The consulting mandate and decision boundary

A consulting assignment may begin with an apparently clear request: “Fix the portal project's delay within two weeks.” That sentence identifies the sponsor's frustration, but it does not specify a testable problem. The delay may be measured against a baseline approved before scope changed, against an informal promise, or against a security approval that has not been granted. A reported percentage complete may omit substantial testing. A consultant should not promise to correct a condition whose mechanism is unknown or whose remedy lies outside the mandate. The first professional output is a decision frame: what will be examined, what evidence will count, which options will be compared, and who can authorize action.

Separate four items in the entry meeting. The symptom is what the sponsor observes, such as a release that did not occur on the announced date. The question identifies what must be learned, such as why the approved baseline and current forecast differ. The deliverable may be a diagnostic memorandum and options, rather than the release itself. Decision authority identifies who may change scope, funding, contractual commitments, or acceptance criteria. Ask for the approved baseline and its approval date, the change log, the benefit owner, and restrictions on data access. If no approved baseline exists, do not manufacture one retrospectively. Record that limitation and propose a formally authorized rebaseline before making firm variance claims.

Draw a boundary between advising and managing. A consultant can inspect plans, interview participants, calculate consequences, and recommend controls. The request “solve the problem” alone does not authorize directing a supplier's staff, approving expenditure, revising a contract date, or assigning blame to the project manager. A contract may expressly delegate some execution power. Record the exact power, grantor, duration, and constraints; do not enlarge it by implication. A compact decision-rights table contains the action, proposer, approver, implementer, and evidence of approval. It keeps disagreement about an option separate from the question of who has permission to choose it.

Specify acceptance criteria for the advisory work itself. For example, within ten working days after receiving essential records, the consultant will deliver an analysis of schedule variance, three feasible interventions under available-resource constraints, and a recommendation showing each option's effects on time, cost, risk, and required approvals. The sponsor can accept that the analysis is complete and its assumptions clear without accepting the recommended intervention. List records the consultant may access and boundaries such as no contact with the end customer until consent is given. Starting the clock at receipt of critical inputs prevents a vague kickoff meeting from consuming the analysis period.

Check independence before examining the case. Compensation linked to the purchase of a tool that the consultant sells, or a close relationship with a supplier manager, can affect an option appraisal even if the arithmetic is correct. Disclose the relationship, separate the recommendation from a sales interest, and ask for an unaffected decision maker to review the choice where appropriate. PMI's Code of Ethics presents responsibility, respect, fairness, and honesty as values for its members and credential holders. Those values provide a useful ethical example for disclosure; the code is not asserted here as a universal law governing every consultant. Limit access to personal or supplier information to the agreed purpose and permission.

Suppose the sponsor asks for “a report proving the contractor caused the delay.” The requested wording fixes a conclusion before evidence is gathered. A defensible alternative is “assess the contribution of supplier dependencies, client approvals, and changes to the variance, identifying what remains unresolved.” If the sponsor will not release the change log or allow the other party to be heard, label the conclusion provisional and conditional. A polished sentence cannot replace missing evidence. Sometimes the proper response is to narrow the task to available records; sometimes it is to decline a mandate that forbids inconvenient findings. Either response protects the practical value of advice.

For practice, draft a brief engagement letter stating the client's description without adopting its proposed cause; the questions to test; the records and interviews required; conditional deliverables and timing; exclusions; the recommendation's decision owner; and a process for correcting disputed facts. Maintain an assumption register with an owner, verification date, and consequence if false. The assumption that 15 June was the approved baseline changes the entire delay conclusion and must be checked before analyzing speed. The assumption that a supplier interview can occur tomorrow affects the work plan but does not establish a cause of delay.