Five signals that a project is in more trouble than the reports suggest

The first signal is a change in the language of the risk register. A risk register that was previously written in specific, measurable terms — 'vendor delivery of Module 3 is 14 days behind schedule, with a projected cost impact on the Phase 2 budget' — and has shifted to vague, general terms — 'vendor performance is being monitored' — is a risk register that is no longer being used as a management tool. It is being used as a compliance document.

The second signal is a change in the composition of the steering committee. When the people who attend the steering committee change — when the senior sponsor is replaced by a delegate, when the CFO stops attending, when the external advisors are no longer invited — the project has lost executive attention. Projects that lose executive attention do not recover without it.

The third signal is a recovery plan that has already been revised. A project that has produced one recovery plan is a project that got into trouble. A project that has produced two recovery plans is a project that did not fix the underlying problem the first time. A project that has produced three recovery plans is a project that does not know what the underlying problem is.

The fourth signal is a change in the definition of success. When the project's success criteria change mid-delivery — when 'on time and on budget' becomes 'delivering core functionality' becomes 'delivering a minimum viable product' — the project is not being managed. It is being renegotiated. The renegotiation is always in the direction of less.

The fifth signal is the absence of a named recovery owner. When a project is in trouble, someone needs to own the recovery. Not the project manager — they own the delivery. Not the program director — they own the portfolio. A named individual whose sole responsibility is to diagnose the problem, design the recovery, and drive the implementation. The absence of that person is the absence of a recovery.