Making Smaller Steps
Making Smaller Steps Part 2/2
Ever struggled with the next step being too big and taking lots of time before we can see the outcome? Not a good situation for progressing by experiment and finding a path in unknown territory. In this and the following episode, I would like to continue discussing this challenging situation for the coach.
Denise remembered a coaching cycle with Michael when he had first joined the team. At that time, Michael had worked on reducing the time for attaching a muffler.
Denise: "Michael, what Obstacles are preventing you from reaching the Target Condition?"
Michael: "Attaching the muffler simply takes too long."
Denise: "And what exactly is the problem?"
Michael: "The muffler often gets stuck when we try to attach it. I think it is caused by the way the team members put the muffler on the tube. If they place it at an angle, it jams and cannot be put on all the way. Then it has to be loosened and completely removed. And then they have to start all over again. That takes a lot of time."
Denise: "So, what's your next step?"
Michael: "I'll discuss with the team how the muffler can be attached without problems. Maybe we need a special fixture for that."
Denise: "And what do you expect?"
Michel:" That we can attach the muffler without any difficulties. That would be a lot faster."
☞ Pause for a moment and consider: What would you do as a coach now?
Denise: "What does that mean in numbers?"
Michael: "I think that we can easily win 30 to 40 seconds here".
Denise: "What is the ‘unwanted effect’ of jamming the muffler on your process metric?"
Michael: "I do not know exactly, but at least 30 to 40 seconds, I would say".
☞ Pause for a moment and consider: What would you do as a coach now?
Denise ignored the estimate and continued: "Don’t worry if you don’t know. What is, therefore, your next step to find out?"
Michael: "Well, I have to take a closer look at the process and measure the time we need when the muffler jams compared to the time when it attaches easily."
Denise: "And what do you expect?"
Michael: "That we can attach the muffler faster.“
☞ Pause for a moment and consider: What is the problem with Michael's answer?
Since the next step contains observing and measuring the current state, this will not improve the process, Denise thought. She flipped to the page in her notebook about the logical connections. The expectation should match the current threshold of knowledge about the Obstacle. In Michael’s case, that is not knowing the unwanted effect.
Reflecting on that coaching cycle, Denise realized that she had asked Michael: "What do you expect regarding the Obstacle?" Michael had answered: "That I know how much time we lose if the muffler jams and what exactly causes it." That was a good expectation, and Michael had written it down in his Experimenting Record.
Similarly, Denise often used the expectation question as a deepening question by adding, “What do you expect to know after taking this step?” This often made it easier for her Improvers to formulate a precise expectation for an analytic step or an observation.
Tip for the Coach: If the expectation does not match the threshold of knowledge, ask*: What do you expect to know after taking this step?
*(especially on experiment type 1 and 2)
When data was missing in the expectation, especially for experiments of type 3, Denise often used the universal question, "what does that mean in numbers?"
Sitting at her desk now, Denise went back to her list of deepening questions for phase 3.
Denise remembered using questions 3.3 to 3.5 when a more detailed analysis regarding the cause of an Obstacle was needed. Denise called that set of questions the “Cause Funnel”.
Question 3.7, in turn, was very helpful for problems that rarely occurred. A coaching cycle with Joe came to her mind.
At that time, Joe had been working on reducing machine errors at one of the automated assembly stations where insulation rings were attached. To achieve that the rings were automatically separated, erected to vertical position, and then picked up by a gripper.
Occasionally, the rings were hooked into the device and then crushed by the gripper. This was rare, but when it happened, it was usually several hours before all fragments of the crushed ring were removed.
Joe had suggested he observe several of these incidents to understand precisely what caused this to happen. That would have been very time-consuming and rather impractical. He would have had to spend some days next to the machine, staring at the gripper. On the other hand, whenever it happened and the team members called him, it was too late to analyze and find the root cause.
Denise remembered that the question: "How can we reproduce the problem?" had helped Joe re-think his approach to analyzing the problem.
As a result, they had split the issue into two parts, ‘erecting to vertical’ and ‘gripping’. First, they had analyzed the ‘erecting to vertical’ part.
Denise remembered that she used the question: “How should the ring be erected for the gripper to grab it correctly?”
Joe had found out about the right position and orientation to avoid the ring being smashed.
Afterward, Denise had continued by asking: “Which parameters influence the process of erecting the ring?"
They found several, such as, the distance between the feeder and the erecting lever, as well as the angle and speed of the erecting lever.
☞ Pause for a moment and consider: What coaching approach is Denise using here? Which deepening questions is she adapting to do so?
After finding out about the parameters, Joe was able to simulate the issue, or rather the effect of each parameter, in separate experiments. Joe gradually changed one parameter at a time until they knew at which setting the ring was no longer correctly erected.
Of course, while experimenting, the rings were not damaged because Joe always stopped the gripper before it crushed a ring. In addition, Joe had progressed through small and very quick steps. He was even able to conduct multiple experiments every day during the shift breaks.
Denise made an additional note:
Tips for the Coach:
If steps tend to be big, find a way to split the problem into smaller parts.
If obstacles occur only rarely but have a significant effect when they do, find a way to simulate the issue.
Thoughts on Making Smaller Steps
☞ If you have taken notes on how to prepare smaller steps while reading this and the previous article, you might want to view them now.
When coaching, I find it helpful to think of the following scenarios for why steps tend to be too big or why it seems we just have to wait.
Scenario 1:
We have not yet understood the obstacle well enough. When we don’t know the cause and effect of an obstacle, it will be difficult to come up with a decent hypothesis. Then, if pushed to do a next step, often by our action bias, we act based on assumption. Of course, we want to get it right, but without understanding the cause, this is nearly impossible. We therefore lean towards bigger steps. The more, the merrier, we think. Or at least bargaining for an extended due date, due to the significance of the measure, we can buy some time.
A situation like this can also be caused by a coach too quickly asking about the next step (Question 4) or a manager or supervisor asking for results. Peer pressure might also cause this chain reaction. In a group situation, we sometimes feel the urge to be the first to come up with a smart solution.
Tip for the Coach: Go Back If confronted with a pre mature step, "Go Back to Obstacle" (Phase 3) to understand cause and effect first. Useful deepening questions:
What Obstacle are you addressing with this step? (Repeat & Add with question 3 of the Coaching Kata)
What exactly is the problem?
What is the unwanted effect of [Obstacle]?
These questions bring us back to the Obstacle and provide a new start for analyzing cause and effect. Often the next step will be some observation or analysis, and much smaller than taking the arbitrary action proposed in the first place.
Scenario 2:
Occasionally, when planing the next step, we think more of the overall outcome we want to achieve than the actual next step that we need to take first.
Example: Obstacle addressed: Long waiting times in the canteen at certain hours Coach: „What is therefore your next step?“ Improver: „Automatically inform people about the length of the queue in the canteen via the home page of our intranet.“
That might be the end we have in mind to solve the problem. However, that is a giant step that will take longer to implement. It is not even a single step, but a series of steps are necessary to achieve this.
How will the length of the queue be measured? Let’s say we use a sensor. What kind of sensor is most suitable for this task? And where can we get it? How much does it cost? Does that fit with our budget? Where and how do we have to install the sensor? How will we evaluate the data? How will the data get into the intranet? In which format should we display it?
There will also be some other obstacles in communication and involving people. How do we inform and engage the canteen team? How do we get the IT team to change the intranet home page?
These are all obstacles, and there will be more we are currently not even thinking of. Each Obstacle needs to be addressed to achieve what we have in mind as an end point.
That will need many steps, and these will be a lot smaller than the „big thing“ we mentioned first.
Tip for the Coach: Slice the Elephant If confronted with a big step, "Slice the Elephant" into multiple steps.
Useful deepening questions:
How will you proceed to do this?
What do you therefore do first?
How can we simulate that?
Scenario 3:
Every so often, however, steps will take time to complete. We might have to wait for a supplier to provide some tools, engineering preparing a design change or a new software mask programmed. However, if I have to painstakingly wait to be able to make progress, I can at least do two things. Make sure that what I am waiting for is making progress, so we will meet the planned due date and prepare everything necessary to get going once I receive what I am waiting for.
Waiting for something and not checking progress might come with a big surprise once the due date arrives. Things are not ready yet! By then the time is gone, we can’t do anything about it and just have to wait again. If we had checked in regularly, we could have supported, helped problem-solve when the unexpected obstacles occurred at our supplier, or added resources. Now it’s too late. That makes for excellent small check-up steps taking place at a regular pace.
And then it arrives, the new software mask we have been waiting for. Ready to be tested. So let’s get started right away. But wait, we first have to define which test to run. And by the way, what exactly do we need to observe, measure, and check? How many test users do we need? Who will they be? How do we evaluate the results? Too late, we realize we are not ready to run the test and should have prepared all this while we waited for the software to be programmed. Again, this makes for nice steps, getting everything ready to go until the thing we have waited for arrives.
Another situation that often causes waiting time is when gathering data, e.g., measuring the actual condition of the Process Metric over several days.
An option is to still meet for daily coaching cycles. Although the measurement is not completed, we can take a look at the date we have gathered so far. There might already be something interesting to learn, or we can already identify an obstacle and address it while continuing to gather further data.
Tip for the Coach: Don’t Sit and Wait, Prepare If having to wait for a step to be conducted, "Don’t Sit and Wait, Prepare" to be ready to go. Useful deepening questions:
How do we know [the thing we are waiting for] is making the necessary progress?
What should the progress be right now?
What is the current progress?
What Obstacles are preventing us from using/testing [the thing we are waiting for] once we have it?
What will be our next step once we get [the thing we are waiting for]?
What do we need to prepare until then?
Scenario 4:
While waiting for something to address a particular obstacle, we might work on another Obstacle rather than staying idle. Of course, these obstacles should not be interconnected or influencing each other. Working on two or three obstacles in different phases might be a practical approach.
While we wait for a fixture to be built, we could be working on developing and simulating a solution for removing the next Obstacle.
If we have to wait for something, we might observe the process to spot the next Obstacle and start analyzing the effect and cause of it.
Having a separate Experimenting Record on the storyboard for each obstacle is helpful.
Figure: This is what Denis’ card looks like now.[1]
[1] Source: Toyota Kata Memory Jogger, GOAL/QPC, 2018, p.105 Adapted from: Mike Rother, The Toyota Kata Practice Guide, McGrawHill, 2018, p.285