Archive : Goldratt – Critical Chain Project Management
Critical Chain – the Critical Chain is defined as the longest chain [not path] of dependent tasks. In this case, ‘dependent’ refers to resources and resource contention across tasks/projects as well as the sequence and logical dependencies of the tasks themselves. This differs from the Critical Path Method.
Estimations – To reduce the behaviours and time wasting associated with having too much embedded safety, Critical Chain Project Management recommends that task estimates are cut to half the length of a “normal” duration.
Safety – Critical Chain Project Management uses safety ‘Buffers’ to manage the impact of variation and uncertainty around projects. The safety at a task level is aggregated and moved to strategic points in the project flow. There are three types of buffer/strategic points necessary to ensure the project has sufficient safety:
Project Buffer – A project buffer is inserted at the end of the project network between the last task and the completion date. Any delays on the longest chain of dependant tasks will consume some of the buffer but will leave the completion date unchanged and so protect the project. The project buffer is typically recommended to be half the size of the safety time taken out, resulting in a project that is planned to be 75% of a “traditional” project network.
Feeding Buffers – delays on paths of tasks feeding into the longest chain can impact the project by delaying a subsequent task on the Critical Chain. To protect against this, feeding buffers are inserted between the last task on a feeding path and the Critical Chain. The feeding buffer is typically recommended to be half the size of the safety time taken out of the feeding path.
Resource Buffers – Resource buffers can be set alongside of the Critical Chain to ensure that the appropriate people and skills are available to work on the Critical Chain tasks as soon as needed.
Execution
Priorities – All resources on a project are given clear and aligned priorities relating to the ‘health’ of the Critical Chain relative to its associated buffer and hence the project as a whole. A resource with more than one task open should normally be assigned to complete any task jeopardising any projects Critical Chain before completing any feeding path task.
Completion – resources on a task are encouraged to follow the ‘roadrunner’ approach. When there is work available it should be progressed at the fastest possible speed (without compromising quality) until completed. Tasks are not left partially complete to remove the temptation to multitask. As task duration estimates have reduced safety they drive resources to meet the more “aggressive” durations and limit the behaviours of Student Syndrome and Parkinson’s Law.
Review
Buffer Management – the amount each buffer is consumed relative to project progress tells us how badly the delays are effecting our committed delivery date. If the variation throughout the project is uniform then the project should consume its project buffer at the same rate tasks are completed. The result is a project completed with the buffer fully consumed on the day it was estimated and committed. Project Managers determine the corrective actions necessary to ‘recover’ buffer time at points in the project where the buffer consumption is occurring faster than the project is progressing.
Remaining duration – tasks are monitored on their remaining duration, not their percentage complete. Resources report upon tasks in progress based on the number of days they estimate until the task will be complete. If the remaining duration stays static or increases, then Project Managers and Resource Managers “watching the buffers” know exactly where a blockage or potential delay is occurring and can take decisive action quickly to recover.
