THE OPERATING PROBLEM
Two production systems, and we only control one of them.
A manufacture-and-install business designs, engineers, manufactures, and installs custom-built products into buildings that are still under construction. It runs two production systems at once. One is a shop the company controls. The other is a building other people are still changing, on their own schedules, for their own reasons.
The shop gets watched closely and all the time. The building usually gets checked loosely and once in a while. The paper argues that difference creates a blind spot in how a company decides when to commit work into production, and that the cost of it shows up later, somewhere else, on somebody else's line.
This is engineer-to-order construction: millwork production planning, specialty-trade operations, and the site readiness question sitting underneath both.
DEFINITION
Production Commitment Point (PCP)
The point at which labor, material, and production capacity become committed to a specific room or product such that reversing the decision becomes increasingly costly.
Where that point sits depends on the shop. It might be first cut, material hitting the saw, a batch going into a run, or ordering material against a job. Before it, changing your mind is cheap. After it, changing your mind costs you. Every shop has that point. Most never name it.
TWO INPUTS, ONE DECISION
Shop information and physical site information are not managed the same way.
SHOP CAPACITY & LOAD
Watched daily. Tracked in detail.
Ask a shop of any size for machine hours by department, load by week, backlog, bottleneck. Most can hand it to you.
PHYSICAL SITE CONDITIONS
Checked now and then. Reported in general.
Ask the same shop how it knows a room is really ready to receive finished product. The answer is usually a call with the superintendent, plus a project manager's read on the schedule.
THE SYSTEM, DRAWN
tracked in detail
reported in general
One decision, two inputs, and we know far less about one of them.
RESPONSIVE RECREATION OF THE DIAGRAM ON PAGE 7
THE ARGUMENT, IN THE AUTHOR'S OWN TERMS
OBSERVATION
In the shops I've worked in, visited, or dealt with, we run the shop on numbers that get updated constantly. We run site readiness on conversations. Phone calls, emails, meetings, a visit now and then, and whatever the project manager has learned over the years.
INFERENCE
We reach the Production Commitment Point with good shop numbers and rough site information. So the decision may not account for what's really happening at the site. What that costs us shows up later as rework, delay, storage, and field changes. It never gets charged back to the commitment that caused it.
HYPOTHESIS
Rooms committed after somebody checked the conditions that matter, while checking them was still possible, will come out better than rooms committed without that. Better meaning fewer rework hours, fewer field change hours, and more dates hit.
The author labels these deliberately. The observation comes from his own experience and he asks for it to be tested against a wider sample. The inference he calls sound but unproven. The hypothesis he calls a proposal.
FOUR THINGS WORTH TAKING AWAY
- Some conditions have a window
- Blocking can be seen after the framer sets it and before the drywall goes up. Miss that window and you do not have bad information. You have no information, and no way to get it.
- Some conditions are a history, not a state
- Conditioned space is the example everybody knows. A reading on Tuesday says nothing about the three weeks before, and by the time the result is visible the panels have already cupped.
- The cost lands on a different line
- Checking costs a person, a day, and fuel, now, on a job that is not installing yet. Not checking costs the crew, later, in a different month. So the trip never gets made and nobody is wrong for not making it.
- Jobs do not get ready. Rooms do.
- Most scheduling systems still treat a job with one room ready and one room six weeks out as a single unit. The room is probably the better unit for commitment, though the author holds that one loosely.
WHAT THE AUTHOR IS ASKING FOR
If you run a shop and you have measured any part of this, he would like to hear about it, especially if what you measured says he is wrong.
If you work in operations research or construction management and know of work that already covers this, point him at it. Operators, researchers, and construction-management professionals are invited to respond with data, criticism, or relevant research.
Respond to the paperThis operating problem is part of the field-to-office gap that led Brian Milburn to begin developing Praxon.
SEE THE PRODUCT →