Archive for November, 2013


Posted November 24, 2013 By Fletcher Groves

(excerpted from The Pipeline)


“Which system requires the most capacity?”, asked the intrepid, results-based consultant, pointing to the two diagrams depicting the effect of multitasking.  “Which system carries the most work-in-process?  Which one has the most complexity?  Which one costs more?”

“The one that kills more of my brain cells”, answered one superintendent.

“Let’s try this again”, she said.  “It’s an illustration of the finite capacity and variability buffering we have been talking about.  If you fix the amount of capacity and increase the amount of inventory, and do nothing about variation and uncertainty, the system has no choice but to lengthen cycle time.  That’s the undesirable effect of multi-tasking.

“But, multi-tasking is not the only behavioral tendency that causes the long durations that we are concerned with”, she continued.  “We pad durations to create safety, but then we waste all that safety, by waiting until the last moment, by letting the work expand to the allowed time, and – yes – by working on multiple jobs at the same time.”

“That might be the case in ‘real’ project management, but not on my jobs”, argued another superintendent, one of the younger ones.  “My framers have full crews on each job.  Same with the electricians, plumbers, masons, and everyone else.”

“Bulls–t”, said the VP of Construction.

“I can’t tell you how many times your framer leaves three guys on your job and sends the rest of the crew to get another job started”, he said.  “It happens all the time.  It happened to you last week.  The truss package showed up a week late on Lot 40.  The same day, the trusses showed up on Lot 47 three days early.

“Two jobs with trusses on the ground, and one framing crew to fly them.  Where’s the crane?  On Lot 47.  Which job is behind schedule?  Lot 40.  So, why don’t we send the crane to Lot 40?  Because we can’t.  Lot 40 has a purchase order issued to another crane operator.

“A lot of times, it’s not even their fault.  We tell them to do something, because we can’t seem to get our own act together on resolving the resource conflicts between open jobs.

“But – it’s not one behavior.  It’s all of it – the multi-tasking, the waiting until the last minute, the pacing of work to consume the allotted time.  There is variation in all its forms.  Different outputs from repeated applications.  There is risk and uncertainty.  Things that are indefinite, indeterminate, and unknown.  Weather.  Delays.  Mistakes.  Failed inspections.

“Things that go wrong.  Murphy.

“And – we step right into it.

“If you can’t see this, then you will have to come up with a different explanation – a different excuse – for why your jobs average 180 days, when the schedule itself specifies 120 days, and everyone acknowledges that it should only take 90 days.”


(The Pipeline: A Picture of Homebuilding Production is available on the publisher website (, through the author website (, as well as,, and


Pipeline Workshops: Compressing the Learning Curve

Posted November 17, 2013 By Fletcher Groves

The intrepid, results-based consultant opened the portfolio, removed a set of game boards, leaned them against the conference room wall, and then opened her notebook carryall, removed the zippered pouches containing the poker chips and sets of six-sided dice, and placed them on conference room table.

Removing one of the tablets from its case, she handed it to the technician, and said, “Do me a favor.  Synch this tablet to the projector and open the Excel worksheet for the standard production simulation, the one with balanced system capacity.”  Then, handing a game board to the two closest members of the team, she said, “I need you guys to clear some space for this board on the conference room table.”

As those tasks were being completed, she explained what was about to happen.

“Change is a necessary condition to any continuous improvement effort.  You can’t expect to get different results by doing the same things the same way.  However, change is difficult, disruptive, time-consuming, and costly;  moreover, the effort can fail to produce the intended result.  We need to be able to figure production management out, without the cost, disruption, and risk of failure associated with doing it in real life.

“What we need to do, is compress the learning curve.

“The new approach that you will use to manage production – and to thereby improve operating and financial performance – becomes intuitive and simple in practice, but there is a lot to understand.

“It is an approach that is counter to what most of you have been taught”, she said.  “You will need to fight your natural tendencies, challenge your paradigms, begin to think systemically.  It can be hard to initially grasp.

“In short – it must be learned.

“That kind of learning is a harsh teacher when it occurs at the cost of real operating performance and actual business outcomes.

“So – we are going to simulate the environment in which production decisions must be made, and we are going to measure the outcome of those decisions, in terms of both operating performance and economic return.  It’s going to be competitive, the pace is going to be fast, the situation will change rapidly.

“There will be uncertainty, risk, and variation;  you will have to learn how to reduce it, and then, how to manage it.

“The purpose is to create learning based on what you experience and do, not simply what you hear and read.  The objective is to compress the learning curve.  The production situations we are going to simulate reflect the circumstances you encounter in the real world, but they will be simple, fast, easy to see and understand.  We will modify them and run them until we learn to see, until we understand how to apply the principles.”

“Here, we have a homebuilding production system,” she explained, gesturing to the game board.  “It represents the resources and tasks necessary to complete a job, i.e., one of your houses;  the chips represent the system’s work-in-process – the jobs/houses at various stages of completion, as they flow through the production system.

“Variation is reflected by the probability in the dice rolls.  The tasks in each resource phase take the same amount of effort on every job, so there isn’t any variation in the amount of work, only in the length of time required to complete it.  We find comfort in the false security of achieving an ‘average duration’, and this game reflects that tendency.

“Each game represents a full year of operations and results;  each round represents a two-week time period in that year.

“The tablet uses its Excel worksheet to accept all of the data for each round – resource work rate, work-in-process, and completions;  it calculates the operating performance – cycle time, productivity, inventory turn;  since it knows the cost of the production capacity that the system incurs, it just has to calculate the business outcome – Return on Invested Assets, and Net Income.”

“Put the money in the bank”, said one of the superintendents, looking toward the Vice President of Sales.  “Start making it rain, Chief.”

The intrepid, results-based consultant smiled.  She had heard it all before.

“We’ll see.”


NOTE:  This post first appeared on Escape from Averageness in March 2012 as an excerpt from The Pipeline, titled “The Game”;  it is adapted here to explain the type of learning that occurs in a Pipeline workshop.  The Pipeline: A Picture of Homebuilding Production is available on the publisher website (, through the author website (, as well as,, and

The first Pipeline workshop is being held at the Ponte Vedra Inn and Club, Ponte Vedra Beach, Florida, on March 12-13, 2014.  Cost is $750.00.  Early registration (through December 10, 2013) is $600.00.  Delivered by SAI Consulting.  Sponsored by BuilderMT and Big Builder (Hanley Wood).

For more details:


Tri Pointe: Learning From The Past?

Posted November 10, 2013 By Fletcher Groves

The immediate thought that came to my mind after reading John McManus’s summation (in of Tri Pointe Homes’ $2.7 billion acquisition of Weyerhaeuser Co.’s homebuilding division?

Fortress Group.

I will acknowledge that other thoughts came to mind, and this particular thought did not come as a comparable transaction, or as a valid/fair comparison for how or why Tri Pointe has been assembled and how it will necessarily be managed, versus how Fortress Group was assembled and was then managed.

Rather, the chronicle of Fortress Group is simply a lesson – or a reminder – for any homebuilding enterprise considering the consolidation route through public ownership.

The story of Fortress Group was detailed by Gerry Donohue in the October 2002 issue of Big Builder:


It is a story that some of us who were variously involved with Fortress – Ed Horne, George Yeonas, Mike Hollister, Jamie Pirrello, myself, among others – have discussed from time-to-time.

One of a number of fashionable rollup consolidations of private companies in the 1990’s, Fortress Group completed its IPO in April 1996.  As Donohue reported, the anticipation was industry domination through economies of scale and access to public money, and the opportunity for owners of the acquired companies to liquefy their investments at a public multiple.

The IPO raised $27 million in equity and raised another $100 million in bonds, small by today’s standards.  After paying all of the expenses of its startup, Fortress had $11 million in cash and $100 million of long-term debt, resulting in a 9:1 D/E Ratio, which – for a startup – is only excessive if it doesn’t work, i.e., if growth doesn’t outpace the debt service.

Fortress started with four building companies and grew to 13 building companies by 1998;  by July 2002, it had sold all of its operating assets and ceased operations.  The terms and the amount of the bond financing was the issue.  During that four-year period, the company cannibalized itself, selling off building operations to service and retire debt, until it was too small to continue.

From early-on, the signs were there.  Fortress was generating nowhere near the margins that it needed.  But, signs are symptoms, not causes.

Donohue identified four factors that led to the demise:  (1) Fortress failed to act like a publicly-held enterprise, taking seriously its quarterly forecasts;  (2) it couldn’t attract the institutional investors that could create the active market Fortress needed to bid up share prices;  (3) other publics – like Pulte, DR Horton, and Lennar – were performing far better;  and (4) most importantly, Fortress wasn’t realizing the operational benefits from consolidation that it should have.

This last factor, Donohue said, was the most important.  Gross Margin, which should have risen due to economies of scale, actually fell.  Net Income Margin, which should have risen because of access to public financing, also fell.

One of the first problems Donohue identified was that the original team were financial people, not operators.  Bob Short, the owner of one of the four original companies that were rolled up, said, “They had little concept of how to run the company after the rollup was done.”

“The original concept of Fortress was for the individual operating companies to have a large amount of operational autonomy”, Donohue reported.  “The central office would merely be a source of capital.”  Short agreed:  “We had a great deal of diversity in the operations . . . on top of that, you had some pretty diverse personalities and strong egos.”

In 1999, the board brought in George Yeonas as CEO.  Yeonas worked to integrate the building companies into a single operational entity, a project involving almost everything in the operating model, including human resources, information technology, culture, and strategy.  This is the point where I was brought into the effort.  My small part was to document, improve, and standardize the business processes across the enterprise.

I can attest to some of  the deficiencies and the difficulties;  every building operation Fortress acquired had its own operating and information system, its own culture and business model, its own way of doing everything, with completely different community and plan portfolios;  the dissimilarities that existed between building operations are hard to overstate.  There was no way to even consolidate the financials – Fortress was a roll-up with no ability to roll-up.

In order to make business sense – in order to make economic sense – consolidation through merger and acquisition has to be about the quantum of value-added, about generating additional benefit in excess of additional cost, about creating/providing opportunity that would not otherwise be available.  Otherwise, it shouldn’t be done.  The task of increasing/expanding value-added might be easier to accomplish through de novo expansion than it is through the acquisition of existing operations, but the requirement still applies.

Donohue concluded:  “The company was a product of the belief that all that was necessary to succeed was public money and growth.  In every industry, small, privately-held companies were rolled up together and sold in the public markets . . . hailed as the model for the new millennium, they failed because of old-millennium realities – too much debt, and a lack of management, integration, and corporate culture.” (emphasis added)

What are the lessons?

I think one of the key issues – the one that must be resolved upfront – is to draw a distinction between the things that work best when standardized (code, in this case, for things that should become national standards), and the things that are deserving of more flexibility, more responsiveness to local requirements.

Homebuilding will never be just another type of manufacturing;  in particular, the supply chain for homebuilding is likely to remain radically different, and the degree of beneficial standardization will always be less than other industries.

Fortress made the mistake of buying whatever companies were available, without regard to how the vast differences would be bridged to build national standards.  Fortress had no plan, but it is not certain whether any plan would have worked.  There were too many differences.

Often, it is a decision of what needs to operate in the foreground, and what can be allowed to operate in the background, what needs to be customer-facing, and what does not.  Once the distinction between what must be standardized and what can be left to regional or local discretion has been clarified, I think it becomes an issue of getting the other things right.

Getting the right things right is not a new admonition.  It is the same one that Treacy and Wiersema were making almost 20 years ago when they wrote The Discipline of Market Leaders:  Choose a value proposition, commit to delivering extraordinary levels of distinctive value to a narrowly-defined segment of the value spectrum, then make the operating model – the processes, the organizational structure, the management systems, the type of employees, the culture – support the value proposition.

The projects and details of an implementation plan flow from there;  it’s involved, but simple enough to understand.  Surely, the management world that has to execute the current wave of IPOs understands the prerequisites and requirements this environment presents, and has prepared to address them.

Fortress failed to do it.  Surely, that lesson has been learned.

Hasn’t it?


“There are no free lunches.”

Posted November 3, 2013 By Fletcher Groves

(excerpted from The Pipeline)


“I have a couple of questions”, said a superintendent.  “The concept of variation buffering makes sense, but how does a production system sort all this out?  You can say that you won’t allow any buffering, but is that even possible?”

“No, it’s not”, replied the intrepid, results-based consultant, answering the second question first.  “It doesn’t matter if you say you won’t allow buffers.  A system will buffer variation.  The Law of Variability Buffering is no different than the Law of Gravity.  The question is only how it will buffer the variation.

“It is easier to take away the inventory and capacity buffer”, she continued.  “You can cap inventory or freeze hiring, and probably make it stick.  But, time is really a self-determining buffer.  Duration is the buffer overflow valve.  Absent some progress on reducing variation and some combination of higher inventory or excess capacity, you will simply wind up mandating cycle times that cannot possibly be achieved.

“There are no free lunches.

“Production performance hinges on these four interdependent measures:  variation, time, capacity, and inventory.”


(The Pipeline: A Picture of Homebuilding Production is available on the publisher website (, through the author website (, as well as,, and