Critical Chain Project Management: A Manifesto for Change

One of the Velocity Accelerators – areas of learning – in a Pipeline workshop™ deals with the benefits-driven imperative for replacing the method currently used to schedule homebuilding production.

In terms of the workflow that has to be managed, the nature of homebuilding production is project portfolio management.  While it also has embedded, enabling, and supporting processes, homebuilding production is essentially the management of multiple projects that share resources.  It is the structuring and the management of a portfolio of job schedules, with interdependencies and interactions of tasks and resources.

The Critical Path Method (CPM) of project scheduling evolved from the Program Evaluation and Review Technique (PERT) in the 1950s;  it has been in existence for more than sixty years.  PERT and CPM were designed for programs with large, complex structures (Polaris weapons system, Manhattan Project), but CPM is now the de facto standard for scheduling all types of projects:  construction, aerospace/defense, software development, product development, research.

The problem with CPM is that it was not designed for managing a portfolio of projects, and it was not designed to function in environments where velocity is important, where faster cycle time and higher inventory turns are critical drivers of business outcomes.  In environments where it must contend with variation and uncertainty, CPM offers only a buffer of additional time – individual task durations lengthened (padded) to protect the completion date of the task, but not the completion date of the project.

(Note:  you do have to hand it to builders that assert fast cycle times, but remain oblivious to the effects of variation on their production systems.  Whether the protection mandated by production physics is in the form of longer duration, higher work-in-process, or excess capacity, the cost of that variation is the same:  the cost is the Gross Income lost from all of the closings that never occurred, from houses that were never built with the capacity that was paid-for;  and, it is Gross Income that likely would have dropped straight to a builder’s bottom-line, as Net Income, and ultimately, Net Profit.)

Moreover, while CPM considers task dependency (the predecessor-successor relationships of tasks) in its work breakdown structure, it does not consider resource conflict;  it does not consider situations in which tasks of different projects/jobs in a portfolio depend on the availability of resources that have insufficient capacity to meet the demand being placed upon them.

These two factors – dealing with variation and resolving resource conflict – are anathema to builders.

CPM was not designed to contend with the production environment homebuilding presents.  It might only be a symptom of the problem (the problem is variation and resource conflict), but CPM is benign to the solution.  ProChain Solutions’ Rob Newbold (Project Management in the Fast Lane) told me that he would go further, saying:  “CPM supports values that perpetuate the problems of homebuilders.”


Which brings us to Critical Chain Project Management.

Developed in 1997, Critical Chain addresses both task dependency and resource contention, and it replaces padded durations intended to protect task completion dates with buffers that protect the completion date of the project/job;  CCPM is much more aware of system constraints.  Most importantly, Critical Chain reduces the duration of projects – the cycle time of houses under construction.

Consider this illustration from The Pipeline: A Picture of Homebuilding Production©:

On a calendar basis, RB Builders has a construction schedule that calls for 120 days, but its achieved cycle time is 180 days.  Yet, there is widely-held agreement within the company that RB Builders should be able to build its homes in far-less than 120 days, because the schedule reflects “highly certain” task durations.

Statistically, the “long tail” of a standard deviation (bell) curve reflecting the difference between 50% confidence and 95% (“highly certain”) confidence lengthens the duration of project schedules by a factor of 1.64.  If RB Builders is highly-confident – 95% confident – it can complete a house in 120 days, it stands to reason that it would also be confident that it can complete half of those houses in no more than 74 days.

Since 50/50 outcomes are not acceptable, CCPM restores the confidence level by taking half of the task time difference between the 95% confidence (120 day) and 50% confidence (74 day) durations, and pooling the risk in a buffer that protects the project completion date from variation.  That results in a schedule of 97 days (74 schedule days + 23 buffer days), 23 days faster than the current schedule, and 83 days faster than the current cycle time.

The resulting reduction in cycle time would double current inventory turn;  RB Builders might have to trade-off margin for higher sales volume, but that would be acceptable, because higher velocity acting upon the same margin would nearly double Return on Assets;  and, depending on the overall condition of the market and market share, RB Builders may not face consideration of a margin trade-off.

Critical Chain Project Management does more than just reduce the length of construction schedules.  It also specifies a set of rules preventing behaviors that consume (and waste) the safety built into task durations.  It installs a release mechanism that pulls starts into the system and keeps work-in-process at the levels required to produce faster cycle times.  It implements simple, visual tools to manage production.

Builders can put some of these practices into place without changing the scheduling algorithms from Critical Path to Critical Chain:  they can change behavior by requiring ALAP starts (Newbold informs me that “requiring meaningful priorities” is better wording), and by eliminating ASAP starts and task due dates;  they can implement pull release mechanisms;  they can instill a relay race mentality to the transition of work between predecessor and successor tasks/resources.

But, builders would still be enslaved to a scheduling methodology that does not work well in its industry vertical.

Builders can be partially excused for a lack of awareness, for not knowing to insist on a CCPM option, for not being capable of developing their own scheduling modules.  But, where are the voices of the management technology companies on Critical Chain?

Why is there a settling for a status quo that doesn’t work for homebuilding, that doesn’t provide licensees, users, and clients with the best solution?  Where is the strategic discipline to provide management technology that contributes to sustainable competitive separation?

In my view, the onus is upon them.

Newbold is harsher:  “In my experience, technology companies will only go where their clients and competitors take them;  until clients demand, vendors will ignore.  You might be better off with a call to arms for the builders to demand better.”