Most organisations do not reset their programmes. They relabel them. A new name, a revised timeline, a fresh steering pack. The same scope, the same vendor, the same unresolved issues. The work resumes. The problems follow. Six months later the conversation repeats.

Real resets are rare. They require the organisation to acknowledge that the foundation is broken, to stop building on it, and to rebuild from a position where the next milestone is actually achievable. That is uncomfortable. It is also the only path that works.

This article is for the executive, sponsor, or PMO lead who has inherited a programme in trouble, been told it has been reset, and suspects it has not been. The distinction matters. A true reset protects the remaining investment. A relabelled one accelerates its loss.

Why organisations avoid the real reset

Two forces are at work. The first is sunk cost. The second is optics.

Sunk cost is visible. The licences are purchased. The SI has been paid. The internal team is committed. The business has communicated benefits to staff. A genuine reset means admitting that some of that spend will not produce the outcome it was meant to. That admission has to travel upward to a board that approved the original case. Very few programme directors volunteer for that conversation.

Optics follow. A reset signals failure. A replan signals continuation. Organisations prefer the language that protects reputations, even when the structure it describes is less truthful. We have sat in rooms where a programme director described re-baselining what was, in every observable sense, a reset by any other name. The preference for softer language obscures the scale of the change required, and that obfuscation eats into the recovery.

What a relabelled reset looks like

Certain signals reveal a reset that is not really a reset.

The new plan inherits the old assumptions. Go-live has moved, but the data migration estimate is unchanged despite known quality issues. Testing windows are shorter than the last time they were compressed. The integration complexity still assumes the original architecture. None of this reflects learning. It reflects a schedule adjusted to fit the desired outcome, not the observable facts.

The governance structure is unchanged. The same sponsor, the same steering committee, the same meeting cadence that was not catching issues before. Governance that failed to diagnose the original problem is unlikely to diagnose the next one.

The scope has not moved. Every requirement from the original business case is still present in the reset. The ambition is preserved. The timeline is expanded. The effort is constant. That is a replan. A reset would reduce the ambition to match the remaining capacity.

The vendor relationship has been smoothed over, not restructured. There has been a difficult conversation, perhaps an executive-level meeting, and the SI has committed to improvements. But the commercial terms have not changed, the escalation path has not changed, and the accountability for specific deliverables has not been written down. Goodwill is not a reset.

Minimum viable first

The discipline underneath a real reset is simple to describe and difficult to accept: get the core live, then build.

This is not a new idea. It is the principle that drives every successful transformation we have seen across our engagements. Organisations succeed when they strip their programme back to the smallest viable outcome, deliver that cleanly, and expand from there. They fail when they insist on delivering the original ambition despite clear evidence that the foundation cannot bear the weight.

We worked on a programme in health and aged care that had run for three years across two failed phases. Phase one had gone live with known defects. Phase two tried to add scope onto that broken foundation. When we were asked to review it, the conversation with the executive team was short. The foundation was not repairable under the current plan. The choice was not between recovery and cancellation. It was between a real reset that stripped back to a minimum viable first release, and a relabelled one that would produce a third failure in due course.

The organisation chose the real reset. It was painful. Scope was reduced by around forty per cent. The vendor contract was renegotiated. The go-live was moved. The board had to be told the truth. But eleven months later the core was live, stable, and being used. The descoped functionality was reintroduced in a controlled second phase. That programme is now considered a success. It would not have been if the reset had been cosmetic.

The four tests for a real reset

Four questions separate a reset from a relabel.

First, has the scope moved. If the original scope is still intact, the programme has not learnt anything from its own history. The reset needs to name what is coming out, not just what is coming later.

Second, has the plan been rebuilt from verified facts, not inherited assumptions. Every milestone in the new plan should have been independently validated. Nothing should survive from the old plan without being re-estimated against actual progress, actual data quality, and actual team capacity.

Third, has governance been restructured. The sponsor model, the steering membership, the escalation path, and the reporting discipline should all have been examined. Some should have changed. Governance that did not detect the original failure is unlikely to protect the recovery unless it has been rebuilt with that purpose in mind.

Fourth, has the vendor relationship been reset at the commercial and behavioural level. Clear deliverables. Clear payment tied to outcomes. Clear escalation. Clear consequences. Without structural change, the relationship will revert to the pattern that produced the original failure.

If the honest answer to any of these four questions is no, the programme has been relabelled, not reset. That is not a minor observation. It is the difference between a recoverable investment and a continuing one.

Compounding cost

The financial argument for a real reset is blunt. The cost of building on a failed foundation is not linear. It compounds.

Every additional phase built onto a flawed base inherits the flaws. Defects multiply. Rework expands. Integration complexity rises. The cost of the eventual clean-up is higher than the cost of the reset would have been. We have seen programmes absorb four to six times the original budget because the organisation refused to reset early.

A programme we reviewed had a programme director managing structural misalignment between business and delivery teams for eighteen months. Each phase added scope, each phase broke, each phase was relabelled. When a forced reset eventually occurred, the cost of the required rework equalled the entire original budget. The reset could have been done at month six at a fraction of that cost. It was not, because the organisation preferred the optics of progress to the reality of a reset.

The point is not that every programme should reset at the first sign of trouble. It is that the cost of delay is systematically underestimated. A reset looks expensive. Not resetting is usually more expensive. The only way to know which is which is to examine the programme honestly, with someone independent in the room.

When to reset, when to recover, when to stop

Not every troubled programme needs a full reset. Some are structurally sound and genuinely recoverable. Some are beyond repair and should be stopped. The judgement depends on three things.

The first is whether the original business case is still valid. If the benefits that justified the programme are still achievable, recovery or reset is worth considering. If the market has moved, the regulatory landscape has shifted, or the organisational strategy has changed, stopping may be the more responsible option.

The second is whether the team and vendor can deliver the reset. Capability gaps that caused the original failure do not resolve themselves. If the SI has lost the people who knew the configuration, or the client team has lost trust internally, the reset has to address that before replanning.

The third is whether the organisation has the absorptive capacity for the revised change. Reset programmes often introduce smaller scope but tighter discipline. The business has to be ready for a different style of delivery. If the fatigue is already deep, the reset may need to be paired with a pause before execution begins.

These judgements are hard to make from inside the programme. They are routine for an independent reviewer. That is the case for involving one before the next relabelled reset consumes another twelve months of budget.

The quiet version of the conversation

Boards and executives are often willing to have the reset conversation when it is framed properly. They are not willing to have it when it is framed as failure. The distinction is subtle and important.

A reset conversation that begins "we got this wrong and need to start again" is rarely well received. A reset conversation that begins "here is what we now know, here is what the programme cannot deliver on the current basis, and here is what a credible path forward looks like" is almost always well received. The first is apology. The second is leadership.

The best programme directors and sponsors we have worked with learn to run that second conversation early. Not because they enjoy it. Because they understand that the cost of not running it falls on them as much as it falls on the organisation. A relabelled reset postpones the difficult conversation by a few months. A real reset closes it.

A programme reset is a governance decision, not a schedule adjustment. The tests are simple. The answers are uncomfortable. The cost of getting them wrong is paid twice.

A real reset strips scope, rebuilds the plan, restructures governance, and resets the vendor. A relabelled reset changes the cover sheet.

If the original problems are still present after the reset, it was not a reset. It was a delay. Independent review before the relabel is usually cheaper than recovery after it.