Visibility, buffers and congestion

With WIP under control, kits ready and daily management running, one question remains: how do we see how each project is doing, and where to act?

The TOC way of scheduling: one buffer, at the end

Traditional plans pad every task. People know what can go wrong, so each estimate includes its own safety. Then that safety is wasted: work expands to fill the time available, effort starts near the deadline, and nobody reports finishing early because it would shrink their next estimate. Delays pass downstream; gains don’t.

Critical chain scheduling takes the safety out of each task and puts it together in one common buffer at the end of the project. The tasks are planned tight; the project as a whole is protected. Because safety pooled in one place covers the project’s uncertainty with far less total time than every task’s padding added up, projects finish earlier and more reliably.

Project status: one chart per project

A project’s status is the status of its longest chain of dependent work. Two numbers describe it:

• % of the longest chain complete
• % of the buffer used up

Plot one against the other over time and you get the project’s fever chart. Using a third of the buffer is fine when a third of the chain is done; using two thirds when only a third is done is not. The trend tells more than any single point:

The portfolio chart for senior management

Put every active project’s latest point on one chart, and senior management sees the whole portfolio at a glance. The rule for reviews: focus only on the few red projects, and review their recovery plans: what will move them back to green.

Two warnings:

• Red doesn’t override the priority sequence. A red project shows one project’s likely delay; it doesn’t jump the queue set by WIP control.
• Check congestion before acting on one project. If several projects are red because they’re all queuing at the same team, the fix is systemic, at that team, not a rescue of one project.

Gearing up project control

With fewer firefights, the planning team gets capacity back for two things:

• Look-ahead planning. Set short-term expectations, such as the next 20% of completion, and list the tasks needed to get there. Contractors know exactly what is expected next, and payments by percentage complete become simple.
• Issue management. Every task manager logs obstacles as they arise and names an owner. The project management group tracks how long issues have been open across the work and triggers resolution or escalation by elapsed time.

Congestion management

Even with good synchronisation, queues build up from uncertainty and variable capacity. So every resource manager watches the trend of queues in their own work centre: design, procurement, a critical machine. A rising queue is an early warning, time to decide calmly:

• Add capacity, outsource, or use overtime
• Multiskill people or reassign them to balance the flow
• At a critical machine with spare time, sell that capacity for extra revenue

The whole solution

1. WIP control: fewer projects open, in a fixed priority sequence.
2. Full kit: no phase starts until it’s ready.
3. High-frequency management: daily, by elapsed time.
4. Planning and visibility: one buffer per project, fever and portfolio charts.
5. Congestion management: watch queues, act on capacity early.

Plug the internal wastages this way, and there is enough time left to absorb the external delays and uncertainty no project can avoid. When time is managed well, cost, scope and quality fall into place.

Check your understanding

Answer all five. Get four right to complete the lesson.

1. In critical chain scheduling, safety time is…

• Added to every task
• Removed from tasks and pooled in one buffer at the end of the project
• Removed entirely
• Given to the customer

Pooled safety protects the project with less total time.

2. A fever chart plots…

• Cost against time
• Tasks against people
• Scope against quality
• % of the longest chain complete against % of buffer used

Those two numbers describe a project’s status.

3. A vertical line upwards on a fever chart means…

• Buffer is being used while the chain isn’t moving: work has stopped
• The project is ahead of plan
• Rework has been found
• Everything is fine

Find what is blocking the chain.

4. Several projects turn red because they all queue at the same team. The right response is…

• Rescue each project separately
• Change their priorities
• A systemic fix at that team: capacity, overtime, rebalancing
• Ignore it

Check congestion before acting on one project.

5. Does a red project jump to the top of the priority list?

• Yes, always
• No: red shows a likely delay; the priority sequence still holds
• Only on Mondays
• Only if the customer is big

Red triggers a recovery plan, not a re-ranking.