≡ Menu

Lean Kanban Central Europe – Day Two – AM

I’ll be live-blogging from Lean Kanban Central Europe in Vienna. This is my second visit to LKCE. Last year I was speaking, health system but this year I’ve decided to just come to learn (and also to scout for speakers for the ACE! Conference). I learned that my retention is higher when I take notes for others (thanks to a blind chemistry student in college who asked me to take notes for her) so I’m going to attempt to increase my focus and retention levels by live-blogging during the event.

I will be updating this post throughout the day with the core takeaways that I get from each talk.

Don Reinersten – Keynote

image

It’s easy to see the cost of underutilization and less easy to measure the cost of delay. Neither can be optimized independently without harming the other. Optimizing the system economically involves minimizing the combined costs of delay and underutilization. WIP limits are one tool to do that.

Random variation does not self-correct in the short term. One example is a coin toss. While heads and tails may come up the same number of times on average, rehabilitation the sum of either will trend away from zero. In another example, health ten large tasks are unlikely to be followed by ten small tasks. Therefore, queues will tend to grow large over a longer period of time than you might be comfortable with, although they’ll average out over eternity.

Don suggests starting with a WIP limit that is twice the unrestrained WIP, and then adjust it down by 20-30% until you find the optimal point at a given time, remembering that WIP adjustments are a powerful tool because they are easy to adjust in BOTH directions.

There are more ways to control WIP then blocking. Don suggests looking to the telecom industry for ideas rather than manufacturing. For example, you can purge WIP by either throwing out work in progress in favor of higher value work or releasing unfinished work if it’s far enough along to deliver value. You can shift resources to adapt to changing demands. You can use rules on buffers or class-of-service WIP limits to ensure that high value work gets through the system in the most economically value-maximizing way.

 


Dominica DeGrandis

Emerging Patters for Kanban in IT Operations

image

Common problems in IT work:

  • Work is unpredictable.
  • Dependancies are difficult to manage.
  • Distributed teams are a fact of life for many teams.
  • Specialization – individuals become overtaxed bottlenecks.
  • Interrupted work is a major cause of job dissatisfaction.

Perhaps it’s no coincidence that three of the top ten most-hated jobs are in IT.

Dealing with distributed teams

  • Improve communication infrastructure (good microphones, cameras, large screens).
  • Use live cams to let remote team members see each other working.
  • Google handout is a good way to do this cheap.
  • Maintain open chat channels and use a common kanban board (web-based).

Dealing with interrupts

  • Have a goalie (a rotating role for dealing with incoming work).
  • Have a swimlane for unplanned work.
  • Set rules for dealing with tasks of different sizes.
  • Mark tickets that are interrupted by making a dot on the card when you are interrupted, to make the problem visible. It’s likely that in a dev team, over 50% of all work items are interrupted.

Specialization

  • If you have individual swimlanes because individuals have critical specializations, then the focus moves away from the work and toward the people.
  • Have some floaters, specialists who can assist in different areas. You can designate them on a kanban board with an avatar, while other ares of specialization have their own swim lanes. Floaters can move between swimlanes.
  • Have a clearly-defined scale of skills for team members, so that everyone knows who can do what, and how well, and what the skill distribution is within the team.
Lack of trust can be due to lack of visibility. Do managers/customers really know what their team members are doing from day to day?
In short – Show the risk in your organization.

Dan Vacanti

Little’s Flaw

image

There’s more to Little’s Law than L = lambda W ,

The value lies in the fact that while two of the components might be easy to measure, Little’s Law allows you to derive the other.

Example: If I add two bottles of wine to my 50-bottle wine rack every week, then I know that the average age of my wine collection is 25 weeks.

Assumes:

  • stable system (the average WIP at the start and end of the calculation must be roughly equal and the average age of the WIP must neither increase nor decrease ove r the period of the calculation)
  • long running averages
  • consistent units

Little called Lead Time, Flow Time, defined as the average time that an item spends as work in progress.

LT=WIP/TH

Lead time (LT) equals work in progress (WIP) divided by throughput (TH).

Littles Law requires:

  • Conservation of Flow: all work that enters the system must flow through and exit the system.
  • The average arrival rate must equal the average departure rate.
Therefore, if your lead time is based on backlog to done, and the backlog grows erratically, you can’t use Little’s Law.
So you can only use Little’s Law to derive average data over time given all of these assumptions are true for your system. It’s predictive ability is limited to averages.
When teams are starting out, understanding the assumptions behind Little’s Law, they can adopt process policies that increase system predictability.
e.g. Try to make arrival and departure rates roughly equal. Try to avoid artificial aging of items in WIP.

From the horse’s mouth, read Little’s Law Chapter Five [PDF]


Torbjorn Gyllebring

Faith, Science and Rightshifting

image

observation: He’s using nonviolent communication patterns. Win!

The goal of the talk is to help us to create coherence between what we know (science) and what we believe (faith) for the purpose of improving effectiveness (Rightshifting).

Reciprocity is about mutual exchange. It can be Symetrical (equal exchange of value), Negative (you hurt me, so I hurt you), Alturistic (give, and the world will reward you).

When you mix faith and reciprocity, you get the golden rule. All major religions have a version of the golden rule, indicating that a sense of reciprocity is a core characteristic of being human.

But people are different, so their needs, values and expectations should be taking into account. Don’t treat them EXACTLY as YOU want to be treated.

Try evaluating your team’s practices according to the golden rule. e.g. Does your method call people farm animals? Do you want to be called a farm animal?

The science of the golden rule. There are evolutionary advantages to the golden rule. Species with a sense of reciprocity are more adapted to changing environments.

If we really believe in “rational man” who maximizes his own utility.

The ultimatum game

Two pirates divide loot. One pirate divides the loot, and the other accepts or rejects the division. If they reject it, then the loot goes to no one. A rational pirate would accept any division, because there’s no cost. But in reality, real people reject non-reciprocal offers. In studies, people tend to offer 50:50 to 60:40 divisions.

We are not purely and solely rational creatures. We are at our core, ethical creatures.

In multiple rounds of the prisoner’s dilema, the rational behavior changes and reciprocity takes over. People tend to act as they remembered the other party acted.

Theory of Mind – the ability to put ourselves in someone else’s shoes

Mirror neuron – give us the ability to sub-consciously put ourselves in the position of others.

Der Wiener Deewan – A cafe in Vienna, near a university, has a pay as you like pricing model. They tend to earn more this way, because of the human mechanisms discussed above.

The mechanism that blocks unethical behavior is that when you think of doing harm to someone, your brain forgets the distinction between yourself and that person and reacts negatively to the thought of that thing happening to you.

Kindness, generosity, gratitude, compassion – the manageable causes of happiness. Do these things describe your organization?

Improve your happiness by keeping a gratitude journal of thinks you are thankful for. It works.

Rightshifting is about making organizations more effectively by making them more humane.

Organizations can find themselves in one of four phases: Ad hoc, Analytic, Synergistic, and Chaordic. The distinguishing characteristics of the process is increasingly kind, adaptive, supportive and humane organizational characteristics creating greater effectiveness.

In a high-trust environment:

  • transaction costs go down
  • heightened demand, because of lower fear
  • no burocracy
  • reduced need for contracts
  • effective communication

These lead to higher productivity and profit. High trust organizations are three times as profitable.

The science is in line with our ethics, and so it’s time to change the world. Someone needs to take the first step and lead the way towards being more human.

Comments on this entry are closed.