By the time I'm called into a troubled programme there are usually three versions of the truth: the status report, the corridor conversation, and the one the team stopped telling anyone months ago. Nobody believes any of them. That is the condition being treated, not the schedule.

A fair share of my thirty years has been spent on remediations and rescues. One banking remediation searched 500,000 home loans to find and compensate more than 10,000 accounts with paperwork errors. ERP programmes brought back in government and health. No programme I've recovered was lost for technical reasons. Not one. The technology was usually behind, but for reasons that had nothing to do with technology: truth that had quietly diverged, governance that had split, people who were exhausted, and a culture nobody had taken the time to read. The pattern is consistent enough to treat as a sequence.

Learn what beat the last team

The most dangerous assumption in recovery is that you're smarter than those who came before. Almost always the last team was smart. They started well. They had highly skilled people not too different to you and yours, and the programme beat them anyway. If you don't find out what actually beat them, it will beat you too, on roughly the same schedule.

Before changing anything, learn. Covey's fifth habit, seek first to understand and then to be understood, works on programmes as well as people. Sit with the outgoing team. Listen without judgment. Read the old decision logs for what was argued, not just what was decided. The diagnosis is rarely in the documents. It lives in what people stopped saying out loud. Respect what the last team got right, and say so publicly. You will need their knowledge, and some of them are still in the room.

Start with one truth

Even if it's ugly. Especially if it's ugly.

Troubled programmes drift into multiple truths, because each version has an owner and owners defend. The first move is a single honest baseline of scope, cost, time, risks and issues. One page if possible, believed by everyone, flattering to no one. Nothing else can start until the numbers mean the same thing in every room. An honest baseline brings something surprising: relief. Teams always know when the report is fiction. Being allowed to tell the truth again is the first act of recovery.

Unify the governance

Two programmes sharing one outcome cannot have two steering committees. I've watched parallel governance run over intertwined work: two sets of priorities, two definitions of done. The work can never be healthier than the governance above it. Merge them. One sponsor group, one decision path, one definition of success. Politically it is the hardest move on this list. It is not optional.

The work can never be healthier than the governance above it.

Stabilise the people before the plan

The instinct on a troubled programme is to replan immediately. Resist it for a week or two. People on a failing programme are exhausted and defensive. They have carried bad news with no safe place to put it, and usually been blamed at least once. A new schedule imposed on that state is just the old failure with new dates. The change models put desire before knowledge for a reason; ADKAR has that right. People rebuild plans; plans don't rebuild people.

People rebuild plans; plans don't rebuild people.

Read the culture, then work with it

Here is the part I've come to believe most strongly: it's always about the culture. Culture truly does eat strategy for breakfast, as Drucker is usually credited with saying, and every remediation I've done has had a cultural element underneath the technical one. A culture that prized heroics and quietly punished bad news. A culture so consensus-driven that no decision survived the meeting that made it. A culture of deep care for the front line that had learned to see head-office projects as the enemy.

Don't work against the culture. You will lose, slowly and expensively. Every culture has great aspects to work with, and like people, most cultures carry dark sides you have to grapple with honestly. The recovery move is to enlist the strengths. Point the pride at the programme's purpose. Make truth-telling safe inside a loyalty culture. Bring the front line's care into the design decisions. As a leader, culture is my job: not an HR abstraction beside the plan, but the medium the plan either grows in or dies in.

And whose recovery is it? As a programme manager, my job is to help my C-level sponsors achieve their goals. A recovery that makes the programme manager look good and the sponsor look rescued has failed politically, even when it succeeds technically. The sponsor carries the outcome to the board. My work is to make that walk safe.

Buy back confidence with visible wins

Recovery is a confidence business. Whatever the replanned end date says, something real must land inside the first thirty days: a fixed configuration, a cleared backlog, a report the CFO can finally trust. Small is fine. Visible is mandatory.

Then teach the team, and every stakeholder watching, how to win again. A string of well planned and deftly executed incremental wins after that first milestone is how team capability grows alongside stakeholder confidence. Some losses will come too, and that's alright. Be soft on the people and hard on the process, so you grow out of the losses as well as the wins. That is what all great teams do.

Be soft on the people and hard on the process, so you grow out of the losses as well as the wins.

The hard hat is optional

Strip it back and programme recovery is change management wearing a hard hat. The technical fix is rarely the hard part. Good engineers fix technology quickly once the conditions let them. The conditions are the job: the last attempt honestly understood, one truth, one governance, stabilised people, a culture enlisted rather than fought, a sponsor made safe, and confidence rebuilt one win at a time.

People fixed every programme I've ever helped bring back. My job was to make it safe for them to do it.