Problem: too much information presented too quickly. I bailed out of a session or two.
Day 3 of the conference, and I am starting to see some themes from the things I hear:
- Check your assumptions when you think it is "too hard."
- It's better to start with a generally positive attitude: people are good, the conflict can be resolved, the solution must be simple.
- Look at the evidence. (This is something I am hearing louder after hearing it several times at LSSC12.)
- We get frustrated when we see gaps or conflicts. Don't stop there - take your passion and do something.
And some entertaining tidbits:
- A change of top management is large enough "antibiotic" to kill the TOC bacterium.
- "Peanut butter management" - smearing 3 people across work that requires 10, ensuring delays across all those tasks and projects.
- Learning from failure is "easier" because success requires many interacting parts, while it only requires one thing to go wrong to fail.
The day started with a case study discussion from Ellwood City Forge. They have implemented the Reliability solution from the Viable Vision process. (Have a look at their website - it is right there on the home page - "Superior due date performance and fast deliveries.") The CEO, a head of Operations, and the head of Sales presented a great story. They have been doing and trying TOC since the early 1990's, and implemented the Viable Vision in 2007. They've gone from 50% due date performance to beyond 90%. Beyond that they highlighted a number of internal beneficial changes that were interesting to hear: the silo between sales and operations dissolved; clients weren't ready for the solution, so it took time to devise the right strategy; this is a business solution, not just an operations or sales solution.
The rest of the day was loaded with breakout sessions organized by theme or topic. It was difficult to choose across the sessions. They are being recorded, so maybe I'll be able to go back and take a look at what happened in some of those other sessions.
I continue to be intrigued about using the TOC Thinking Process more and more in my work and my life. It is not something I regularly practice, and I appreciate the results that I see other people getting from using the techniques. One of the sessions today was on "TOC in personal life," which I assumed was going to be along the lines of my personal effectiveness thinking. It turned out to be more of a discussion of how TOC helps the panelists think about problems, whether those were personal or local or even larger. Many of the pieces that they suggested reflected back on what I heard on the first day around the mistakes of management or the four obstacles to personal growth. And now I am asking the attendees about the Odyssey program and the Jonah program. Or maybe I should just start trying what I know of the tools and see where I go.
I sat in on the Buffers track with a talk by Roy Stratton of Nottingham Trent University on "Buffer Management in Context." The context is that of other researchers and implementers who have looked at similar problems. He talked about the four main functions of buffer management as
- Improving Flow
- Prevent overproduction
- Abolish local efficiencies
- Focus activity to balance flow
And then he used this to talk about how buffer management relates to Shewhart's control charts. Stratton also talked about how the concepts of buffer management fit with the "functions of Kanban" as defined by Ohno. There is a lot of overlap and similarity - the idea of "buffer" is not antithetical to Lean, as many people believe. There was some suggestion that buffers can hide the source of problems, whereas production kanban highlights the specific areas right away. Of course, this generated some discussion.
Dee Jacobs of the AGI-Goldratt Institue gave an informative talk called "Goldilocks and the Three Buffers" with an emphasis on project management and sizing the buffers. You don't want buffers that are too small or too big. You want them just right. Her talk was full of entertaining anecdotes and her experiences in implementation. The sizing discussion talked about the various statistical approaches and the common Rule of Thumb used in CCPM, and talked about their assumptions. In short, none of the approaches are perfect. You need to apply your brain and the intuition of your clients ("people are good"). She was particularly interested in situations where the amount of task uncertainty varies within the project. For particularly uncertain tasks, these should contribute proportionally MORE to the overall project buffer. In some cases, maybe all the uncertainty of the task should be incorporated into the buffer. I wonder if there might be other ways to protect the project - is it possible to allocate the buffer directly to the area where the high-uncertainty task lies? In case it isn't obvious, she is definitely in the camp of explicitly requesting at least two durations for every task ("highly probable" and "aggressive, but doable").
An aspect of Jacobs' talk I particularly liked was what she termed "immunization" of projects. As in a CCPM implementation should successfully immunize projects, so that they finish on time. If the immunization does not "take," then you need to look at the elements of that immunization to see what was missing. Look for the evidence. The elements of immunization: good project networks, a defined critical chain, buffer sizing, following task prioritization, active buffer management. In other words: Don't just watch the buffer go from green, to yellow, to red. Do something!
Of course, all this is topped off with interesting, entertaining, intriguing hallway conversations.
[Photo: image from classic Memorex commercial found here]