On Time, On Budget

Clark Ellis, a Pipeline workshop™ colleague of mine, recently published an article in Big Builder, titled “Budget Buster No. 1,” about the inattention paid to production capacity in the annual budgeting process. It is a worthwhile read. In it, Clark attributed The Pipeline: A Picture of Homebuilding Production, Second Edition© for the important explanation of calculated cycle time.

Here is the full picture, excerpted from the book itself, and covered in every Pipeline workshop™.

The intrepid, results-based consultant walked to the front of the conference room, now filled with a cross-section of teammates and leaders responsible for production.

On the erasable board, she started a list:

Pipeline

“We have a lot of work to do”, she said. “The clearest picture–the best visual image–we can convey of RB Builders’ production system is that of a pipeline. That being the case, what is the purpose of this pipeline?

“What does it do? What does it carry, and what does it deliver?”

“I would say that the pipeline does two things”, offered the VP of Construction. “It carries our work-in-process–it carries houses under construction–and it delivers closings – completed homes. So–its purpose is to produce completed homes, and generate Revenue from the closings that ensue.”

The intrepid, results-based consultant added to her list.

• Pipeline
• Pipeline Size v. Pipeline Capacity
• Cycle Time
• Work-in-Process
• Throughput

“Okay. Then what is the capacity of the pipeline to do that? How many houses – how much work-in-process – can the pipe carry?”, she asked, “How do houses get into the pipeline? And – how many closings is it supposed to produce?”

“As many as we can put in it. However we want to put them in it. Whenever we want to put them in”, one of the superintendents quipped. “Okay. We’re told we’re supposed to generate an even and sufficient rate of sales, starts, and closings.

“That part makes sense. We just can’t seem to achieve it.

“And–if we could smooth-out our rate of sales, starts, and closings – then we could probably also manage to maintain a consistent level of work-in-process in the system.

“That part makes sense, too.

“But–as for the capacity of the pipe–apparently we think it has unlimited capacity, because every start we put in the pipeline will eventually be completed and closed. As for the output–the throughput–that’s a budgeted number of completed houses that turn into closings every year, sometimes we make it, sometimes we don’t.

“As far as how houses actually get into the pipeline, there is a start matrix, which acts as the pipe’s control valve. Under the old production system, the start matrix prescribed both the order and rate of starts, and ‘pushed’ the starts into the system. Under the new production system, the start matrix only prescribes the order; houses are supposed to be ‘pulled’ into the system at the rate of closings.”

“Okay”, said the intrepid, results-based consultant. “I have several questions. First–is there a difference between the size of the pipe and its capacity? Second, how many homes should you have under construction–how much WIP do you need–in order to reach your budgeted closings? Third–how long is it supposed to take you to build a house, and how long does it actually take you?”

“Yes–I suppose there is a difference between size and capacity”, said the VP of Construction. “The size of the pipeline would be defined by the amount of work-in-process, while the capacity of the pipeline would be a function of output in relation to size.

“There is a limit to how much it can hold, so–yes–the size of the pipe is finite.

“Yes–there is a connection between how much the pipe can hold and how much it can produce. We think the pipe should be able to hold 100 houses, and we think the pipeline should be able to produce 240 completed houses a year–at least, that’s the budget – which is 20 closings per month. So, you could say that the size of our pipeline is 100 houses, and its capacity is 20 completed houses per month.

“In terms of our cycle time, it varies depending on the house plan, but our construction schedules call for an average of 120 days”, the VP of Construction continued. “We know we are nowhere near that fast. Most of our homes finish late. I would say that eighty (80%) percent of our houses take between 160 and 200 days.”

“If you close 240 homes with 100 houses in WIP, your cycle time is 150 days”, said the intrepid, results-based consultant, “30 days longer than the schedule. But, this year, you are only on track to close 200 homes, meaning your cycle time is pushing 180 days.”

“Then we agree”, replied the VP of Construction. “The way we measure it, our cycle time has been averaging around 180 days. There is considerable variation, particularly on individual jobs; some take more time, some take less time. But, the overall average is around 180 days.”

The intrepid, results-based consultant moved to the erasable board at the front of the conference room, selected an erasable marker, and wrote:

CT = 120 days WIP = 80 houses T = 240 homes

CT = (WIP ÷ T) x Days (80 ÷ 240) x 360 = 120 days

WIP = (CT x T) ÷ Days (120 x 240) ÷ 360 = 80 houses under construction

T = (WIP ÷ CT) x Days (80 ÷ 120) x 360 = 240 closings

“There is, in fact, a direct connection”, she said. “There is an accepted, proven mathematical relationship between cycle time (CT), the level of work-in-process (WIP), and the throughput (T) of a process. So–if you know two of the values, you can always calculate the third value.

“There are two laws of production that deal with this relationship. The one I just mentioned, called Little’s Law, and one which we call the Law of Variability Buffering, which tells us that every system will protect itself from unplanned variation and uncertainty with some combination of – you guessed it – longer cycle time, more inventory (work-in-process), or excess/unused capacity.

“Regarding the difference between calculated cycle time and measured cycle time, there are two ways to determine it. You have measured cycle time, which is the average, or mean duration, of a set of jobs; its value lies in examining the forensics of individual jobs, in order to eliminate the causes of problems, variation, and waste. It’s where we use kaizen, PDCA.

“The value of measuring average cycle time lies in the selection of individual jobs to examine, because they are outside the lines of statistical control.

“And, you have calculated cycle time, which is the relationship between completed jobs and work-in-process; the value of calculated cycle time is in providing a picture of the condition and performance of the entire system.

“In terms of what is actionable, measured cycle time is actionable at the level of every future job based on the forensics of past jobs, while calculated cycle time is actionable in ways that affect the performance of the entire system.”