Transition to Critical Chain Multi-Project Management for Long Duration Projects
What to Do Until Buffer Management Kicks In
The transition from traditional project management to Critical Chain Project Management (CCPM) in a multi-project environment presents a formidable problem with projects of long duration. A simple method is presented for that transition and provides the metrics necessary to directly encourage and cement the behaviors needed for Critical Chain Multi-Project Management. This paper assumes the reader is familiar with CCPM.
The Multi-Project Implementation
This paper focuses on the period of time from planning the first Critical Chain (CC) project, the cut-over project, to completion of the last traditionally managed project. This can be a long period of time before the company has fully implemented Critical Chain Project Management. Theory of Constraints (TOC) practitioners involved in Critical Chain Mulit-Project Management (CCMPM), often find this transition to be the toughest part of an implementation.
The Implementation Conflict
In order to successfully implement Critical Chain Multi-Project Management, we must obtain support for it. Everyone expects that CCPM will be another flavor-of-the-month implementation that fades away if properly ignored. To obtain that support, we must start with one project to prove that CCPM works. And to be successful, we must change the whole project system to CCMPM. Because Critical Chain requires Buffer Management and traditional projects can’t use it, we must implement CC on all projects at the same time.
Implement One Critical Chain Project First
Even though we know it works, we must prove that it works “here!” A common solution is to use a pilot (trial) project as a way to demonstrate CCPM and get the bugs out of the existing system. One project at a time is much simpler to implement than many. The pilot project should not be thought of as a trial. It’s really the first Critical Chain (CC) project, the cut-over project. Every new project following it will also be a CC project.
Typically, for a transition, the cut-over project is planned while the work-in-process is ignored. But in a multi-project management environment, that means that some or many shared resources will be fought over by the CC and non-CC projects. The resources are usually expected to multitask and have several projects in work at one time. Multitasking is a huge factor in projects being slow. How can scarce resources be assigned where they are most needed, if the statuses of these projects are measured differently?
The common approach to adding a new project to the pipeline of projects is to commit to a date and put it in the system. With little understanding of the amount of work in the system and the system’s capacity, work is pushed in with the expectation that it will get done.
With a system full of work-in-process projects, it will take a long time to complete this first CC project. Continued multitasking between projects will assure it. The reality is that people are asked to not multitask on the CC project while they are multitasking on the others. The non-CC projects will delay the faster, CC project. It will be difficult to determine and measure the Critical Chain project’s success compared to the others. Some people will believe it gets special attention and will demand to share its resources.
The more difficult problem is the lack of Critical Chain buffer management. Lacking CC project buffers, traditional projects can’t use buffer management. Priorities among the projects may be determined by perceived urgency as expressed by the project managers. Implementing the first Critical Chain project has not always been easy.
Big Bang Approach
The whole project system can be changed in one massive replan of all projects. It may make a lot of sense since we know we won’t be done until all the projects are CC projects. All projects are measured the same way and they quickly get up to speed. Or do they? How does the whole system get changed? All of the projects must be re-planned and changed to CCPM by shortening the duration of many, many tasks of many projects.
In a small system, the big bang approach is a real option. In a large system, it is definitely much more challenging and probably not possible. To change all the projects to be Critical Chain projects requires re-planning while they are in progress. The same people that are working the projects are need to do the replan. It’s likely to be chaotic and it won’t happen overnight. Re-planning will delay the implementation, delay current projects and may jeopardize an initial (or any) success. Just the opposite of what was intended.
Delay Until the System is Ready
Do not insert the cut-over project until the resources can focus on it. Prioritize the projects. Since any prioritization is effective in increasing the speed of a system, use the commitment dates as priorities to help determine what to focus attention on. Propose a drum resource and plan the release of the cut-over project to be synchronized with this drum. That sets up the next issue. How do resources (and management) know what to work on next? We need buffer management. We still can’t have it.
Unfortunately, it is not possible to start with a clean slate, no projects. We must deal with the work as it is in the system. It looks like we have to wait to use buffer management until after all projects in the system are CC projects. We still have an implementation conflict.
A New Approach
Create a method of comparing a Critical Chain project’s status with a traditionally managed project’s status, while promoting better behaviors.
(1) Prioritizing the work allows us to recognize that some work may be low enough priority to be delayed or canceled. Use buffer management on the first CC project, and create a kind of virtual buffer for the other projects. Then use virtual buffer management on all of those projects without re-planning them.
(2) Collect status for all projects as “How long until you are done with your task?” If percent complete is provided, accept it and restate it back as, “Does that mean you have 5 days of work remaining and you expect to be finished by next Wednesday?” Also ask, “Is there anything else you are working on?” Be consistent and persistent in asking for work remaining. Don’t argue about it. Accept whatever they give you. Reality will show up eventually.
(3) For each main chain of tasks (the Critical Path) and each feeding chain, compare the planned (base) finish with the current expected finish. The status (days ahead or behind) relative to the plan indicates how it is doing. This same calculation is done for Critical Chain’s buffer management and is called buffer incursion (in days).
(4) This information is used to manage the existing projects with their current due dates, without adding buffers to them, to create an unbuffered management report. The process is to prepare the existing projects by inserting a milestone at the end of the project, and between each feeding chain and the critical path. The milestone, being the last task in the chain, indicates the planned finish of the chain. As status is added, the expected finish of the current task pushes all successors to the future or pulls them earlier. Do NOT recalculate the critical path unless it makes a significant difference to the flow.
(5) Compare the current expected finish date with the base milestone (planned) finish date. This becomes an unbuffered incursion and can be reported and/or plotted for each chain of the project. Unbuffered Management can be used for all the projects, including the Critical Chain project. This provides a way to compare the health of all of the projects and a gives a basis for assigning scarce resources. The Critical Chain project would also have a Critical Chain Fever Chart and Buffer Report.
Create a chart with % Complete on the X-axis and Days Ahead/Behind on the vertical axis. The chart will have characteristics like a fever chart. Place a zero line horizontally (exactly on schedule), and plot days behind above and days ahead below the line. Like the fever chart, it is a visual indicator that the projects are gaining or losing ground. The chart indicates how each the project is doing and its likelihood of completing on time. It has a virtual buffer. The buffer is really not there, but its usefulness is.
Traditionally managed projects typically have significant safety in each task in a futile effort to get every task completed on time. Most project managers either believe they have little or no safety in their projects or they believe that their safety is a minimal requirement to maintaining their schedule. They have substantial experience to prove it. They know that time and Murphy are very fickle. By using unbuffered projects, they keep their original task estimates and project due date. By adjusting behaviors toward Critical Chain requirements, task safety is much less needed and will accumulate at the end of the project. All projects are likely to go faster than they were. Project Managers see real results on their existing projects and look like heroes.
Critical Chain Buffer Management provides focus for management attention to significantly improve project performance. Since it is extremely difficult to transition from a traditional project management system to CCMPM, a transition methodology providing tools similar to Critical Chain Buffer Management is a significant bridge for that gap. With prioritization and unbuffered management, attention is focused where needed. Then good behaviors and a Road Runner ethic are developed, with the focus on completing as soon as possible, rather than on meeting the due dates. All of the work takes advantage of unbuffered management and the whole system flows faster during the transition.
This methodology is only for the transition to Critical Chain Multi-Project Management. It is not to eliminate buffers. It puts all of the projects on a level playing field until the transition is complete.
What to do until Buffer Management kicks in? Be doing Unbuffered Management!
Copyright Skip Reedy, 2002, 2011
Reprint allowed with credit