The Information Asymmetry That Keeps Complex Programmes Green Until They Fail
A programme report is never just a description of delivery; it is the product of negotiations over which version of delivery the organisation is prepared to hear.
The green report and the red corridor
At 7:20 on the morning of a steering committee, the programme report is complete. Overall status: green. Delivery is shown as 72 per cent complete, the principal supplier has met the month’s milestones, and the remaining contingency appears adequate. The pack has been checked by the programme office, approved by the programme director and circulated to the sponsors.
In a project room two floors below, the picture is different. Of eighteen interfaces described as “built”, only seven have processed representative volumes. The data-conversion team has found that one customer record in six cannot be matched without manual investigation. A supplier manager has already calculated that the current testing sequence requires eleven weeks more than the approved plan contains. None of these facts is wholly absent from the report. They have been translated: incomplete volume testing is “performance assurance in progress”; unmatched records are “data cleansing within tolerance”; the eleven-week exposure sits inside an amber workstream whose recovery plan is “being finalised”.
Every sentence is defensible. The conclusion is false.
This is the information asymmetry problem in complex programmes. Those closest to delivery possess the richest knowledge of actual progress, uncertainty and failure. Those who provide capital, authority and organisational protection must decide with a thinner and more managed account. Between them sit layers of reporting, commercial interest and professional self-preservation that progressively change what the information means.
The problem is rarely simple dishonesty. It is more durable than that. It arises because different actors know different things, carry different consequences and are rewarded for different outcomes.
Distance changes the meaning of information
A sponsor and a delivery team do not occupy the same programme.
The sponsor experiences commitments: a date announced to the board, a benefit promised in the business case, a budget competing with other investments, a regulatory or customer obligation that cannot easily move. The delivery team experiences mechanisms: an interface that fails under load, a design decision waiting for an owner, a supplier dependency omitted from the schedule, a process that works in one country but not another.
The sponsor cannot inspect all of this directly. Complex programmes now span functions, locations and corporate boundaries. Enterprise-system implementations, shared-service consolidations and outsourced operations may involve hundreds of staff, several suppliers and thousands of linked tasks. Reporting is therefore unavoidable. The organisation must compress reality before senior people can use it.
But compression is not neutral. Each layer decides what to omit, aggregate, defer or rename. A fact that is operationally alarming may look statistically small. A delay that is obvious in one workstream may disappear inside programme contingency. A risk without a precise financial value may lose to one that can be entered cleanly in a register.
The principal-agent difficulty begins here. The principal—the sponsor or governing body—delegates delivery to agents who possess more information about the work. The agents know whether the plan is credible, but they also know that admitting it is not may threaten funding, reputation, contractual position or employment. The principal wants candour, yet may punish the messenger, refuse the implication or insist that the original commitment remains unchanged.
Information is filtered not merely because people wish to conceal. It is filtered because truth has consequences, and those consequences are unevenly distributed.
A programme report is never just a description of delivery; it is the product of negotiations over which version of delivery the organisation is prepared to hear.
Three translations before the truth reaches the room
Bad news often changes character three times before it reaches a steering committee.
At the first translation, a technical or operational fact becomes a workstream issue. “The conversion routine corrupts historical tax codes” becomes “additional mapping required”. The new phrase is not wrong, but it removes the failure and foregrounds the remedy.
At the second, the workstream issue becomes a programme risk. Its direct effect is combined with probability, contingency and a recovery action. “Additional mapping required” becomes “medium likelihood of delay, mitigated through parallel working”. The fact has moved from present to possible.
At the third, the programme risk becomes an executive message. Several risks are balanced against progress elsewhere, and the status is chosen according to escalation rules. “Medium likelihood of delay” becomes “overall delivery remains on track”. The sponsor receives assurance precisely because the original fact has been processed by the programme’s controls.
This sequence explains how formal governance can produce less truth rather than more. Each person improves the information for the needs of the next level. Detail is removed, ambiguity is resolved and an action is attached. By the time the report is concise enough for decision, the uncertainty that required a decision may have disappeared.
| What delivery knows | Reporting translation | What the sponsor may infer |
|---|---|---|
| Seven of eighteen interfaces have passed representative-volume tests | Interface build substantially complete; performance testing under way | The remaining work is routine verification |
| One customer record in six requires manual matching | Data cleansing progressing within agreed tolerance | Data quality is controlled and resourced |
| The current sequence needs eleven additional weeks | Recovery options being evaluated within contingency | The approved date remains achievable |
| Two suppliers dispute ownership of a defect | Commercial clarification in progress | Accountability is clear and resolution is imminent |
The table shows that the gap is not between data and no data. It is between evidence and inference. The governing body sees statements that are individually accurate but collectively invite a confidence the evidence does not support.
A programme in three versions
Consider a composite programme replacing order, warehouse and finance systems across fourteen operating units. The approved case is £68 million over thirty months, with savings from retiring local systems and consolidating support. A systems supplier owns configuration, an internal team owns process design, and a second supplier provides testing.
By month eighteen, three versions of the programme coexist.
The integrated schedule shows 61 per cent complete. This number is based on tasks finished, not outcomes proven. Configuration workshops and document approvals carry substantial weight; end-to-end tests carry relatively little because they occur later. Completing another design document moves the percentage. Discovering that the design fails a real transaction does not reverse it.
The commercial report shows the principal supplier broadly on budget. Yet £3.4 million of disputed change requests sits outside the forecast while responsibility is negotiated. The supplier has little incentive to include costs it argues are caused by client delay; the client team has little incentive to accept costs that would consume most of the contingency. The forecast is therefore accurate only within a boundary that neither party has agreed.
The workstream reports show data migration as amber. The migration team has profiled 2.7 million customer records and found 430,000 that breach at least one new validation rule. Its manager estimates twelve people will need sixteen weeks to investigate the highest-risk cases. The programme plan contains six people for eight weeks. The gap is not hidden: it appears in a resource request awaiting approval. But the steering pack treats approval as part of the recovery plan and therefore reports the risk as managed.
The sponsor sees 61 per cent complete, budget broadly stable and an amber data issue with mitigation. The delivery teams see a schedule whose percentage rewards early paperwork, a forecast excluding disputed exposure and a recovery plan requiring four times the funded effort.
At the next review, the sponsor asks whether the date is still achievable. The programme director answers yes—with prompt decisions, supplier cooperation and no further material data defects. The conditional sentence is recorded simply as yes.
No individual has necessarily lied. Every actor has protected a legitimate interest:
- the supplier protects its contractual entitlement;
- the workstream manager protects scarce staff and avoids being labelled obstructive;
- the programme director protects confidence while seeking time to recover;
- the sponsor protects a commitment already made beyond the programme.
The aggregate effect is organised optimism.
The strongest case for compression
It is easy to conclude that sponsors should demand raw detail, speak directly to every team and distrust all summaries. That would replace one failure with another.
Senior governance cannot operate at transaction level. A steering committee that spends three hours debating defect classifications may miss the decision only it can make. Unfiltered information creates its own asymmetry: technical specialists can overwhelm non-specialists with volume, obscure weak performance behind complexity, or elevate local concerns above the programme’s wider purpose. Direct access can also undermine accountable managers if every disagreement becomes an appeal to the sponsor.
The case for disciplined reporting is therefore strong. Complex programmes need common definitions, thresholds, integrated plans and concise decision papers. Agents are employed partly to interpret detail. Delegation without trust becomes paralysis.
The answer is not to abolish compression. It is to preserve the lineage between the compressed message and the evidence beneath it.
A useful report should allow a sponsor to ask: What observation produced this status? Which assumption converts that observation into a forecast? Who disagrees with the interpretation? What decision would change the exposure? These questions do not demand every detail. They reveal where judgement has entered the chain.
The difference is subtle but important. Weak governance asks whether the report follows the standard. Strong governance asks whether the standard has preserved the meaning.
Incentives speak louder than escalation rules
Many programmes respond to information failure by adding controls: a new reporting template, a tighter issue threshold, another assurance review. Controls matter, but they cannot overcome incentives that reward concealment.
If the first person to declare red must also produce an immediate recovery plan, teams will delay red until a plan exists. If a supplier’s performance payment depends on milestone acceptance, completion will be defined at the most favourable boundary. If sponsors react to early warnings by demanding certainty, estimates will become promises. If programme directors are praised for calm rather than accuracy, turbulence will be managed rhetorically before it is managed operationally.
I have come to treat the time between private knowledge and formal escalation as one of the most revealing programme measures. In one composite case, a testing team recorded a critical capacity constraint forty-three days before it appeared in the steering report. During those forty-three days, the issue passed through six meetings, acquired two recovery plans and lost nine weeks of potential supplier lead time. The final escalation was professionally written and procedurally correct. It was also too late to protect the date.
The mechanism was not a missing rule. The team knew the escalation route. It delayed because each level believed the next meeting might produce a solution and because no one wished to escalate a problem without one.
Candour therefore requires more than permission to speak. It requires a governance culture that separates early disclosure from failure of ownership. The person who surfaces uncertainty should still help resolve it, but should not be made responsible for the existence of every fact disclosed.
The sponsor is inside the problem
Principal-agent language can tempt sponsors to imagine that asymmetry is something done to them by delivery teams. In practice, principals help create it.
A sponsor who asks, “Are we still on track?” signals that confirmation is the desired answer. A sponsor who asks, “What would have to be true for this date to hold, and which of those conditions have we not proved?” invites evidence. A committee that receives a red status calmly but reacts angrily to the consequence teaches teams that colour is tolerated and truth is not.
Sponsors also hold information that agents lack. They may know that funding beyond the year is doubtful, that an acquisition is under consideration, or that a public commitment cannot move. When this context is withheld, delivery teams optimise against an incomplete purpose. Information asymmetry runs both ways, even if authority does not.
The governing relationship becomes healthier when each side makes its constraints visible:
- delivery exposes uncertainty, dissent, disputed cost and conditions behind forecasts;
- sponsors expose decision boundaries, non-negotiable commitments and the real tolerance for cost, time and scope;
- suppliers expose the commercial assumptions on which estimates and milestones depend;
- assurance functions expose where evidence is absent rather than merely where procedure is incomplete.
This is not radical transparency. Some information must remain restricted. It is purposeful transparency: sharing what another party needs to make a truthful commitment.
Truth early enough to act
Complex programmes will never eliminate information asymmetry. Expertise will remain distributed, interests will differ, and senior decisions will always depend on compressed accounts. The ambition of governance is not perfect visibility. It is to make distortion difficult and correction early.
That requires a different view of reporting. The report is not the truth delivered upward; it is a proposition about the truth, supported by evidence and open to challenge. Green is not a fact. Completion percentages are not facts unless the unit of completion has operational meaning. A recovery plan is not mitigation until authority, money and time have been committed to it.
The recurring pattern is that programmes fail visibly only after they have failed informationally. Long before a date is missed or a budget exceeded, the organisation loses the ability to distinguish an achieved outcome from a completed task, an exposure from a forecast boundary, and a proposed remedy from a funded decision.
By then, the red status is merely the last translation.
The real discipline is to protect the uncomfortable interval when evidence has emerged but its implications are not yet settled. That is when governing bodies are most tempted to demand reassurance and delivery teams are most tempted to manufacture it. It is also when choice still exists.
A programme is not well governed because its leaders receive concise information. It is well governed when concision has not cost them the truth they alone have the authority to act upon.