The most important information in an estimate is often not the amount, but the evidence, assumptions and uncertainty that produced it.
An executive asks what a proposed initiative will cost. The answer arrives as a precise number, presented in a business case and repeated through portfolio reporting. Precision creates comfort. It also creates a dangerous illusion: that uncertainty has been resolved because a spreadsheet contains two decimal places.
The estimate may have been prepared before scope, design, supplier strategy or delivery conditions were mature. Its apparent certainty does not reflect the quality of the underlying decision. When the final cost changes, leaders call it an overrun even though the original number was never capable of supporting the commitment made against it.
The Strategic Context
Cost estimates influence whether an initiative enters the portfolio, how it competes for capital and whether leaders believe the organisation can afford its strategy. They also affect pricing, funding, contracts, capacity and stakeholder expectations.
The estimate is therefore not merely a project-control input. It is a statement about the probable resource demand of a strategic choice under stated conditions.
As an initiative develops, the evidence changes. Early estimates may rely on analogous experience, broad parameters and assumptions. Later estimates can use defined work packages, supplier quotations and detailed resource plans. These estimates should not be treated as interchangeable. Each is suitable for a different decision.
The critical governance question is: does the maturity of the estimate match the irreversibility of the commitment?
What Leaders Commonly Misread
The first error is assuming that a single number is more rigorous than a range. A narrow number can conceal wide uncertainty. A range, supported by clear drivers and confidence, may be more decision-useful.
The second is confusing an estimate with a budget. An estimate forecasts expected resource demand. A budget is an authorised allocation and control baseline. They are connected, but one describes expectation while the other represents a governance decision.
The third is comparing estimates without normalising their basis. One may include internal labour, escalation and contingency; another may include only external expenditure. Apparent cost advantage can be a classification difference.
The fourth is treating contingency as evidence of poor planning. Some uncertainty cannot be eliminated economically before a decision. Contingency recognises exposure; hidden optimism does not remove it.
Reframing the Issue
Cost estimating should be reframed as evidence-based confidence management.
The purpose is not to predict one exact future. It is to help leaders understand the likely cost, the range of credible outcomes, the conditions that could move the result and the actions available before or after commitment.
This shifts attention from defending the number to improving the decision. Leaders can choose to gather more evidence, reduce scope, change delivery strategy, stage the commitment, transfer selected exposure or accept uncertainty because time-to-value matters more.
An estimate becomes decision-grade when it explains what is known, what is assumed and what remains sensitive.
Estimate Maturity Must Match Decision Maturity
Different estimating approaches serve different stages.
Expert and analogous estimates are useful when little detail exists, particularly for comparing broad alternatives. Parametric approaches can provide discipline where reliable relationships and relevant data exist. Bottom-up estimates become valuable when work is sufficiently defined. Supplier quotations test market conditions but can still omit client costs, interfaces and change exposure. Three-point estimates expose uncertainty around individual elements, although their value depends on realistic inputs.
No method eliminates judgement. The question is whether the method is appropriate to the evidence and decision.
A reversible discovery experiment does not need the same estimating detail as a fixed-price production contract. Conversely, a major irreversible commitment should not proceed on an early analogy simply because executive timing demands an answer.
Related article: Single-Point Schedules Hide the Risk Leaders Need to See
The Cost Boundary Must Be Explicit
Every estimate needs a defined boundary. It should distinguish, where relevant:
- Direct and indirect costs.
- Internal labour and external expenditure.
- Fixed and variable costs.
- Capital and operating consequences.
- Base estimate and contingency.
- Project cost and wider enterprise cost.
- Initial acquisition and through-life cost.
The widest source of surprise is often not poor arithmetic but omitted work. Data cleansing, operational downtime, transition support, spares, training, maintenance, cyber controls and decommissioning can sit outside the project estimate while remaining unavoidable consequences of the investment.
A low acquisition cost may therefore create a high enterprise cost. Portfolio leaders need the economic boundary that matches the strategic decision, not merely the boundary easiest for the project to report.
Assumptions Are Part of the Estimate
An estimate without an assumption register is incomplete. Assumptions may concern productivity, exchange rates, design stability, supplier availability, access, inflation, approvals or resource continuity.
Leaders should focus on assumptions that are both influential and weakly evidenced. Those are the points where targeted discovery creates the greatest decision value.
For example, a hypothetical manufacturer may estimate installation on the assumption that shutdown access is available for two uninterrupted weeks. If operations can provide only weekends, the cost and schedule model changes materially. The issue is not estimating technique; it is an unresolved interface between production and delivery.
Assumptions should have owners, evidence and review dates. When an assumption changes, the estimate should change visibly rather than preserve the original number for political convenience.
Decision Framework
Before relying on an estimate, leaders should review six dimensions.
| Dimension | Decision test |
|---|---|
| Purpose | What decision is this estimate intended to support? |
| Maturity | How developed are scope, design, schedule and commercial strategy? |
| Boundary | Which costs and lifecycle consequences are included or excluded? |
| Method | Is the estimating approach suitable for the available evidence? |
| Uncertainty | What range is credible, and what drives the extremes? |
| Governance | Who can revise the estimate, draw contingency or change the commitment? |
Leaders then have four broad pathways:
- Commit: evidence and confidence are sufficient for the decision.
- Stage: release limited funding while preserving later choices.
- Reduce uncertainty: fund design, trials, market testing or investigation.
- Decline or defer: expected value does not justify cost or uncertainty.
This framework recognises time as a strategic variable. Waiting for perfect information can destroy opportunity, but proceeding with false certainty can lock the organisation into an uneconomic path.
From Strategy to Execution
Immediately, require estimates to state purpose, maturity, boundary, assumptions and range alongside the central value. Executives should stop comparing numbers that were prepared on different bases.
Over the medium term, connect estimate development to decision gates. Confidence should increase as the organisation moves toward harder-to-reverse commitments. Estimate changes should be reconciled to scope, market, schedule, assumptions and identified risks.
Long-term capability requires an estimating evidence base. Actual costs, productivity, supplier outcomes and change causes should be captured in reusable form. The objective is not to punish variation; it is to improve future forecasts and identify systematic optimism.
Portfolio governance should also compare forecast uncertainty across initiatives. A portfolio made entirely of individually attractive but highly uncertain investments may exceed the enterprise's risk capacity.
Related article: The Work Breakdown Structure Is a Control Architecture
Signals to Monitor
Leaders should challenge estimates when:
- Precision increases without corresponding design or scope maturity.
- Contingency is removed to meet an affordability threshold.
- Internal labour or operational disruption is treated as free.
- The estimate remains unchanged while major assumptions move.
- Supplier quotations are accepted without client-side and interface costs.
- Early estimates become permanent performance commitments.
- Actual-cost learning is not improving future estimates.
Questions for the Leadership Team
- What decision is this estimate sufficiently mature to support?
- Which cost drivers are both material and weakly evidenced?
- What enterprise costs sit outside the project's reporting boundary?
- Which assumptions could be tested before we make an irreversible commitment?
- Are we using contingency to govern uncertainty or to conceal optimism?
- What opportunity are we sacrificing by funding the upper end of this estimate?
Closing Perspective
Executives do not need estimates that pretend to know the future. They need estimates that reveal the quality of current knowledge and support proportionate commitment. A credible estimate makes uncertainty visible, connects cost to scope and conditions, and changes as evidence improves. The number matters, but the confidence behind it determines whether capital is being allocated intelligently.
About EraNorth Insights
EraNorth Insights publishes practical analysis on strategy, projects, operations, transformation and decision intelligence for professional and organisational use. About EraNorth.