TOC course · Lesson 3 of 6

Exploit the constraint, then subordinate the system

Get more goal output from the existing constraint—and stop the rest of the organization from undoing that decision.

Time: 12 minutes Previous: identify from evidence Reference: field guide

Two different decisions

Exploit

Decide how the constraint should spend its existing scarce time. Remove avoidable waste, bad inputs, preventable interruption, and low-value work.

Subordinate

Align release, priorities, upstream preparation, and downstream behavior with the exploitation decision—even when local utilization falls.

The canonical sequence matters: exploit before spending to elevate, and subordinate everything else to the exploitation decision. See the Five Focusing Steps.

Apply it to RFC clarification

MoveOperating ruleWhy it helps flow
ExploitRun two 45-minute decision clinics each week for RFCs with complete inputs.Protects scarce decision attention from fragmentation and creates a reliable cadence.
ExploitRequire next decision, owner, deadline, and dependency for every clarification item.Turns passive waiting into managed waiting.
ExploitRemove closed, duplicate, or commercially rejected RFCs from active WIP.Prevents false demand from consuming review attention.
SubordinateDo not create a clarification RFC without an owner and next decision date; do not release incomplete work into the clinic.Upstream must prepare constraint-ready inputs and active WIP stays actionable.
SubordinateLimit new RFC starts when the clarification buffer is above its control limit.Stops excess WIP from lengthening feedback and lead time.
SubordinateProtect a small execution block only for RFCs already in plan-ready.Prevents improvement upstream from immediately stalling downstream without reserving capacity for unapproved work.

The local-efficiency trap

If sales is rewarded for RFCs opened, engineering for hours utilized, and finance for avoiding any commercial uncertainty, each function can look productive while the system produces fewer delivered and collected changes.

Subordination can look inefficient locally: intake may pause, non-constraint people may wait, and some specialists may help prepare better inputs instead of maximizing billable activity. Judge the decision by system throughput and guardrails.

Practice: classify the action

Type exploit or subordinate.

Your operating rule

We will exploit [constraint] by protecting it from [lost time].
Therefore [upstream/downstream group] will [subordination rule].
We will judge this by [goal-throughput measure], guarded by [necessary condition].
Ask follow-up questions: Bring one proposed rule to the agent. We can check whether it truly subordinates the system or merely shifts work elsewhere.

Primary reading