by doing experiments. This can be done using the retrospective format, but making sure that all action points aim for the target condition.
The image below is a classic in a LEGO version by Håkan Forss. If you feel that you are too busy to improve, you really need to stop what you doing and start to improve.