Choose the intervention the evidence supports

A delivery problem can prompt several plausible responses: limit new work, add capacity, redesign a hand-off. Choosing among them requires an account of what is happening. The cheapest option is only useful when it addresses the observed mechanism.

Reversibility is a practical consideration. If the diagnosis is wrong, a bounded trial may be easier to stop than an organisation-wide redesign. But reversibility does not establish that the trial is relevant.

Evaluate before prescribing

Examine whether excess intake, scarce expertise or a structural hand-off problem is supported by the evidence. These explanations can call for different actions. A team does not need to perform every cheaper intervention before it is allowed to address a demonstrated capacity or workflow problem.

The order in which questions are considered can reduce the cost of a mistaken diagnosis. It should not override the answer to those questions.

A policy change still has a cost

Changing an intake policy may require little procurement spend. It can still consume leadership attention, change priorities and displace planned work. Calling it free conceals the decisions needed to make it operate.

A useful proposal names the cost and trade-off in the organisation's own terms. It also names who has the authority to act. A policy written on a board has little effect if a different person can bypass it whenever the next urgent request arrives.

Size the commitment to the evidence

Where the evidence is incomplete, consider a small intervention with a declared purpose and a review point. Define what would support retaining it, what adverse effects matter and what would cause the team to reconsider.

An intervention can be useful without proving a universal causal rule. The claim needs to remain within the observed work, period and conditions. Simultaneous changes make that interpretation harder, so acknowledge them rather than assigning the whole result to the most visible action.

Carry the decision through

Implementation is part of the intervention. Time has to be allocated, ownership established and the changed practice used. A recommendation that never changes how work is handled has not tested the underlying hypothesis.

The next review should ask what actually happened and whether the result supports the next commitment. That keeps a reversible experiment from becoming an unexplained permanent rule.