The first job is not to make the programme look recoverable. It is to discover what the organisation has been unable—or unwilling—to say.
Most programme rescues begin with a familiar ritual: the incoming programme director asks for the latest status pack.
I think that is often exactly the wrong place to start.
By the time a rescue is necessary, the organisation usually has plenty of reporting. What it lacks is a shared, reliable account of what is actually happening. The pack may be the most polished artefact on the programme. That does not make it the most truthful.
A status pack can tell you that a milestone is amber. It cannot, by itself, tell you whether the remaining scope is understood, whether a dependency has an owner, or whether anyone still believes the recovery date. Those things require evidence and conversation.
This is not necessarily a story of incompetence. Troubled programmes often operate inside incentives that reward confidence over evidence and escalation over uncomfortable truth. By the time the facts become undeniable, the organisation has spent months managing the narrative instead of the delivery.
The first 48 hours are therefore not about producing a definitive recovery plan. That would usually be premature. The aim is to establish enough ground truth to form a credible recovery hypothesis—and to identify what must be tested next.
Start with evidence, not the rescue narrative
Every troubled programme has a narrative: the vendor is late, the business keeps changing its mind, the architecture is wrong, or the previous leadership team was too optimistic.
Some of these explanations may be accurate. None should become the starting assumption.
I want to see the artefacts that describe the programme as it currently exists: the scope baseline, decision log, dependency map, release plan, outstanding defects, and recent acceptance evidence. I am less interested in whether these documents look complete than in whether they agree with one another.
Contradictions are useful. A release date that depends on a decision not yet scheduled, or a component marked complete while its acceptance criteria remain untested, tells you where to look next.
Interview for differences in reality
The first interviews should include people who can explain different parts of reality: the sponsor, programme leadership, business ownership, delivery leads, architecture, operations, and the supplier where relevant.
I am not asking everyone for a general opinion on the programme. I want to know:
- What must be true for the next meaningful outcome to be delivered?
- Which of those conditions is currently uncertain?
- What evidence would resolve that uncertainty?
- Who has the authority to act on the answer?
Different answers to the same question are more informative than polished alignment. They show where the programme has been coordinating appearances rather than resolving substance.
Do not redate a broken plan
A missed date is visible. Broken delivery logic is more dangerous.
If a milestone slipped because one discrete activity took longer than expected, the programme may recover through focused intervention. If the date depended on untested assumptions, unclear ownership, or work that could only ever have run in parallel on paper, moving the date changes nothing.
This is why I resist discussing a new timetable before understanding the old one. A recovery plan that preserves the logic that caused the failure is simply a more recent version of the same plan.
Turn unresolved questions into a decision agenda
Programmes often accumulate unresolved questions without recognising them as decisions.
Can the first release exclude a difficult segment? Who accepts the operational risk of a temporary workaround? Which team owns a cross-system data issue? Is the business prepared to change a process, or is the technology expected to reproduce it?
When these questions have no named owner and no deadline, delivery teams either stop or make implicit choices. Both outcomes are expensive.
A rescue needs a short decision agenda: the decision, its owner, the evidence required, and the date by which inaction becomes a delivery constraint.
The output of the first 48 hours
By the end of the first 48 hours, I want a concise view across five areas:
- Scope: What outcome is genuinely being delivered, and what is no longer certain?
- Decision ownership: Which choices are blocking movement, and who can settle them?
- Dependency reality: Which external conditions have evidence, owners, and dates?
- Delivery evidence: What has been demonstrated or accepted, rather than merely reported complete?
- Stakeholder confidence: Where do expectations differ so materially that they will undermine recovery?
This is not yet the recovery plan. It is a tested explanation of why the programme is stuck and what must change for movement to resume.
The distinction matters. A rescue should create clarity before it creates momentum. Otherwise, the organisation simply becomes better organised around the wrong recovery.
The first useful promise is not “I have a new date.” It is “I now know what we need to establish, who must act, and how we will know whether recovery is working.”