A program can remain inside its cost and schedule tolerances while moving into a strategically unacceptable position.
Many governance systems are built around variance.
Project managers can operate within an approved tolerance. A schedule movement beyond a defined percentage is escalated. A cost overrun above another threshold goes to the sponsor or Board. The logic is sensible: executives should not be involved in every delivery adjustment.
The weakness appears when numeric tolerances become the main definition of materiality.
Imagine a high-consequence transformation in which a project manager may accept schedule movement within a five per cent tolerance, while the same program faces risks involving safety, welfare, reputation, funding continuity, quality and public confidence.
A two per cent delay caused by evidence of serious harm may deserve immediate executive intervention.
A six per cent delay on a low-consequence activity may be recoverable within the program.
Variance tells leaders how far performance has moved from plan.
It does not tell them how much the movement matters.
The Strategic Context
Delegated authority is essential for speed.
Without thresholds, every decision travels upward and governance becomes a bottleneck. Project and program leaders need freedom to manage routine variation.
But thresholds must reflect the nature of the decision.
Cost and schedule are easy to quantify, which makes them attractive governance controls. Strategic consequences are harder. Reputation can deteriorate before it appears in a financial report. Customer trust can decline while the project remains on budget. A critical capability can be lost while schedule variance remains green. A quality problem can be irreversible even if its direct cost is small.
The organisation therefore needs two forms of escalation:
variance-based escalation for controlled delivery management; and
consequence-based escalation for strategic protection.
They are complementary, not competing.
What Leaders Commonly Misread
The first error is treating a percentage tolerance as a proxy for significance.
Five per cent is not inherently material. It becomes meaningful only in relation to context.
A five per cent budget variance on a minor work package and a five per cent variance on a constrained transformation do not create the same decision.
The second error is making thresholds symmetrical.
A ten-day delay can be harmless in one part of the critical path and catastrophic in another. A small quality deviation can be insignificant in a temporary internal process and unacceptable in a safety-critical interface.
The third error is escalating only after an outcome has occurred.
Good governance escalates deteriorating conditions when intervention can still preserve options.
The fourth error is assuming that red status equals decision.
An issue can be severe yet remain at project level if the response is clear and within delegated authority. Conversely, an apparently minor issue may require executive action because it changes strategic assumptions, crosses organisational boundaries or requires a trade-off the project manager is not authorised to make.
Related article: Portfolio Reporting Should Change Decisions, Not Produce More Data
Reframing the Issue
Escalation should answer a different question:
Does this development require authority, judgement or resources beyond the current level?
That can occur for several reasons:
- the potential consequence is severe;
- the decision is difficult to reverse;
- the issue threatens a non-negotiable outcome;
- the strategic thesis may no longer hold;
- several programs or functions are affected;
- additional capital or capacity is required;
- stakeholder legitimacy is at risk;
- external commitments must change;
- the accountable leader needs to accept a new trade-off.
The percentage variance may still matter, but it becomes one indicator among several.
Four Dimensions of Materiality
A stronger escalation model considers at least four dimensions.
1. Magnitude
How large is the deviation from approved cost, time, scope, benefit or performance?
This is the traditional variance view.
2. Consequence
What happens if the issue is not resolved?
Consider safety, customer, financial, regulatory, reputational, workforce, operational and strategic consequences.
3. Reversibility
Can the decision be corrected later at reasonable cost?
Irreversible or expensive-to-reverse decisions deserve earlier escalation.
4. Authority boundary
Does the response require a decision outside the team's delegated mandate?
Even a small issue should escalate when resolving it requires changing strategic scope, accepting a lower benefit, renegotiating an external commitment or moving resources between portfolios.
Together these dimensions create a more decision-relevant definition of materiality.
Decision Framework
Use a Dual-Threshold Escalation Model.
Layer 1: Delivery tolerances
Define normal delegated tolerances for:
- cost;
- schedule;
- scope;
- quality;
- resource usage.
These protect management speed.
Layer 2: Consequence triggers
Define immediate escalation triggers independent of numeric variance.
Examples include:
- credible risk of serious safety or welfare harm;
- potential breach of law, regulation or ethical obligation;
- material damage to stakeholder trust or reputation;
- failure of an assumption essential to the business case;
- reduction of a strategic benefit below its minimum threshold;
- loss of critical organisational capability;
- need to change an external commitment;
- an irreversible decision with material future consequences;
- conflict between major objectives that requires executive trade-off.
A simple matrix helps:
| Condition | Within delivery tolerance | Outside delivery tolerance |
|---|---|---|
| Low consequence, reversible | Manage locally | Escalate for control |
| High consequence, reversible | Escalate for judgement | Escalate urgently |
| Low consequence, irreversible | Escalate before commitment | Escalate before commitment |
| High consequence, irreversible | Immediate executive escalation | Immediate executive escalation |
This model prevents a "green" dashboard from hiding a strategic problem.
Related article: When Transformation Objectives Conflict, Decide What Cannot Be Traded
From Strategy to Execution
Immediate action
Review current escalation rules.
For every threshold, ask what type of decision it is designed to trigger. If the only rules concern budget and schedule, governance is incomplete.
Add consequence triggers to the program charter and risk framework.
Then clarify who receives each type of escalation. Not every issue belongs with the Board. Some require the sponsor, operational executive, risk owner, regulator-facing executive or portfolio committee.
Related article: Program Governance Begins With Decision Rights, Not Committees
Medium-term capability building
Train teams to escalate decisions, not just problems.
An effective escalation should state:
- what has changed;
- why it matters;
- what happens if no action is taken;
- which options exist;
- what trade-offs each option creates;
- what authority is required;
- when the decision must be made.
This improves both speed and quality.
Governance forums should also distinguish forecast from actual variance. Waiting for an actual threshold breach can be too late when the trend is clear and the response lead time is long.
Long-term strategic positioning
Portfolio leaders should analyse escalation patterns across programs.
If the same issues repeatedly reach executives late, delegation rules may be wrong. If senior forums are flooded with minor variances, authority may be too centralised. If major consequence events remain buried at project level, the organisation's materiality model is too financially narrow.
These patterns are evidence about governance design itself.
Related article: Governance Can Reduce Complexity or Become Another Layer of It
Signals to Monitor
Warning signs include:
- green cost and schedule status alongside deteriorating customer or risk indicators;
- repeated surprises at Board level;
- teams delaying escalation because tolerance has not technically been exceeded;
- different functions applying inconsistent meanings of "material";
- significant irreversible commitments being made within routine delegated authority;
- escalation papers that describe the problem but not the decision required;
- sponsors routinely learning about issues only through monthly reports;
- high-consequence risks being assigned the same cadence as ordinary project variance.
A strong escalation system creates earlier visibility without pulling every operational decision upward.
Questions for the Leadership Team
- Which consequences require escalation regardless of cost or schedule variance?
- Which decisions are difficult to reverse and therefore deserve earlier scrutiny?
- Do our delegated tolerances match the risk appetite of the enterprise?
- Are teams escalating when they need authority, or only when a metric turns red?
- Which strategic assumptions would trigger immediate reconsideration if contradicted?
- Are our governance forums receiving problems, or decision-ready choices?
- What do recurring escalation patterns tell us about whether authority is placed at the right level?
Closing Perspective
Numeric tolerances are useful because they create clarity.
They become dangerous when clarity is mistaken for completeness.
A transformation does not become strategically safe because cost and schedule remain within five per cent of plan. Materiality is shaped by consequence, reversibility, strategic value and authority as much as by variance.
The purpose of escalation is not to punish deviation.
It is to place consequential decisions with the people who are accountable for their effects while there is still time to choose.
Good governance therefore asks more than, "How far are we from plan?"
It asks, "What has become important enough that a different level of leadership must now decide?"
About EraNorth Insights
EraNorth Insights publishes practical analysis on strategy, projects, operations, transformation and decision intelligence for professional and organisational use. About EraNorth.