Single-Point Schedules Hide the Risk Leaders Need to See

A completion date without a confidence level is not a forecast; it is one point selected from a range of possible futures.

Senior leaders want a date. Customers need commitments, funding decisions require timing and dependent initiatives must plan. The organisational pressure for one number is real.

The problem begins when a single date is treated as if uncertainty has disappeared. Activity durations are estimates. Dependencies may change. Resources may not be available. Risks can affect several paths at once. The project can calculate one finish date while the underlying range remains wide.

False precision does not reduce uncertainty. It transfers uncertainty into later surprise.

The Strategic Context

Deterministic schedules assign one duration to each activity. This is appropriate for building logic and managing day-to-day work, but it does not describe the range of possible completion outcomes.

Probabilistic methods use ranges or distributions for uncertain activities and model how uncertainty combines through the network. Three-point estimating and PERT are accessible entry points. More comprehensive schedule risk analysis uses repeated simulation to show the distribution of possible finish dates and identify the risks or activities driving that distribution.

The US GAO Schedule Assessment Guide identifies schedule risk analysis as a characteristic of a reliable schedule. It is intended to estimate confidence in completion dates and the contingency required—not merely produce another sophisticated chart.

What Leaders Commonly Misread

The first misreading is that the most likely duration for each activity produces the most likely project finish. It may not. Uncertainty accumulates, parallel paths can become critical and delays are not always offset by early finishes elsewhere.

The second is that contingency should be hidden inside every activity. Local padding is difficult to see and tends to be consumed without regard to overall need. It also makes it hard to distinguish a realistic duration from a protective allowance.

The third is that an 80 per cent confidence date means the project will be on time 80 per cent of the time. The result is conditional on the model, data, correlations and risks represented. Model confidence is not certainty about reality.

The fourth is that statistical analysis repairs a weak schedule. Simulation of incomplete scope, poor logic or unrealistic resources creates a quantified version of the same weakness.

Reframing the Issue

Schedule uncertainty should be governed as a range of outcomes connected to decisions.

Leaders do not need to become statisticians. They need to understand:

• the date produced by the deterministic schedule;
• the confidence associated with alternative commitment dates;
• the principal drivers of variation;
• the contingency embedded or held separately;
• the events that could move the distribution; and
• the consequence of finishing later or earlier.

The right date depends on the decision. An internal planning target may use an aggressive confidence level to drive action. A contractual promise, public launch or operational shutdown may require substantially greater protection. Confusing the two invites either complacency or avoidable failure.

Related article: The Critical Path Is a Forecast, Not a Promise

Three-Point Estimates Are a Conversation, Not an Answer

A three-point estimate asks for an optimistic, most likely and pessimistic duration. The value is not only the weighted calculation. It forces estimators to explain the conditions behind the range.

For each activity, leaders should ask:

• What must occur for the optimistic case?
• Which conditions define the most likely case?
• What credible events produce the pessimistic case?
• Are the bounds based on evidence or negotiation?
• Does the same risk affect several activities?

Traditional PERT uses a weighted mean and a simplified measure of variability. These formulas are useful for teaching and rough analysis, but their assumptions are restrictive. Real project durations may be skewed, estimates may be correlated and the critical path may switch between iterations.

For material commitments, the organisation should avoid treating the basic PERT formula as a complete risk model.

Correlation Changes the Result

Activities are often treated as independent because the mathematics is easier. In practice, common drivers create correlation.

A delayed design decision may affect procurement, testing and regulatory submission. Low workforce productivity may influence several work packages. Severe weather can delay multiple site activities. A supplier failure can affect an entire subsystem.

If correlated risks are modelled as independent, apparent diversification may be overstated. The simulation can suggest that early and late outcomes cancel when the same event moves many activities together.

Risk analysis should therefore connect the risk register to schedule activities and model common-cause effects where they are material.

Critical-Path Switching Matters

Deterministic CPM identifies the path controlling completion under one set of durations. Under uncertainty, another path can become longer.

This is particularly important where several paths have similar length or converge at a major integration milestone. Focusing only on today’s critical path can leave a near-critical path unmanaged until it becomes the driver.

Schedule risk analysis should reveal criticality sensitivity: how often an activity or path becomes critical across simulations. An activity with float in the deterministic plan may still deserve attention if it becomes critical in many plausible scenarios.

Contingency Is an Enterprise Choice

Time contingency is not free. A later commitment may delay revenue, extend temporary operations or reduce market advantage. Too little contingency increases the probability of failure; too much can weaken urgency and make the investment unattractive.

Leaders should decide where contingency sits and who controls it. Options include:

• project-level buffer protecting final completion;
• feeding buffers protecting critical work from converging paths;
• management reserve for identified uncertainty;
• contractual allowance; and
• portfolio contingency for shared external events.

The design should prevent local teams from consuming protection without understanding the consequence for the whole project.

Decision Framework

A probabilistic result should change a decision. If leadership receives a probability curve but continues to manage only the original date, the analysis has become decorative.

From Strategy to Execution

Immediate action: Ask schedule owners to separate activity estimates from contingency and identify the assumptions behind the most important ranges.

Medium-term capability: Introduce schedule risk analysis for projects with material uncertainty, irreversible commitments or high consequences of delay. Develop historical duration data and calibrate estimates against actual performance.

Long-term positioning: Aggregate schedule exposure across the portfolio. Shared suppliers, approval bodies, facilities and specialists can create correlated delay that individual project models do not see.

Related article: The Hidden Portfolio Cost of Multitasking

Signals to Monitor

• Every activity has a precise duration unsupported by evidence.
• The external commitment equals the deterministic earliest finish.
• Pessimistic estimates are rejected because they are “too negative”.
• Contingency is hidden across activities and cannot be reconciled.
• Risk analysis assumes independence despite common drivers.
• Only the current critical path receives risk attention.
• Confidence improves without a change to scope, logic, resources or risk.

Questions for the Leadership Team

1. What confidence level supports the date we are communicating?
2. Which assumption contributes most to the width of the forecast range?
3. What risks affect several activities or projects simultaneously?
4. Which near-critical path most often controls completion under uncertainty?
5. Where is schedule contingency held, and who can consume it?
6. What is the commercial consequence of choosing a safer commitment date?

Closing Perspective

Leaders will always need dates. Responsible forecasting does not avoid commitment; it makes the basis and risk of commitment visible.

The objective is not statistical sophistication for its own sake. It is to choose targets, promises and contingency with an honest understanding of uncertainty—before the organisation loses the ability to respond.

About the author
Kevin Jogin is Founder & Principal Advisor at EraNorth. Meet the Founder.

Copy link Print