The last note defined objective completion gates for the work moving through the traveling factory. With those gates applied consistently, the team can see how quickly each stage processes completed scope and where work begins to accumulate.
Each stage has its own crew, capacity and production rate. The stage that limits completed-work throughput is the bottleneck: the system's constraint.
Eliyahu M. Goldratt developed the Theory of Constraints and introduced many of its fundamental ideas in The Goal, the business novel he wrote with Jeff Cox.
One of the most memorable parts of the book has nothing to do with a factory. It takes place during a Boy Scout hike.
The bottleneck controls the pace.
As the scouts hike, the group continues spreading out. Some boys move quickly while others fall behind. The faster hikers can increase their pace, but that does not get the entire troop to its destination any sooner.
The troop can only move as quickly as its slowest hiker.
Once the slowest hiker is identified, the group moves him to the front and organizes the other hikers around his pace. They also remove weight from his backpack so he can move faster.
Improving the slowest hiker improves the pace of the entire troop. Making the hikers ahead of him faster only increases the distance between them.
That is the basic idea behind the Theory of Constraints. Every system has something limiting its ability to reach its goal. Improving anything other than that constraint may create more local production, but it will not necessarily improve the output of the system.
The hike finishes with the last person.
One trail, four hikers, and no passing. The slowest hiker limits everyone behind them. Compare a faster leader, a group paced to Herbie, and a lighter pack for Herbie. The last hiker determines how much trail the whole group has completed.
The fourth hiker cannot pass Herbie. The group finishes when the last person reaches camp.
Start by defining the goal.
Before we can find the constraint, we need to define what the erection operation is supposed to produce.
The goal is not maximum crane utilization. It is not the most tons, picks or welding hours possible from each individual crew.
The goal is to move completed and accepted work through the erection operation safely, correctly and profitably.
Goldratt used three measures to describe the performance of a production system:
- Throughput: The rate at which the system produces value.
- Inventory: The work and material that have entered the system but have not yet become throughput.
- Operating expense: The cost required to convert that inventory into throughput.
For this discussion, we can treat completed and released work packages as an operational measure of throughput. Sequences that have been started but not completed are work in progress. The labor, equipment and support required to complete them are operating expense.
Safety and quality remain necessary conditions. Improving throughput does not mean trading either of them for speed.
What if welding is the bottleneck?
Suppose the erection operation has the following capacities:
| Operation | Capacity |
|---|---|
| Erection | 1.0 sequence/week |
| Plumb and bolt | 1.1 sequences/week |
| Welding | 0.6 sequence/week |
| Decking | 0.9 sequence/week |
| Detail and closeout | 0.8 sequence/week |
Each operation has its own production rate. In this example, welding is the bottleneck.
System throughput: 0.6 sequence per week
The raising gang can erect one sequence per week, but welding can process only 0.6.
That makes welding the constraint. The entire system cannot consistently produce completed work any faster than welding can process it.
After five weeks, the raising gang may have erected five sequences while welding has processed the equivalent of only three. The difference becomes work in progress accumulating in front of the constraint.
The project may report strong tonnage and pick production. The crane may appear highly utilized. The operation has still produced only three welded sequences.
Would another raising gang help in this example? If welding stayed at 0.6 sequence per week, more erected steel would add to the backlog without increasing completed-work throughput. That suggests we should look at what is holding up welding before adding erection capacity.
Focus improvement on the constraint.
The Theory of Constraints gives us five focusing steps for improving the system:
- Identify the constraint.Determine which operation is limiting progress toward the goal.
- Exploit the constraint.Make the best use of the capacity already available. If welding is the constraint, keep it supplied with ready work, clear access, consumables, working equipment and timely inspection.
- Subordinate everything else.Organize the other operations around the constraint. Upstream crews should provide enough work to protect welding without releasing so much that incomplete sequences overwhelm the project.
- Elevate the constraint.If coordination alone is not enough, increase its capacity. That could mean additional qualified welders, another shift, better equipment, improved access or a change in how the work is sequenced.
- Repeat the process.Once welding is no longer the constraint, another part of the operation will control throughput.
“Exploit” does not mean pushing a crew beyond a safe or reasonable pace. It means removing avoidable waiting and making effective use of the capacity the operation already has.
On the Minneapolis stadium job, I was pulled into the paint yard because roof-truss painting was holding up the project. At the beginning and end of each shift, I climbed to the bridge crane platform, photographed the floor and logged the photos. I put reports together for the project management team, and every few days we reviewed them to understand how the work was moving.
One bottleneck was getting pieces flipped so painting could continue. We used timers, and when a piece was ready, available people paused their other work to help make the flip. We went from about a dozen trucks a week to roughly 25–28. We never reached the target of 30, but we substantially improved the projected schedule. What stuck with me was how a brief, coordinated effort at the right point could do more for throughput than everyone staying busy with their own task.
The bottleneck can move.
The constraint may not remain in the same place throughout the project.
Heavy lower-level steel may make the crane and raising gang the constraint. A later sequence with extensive field welding may shift it to welding. Access, deliveries, engineering, inspection or detail work may control another portion of the project.
The traveling factory changes as it moves through the building, and the bottleneck can move with it.
That is why the objective is not to permanently maximize any individual operation. It is to continually identify what is limiting completed-work throughput and organize the rest of the system around it.
The constraint is not necessarily the crew working the hardest or the operation attracting the most attention. It is the part of the system currently limiting progress toward the goal.
Once the constraint has been identified, the next step is to test the planned sequence, access, resources and handoffs to see if they actually protect it. That is where 4D becomes useful.