All projects come with risks—but scale them up, and the stakes rise fast. In fact, a 2023 survey by McKinsey found that large capital projects overrun budgets and schedules by 30-45%. Traditional contingency planning— adding a fixed percentage to the budget as a buffer—just doesn’t cut it for complex projects, especially not in a post-COVID world where costs are so volatile. So how can risk managers effectively analyze risk for projects that can shape industries, economies, and communities for decades.
Fortunately, there are smarter ways to approach capital project cost risk modeling. In the webinar, “Mastering Cost Risk: Smarter Models for Large-Scale Projects,” Dimitris Leimonis, Founder and Manager Director at Financial Arena, demonstrated a process for using @RISK to create cost risk models that evolve with your project—no matter how large.
Continue reading to gain the session highlights or watch the webinar on-demand.
Traditional cost risk models vs. probabilistic models
Traditional cost control methods treat uncertainty and risk as exceptions, not the rule. Typically, risk managers create a single-point estimate for their project with a fixed contingency applied. The model is hosted in a spreadsheet that must be adjusted manually to conduct sensitivity testing.
Most importantly, the estimate in that model is deterministic – it’s based on everything going to plan. When the project inevitably doesn’t go to plan, the result is overruns that lead to unpleasant surprises like delayed ROI, reduced opportunities, and erosion of stakeholder confidence.
The limitations of traditional cost control methods
Below are common cost-control methodologies and the reasons why they are unsuitable for large projects. These methods treat uncertainty as an exception rather than the rule:
- Bottom-up estimating: Detailed work breakdown with unit costs. This assumes perfect information and stable conditions.
- Historical benchmarking: Using past project costs as reference. This ignores the fact that every project has unique complexities and contexts.
- Fixed percentage contingency: Adding 10-20% buffer to base estimates. This method contains no connection to actual project risks.
- Three-point estimating: Optimistic, pessimistic, and most likely scenarios. This method still produces single point estimates.
- Earned value management: Tracking planned vs. actual progress. This is reactive rather than predictive.
The advantages of probabilistic cost models
Probabilistic models, on the other hand, simulate multiple scenarios that account for uncertainty. Sensitivity testing is carried out through automated Monte Carlo simulation, leading to outputs that offer ranges for costs based on the most likely outcomes of the simulation.
With probabilistic cost models, risk management teams can quantify how likely different outcomes are, leading to risk-informed contingency planning that reflects the complexity of the project at hand – not a fixed buffer percentage.
Case study: Building a light rail extension
To demonstrate how @RISK helps risk managers build more robust cost risk models for large projects, let’s look at a hypothetical light rail extension program project outline that costs $50 million over five years.
Project details:
- Cost: $50M base infrastructure investment
- Timeline: 5-year construction timeline
- Categories: 17 major cost categories (NRM/CBS structure)
- Risks: 73 individual risk factors
- Scenarios: 3 economic scenarios
Why this example:
- Represents typical large-scale complexity
- Shows real-world risk register integration
- Demonstrates stakeholder communication challenges
This large, complex assignment is typical of infrastructure projects. Keeping track of all aspects of the program is a crucial step in transforming cost risk models from static ones to dynamic ones. In the webinar, Dimitris showed one possible approach to assigning codes to different cost categories within the project.
Cost model coding system, bringing order to complexity.
Potential cost category code “buckets” and naming conventions for this project. Precise coding, like a good filing system, creates clarity out of chaos.
A well-defined coding structure can produce clearer reports, allowing more productive discussions with stakeholders. Codes also help with developing an integrated risk register for the project, which is another advantage of using @RISK risk management software.
Another example of coding inputs that demonstrate the potential granularity of cost risk models using @RISK.
However you choose to code your project, the goal is to transform a static, deterministic cost model into a dynamic, probabilistic model.
Building a framework to define scenarios and cost categories
The first step is to gather base estimates and build a framework for your cost categories. Many deterministic cost risk models use a three-scenario framework: one that covers the base estimate, plus an optimistic and pessimistic scenario in which conditions (for example, inflation or material costs) are better or worse than planned.
A potential issue with this framework is determining what “optimistic” and “pessimistic” conditions look like—for example, teams may argue about which inflation rates to use.
In this example, the three-scenario economic framework looks as follows:
Optimistic:
- Inflation: 2.0% -> 1.0% -> 0.5%
- Risk level: Low
- Use case: Best-case planning
Base:
- Inflation: 3.0% -> 2.0% -> 1.5%
- Risk level: Medium
- Use case: Most likely outcome
Pessimistic:
- Inflation: 4.0% -> 3.0% -> 2.5%
- Risk level: High
- Use case: Stress testing
With @RISK, cost risk managers can model all scenarios and determine how likely each one is. The result is a model that replaces single-point estimates with distributions, includes time-series elements, and accounts for correlations between risks (more on correlations shortly).
Example of a Pert distribution in @RISK showing the main contractor’s preliminaries ranges, determined by estimators.
Integrating the risk register into the cost model
The next step is to develop a qualitative risk register that can be integrated with the quantitative cost model. Traditionally, qualitative assessments take place in workshops that are separate from the development of the quantitative cost model.
This is where @RISK shines. Risk managers can connect inputs from the quantitative @RISK model with the qualitative risk register, unifying them into one central resource. Finally, building a granular risk register makes it possible to analyze specific cost elements or cost scenarios by “switching” them on and off within Excel before running the simulation.
Excel spreadsheet for the sample project showing “switches” that allow for risk modeling based on specific project elements.
This level of control allows cost risk managers to model the impact of specific sets of circumstances and then determine how different decisions could mitigate or exacerbate risk.
Using a carefully designed dynamic model gives managers opportunities to understand potential outcomes better.
Defining correlations between risks
When one thing goes wrong, other things are likely to follow as certain risks are often linked to one another. For example, commodities market fluctuations may impact the cost of copper wiring. Multiple requests for design changes may lead to higher fees from the professionals involved.
Correlations like this can be difficult to model deterministically. Calculating the independent cost of each possible risk and adding them up – a common deterministic approach – often leads to an underestimation of the impact.
Example of independent risks vs. correlated risks
Independent risks (wrong)
- Weather delay: +$2M
- Labor shortage: +$1.5M
- Total impact: +3.5M
Correlated risks (reality)
- Weather delay causes labor shortage
- Combine impact: +$5.2M (48% higher)
Correlations can act as force multipliers on one another – a probabilistic model is more likely to reflect this accurately.
With @RISK, cost risk managers can build a model that helps them understand the potential impact of a cascade of contingencies.
Give the right information to the right stakeholders
All large projects have three main types of stakeholders: executives, finance team members, and project team members. With @RISK’s built-in reporting capabilities, cost risk managers can effectively communicate with all three groups.
Below are three stakeholder audience types, with examples of information they may need from cost risk models:
Executives: ‘What’s the bottom line?’
- P50 cost: $58M, P90 cost: $67M
- 35% chance of exceeding $60M
- Key driver: Permitting delays
Finance team: ‘Show me the details’
- Full tornado chart (25 risk factors)
- Monthly cash flow distributions
- Sensitivity analysis tables
Project team: ‘What should we focus on?’
- Top 5 risk mitigation priorities
- Correlation impacts
- Scenario planning insights
Quickly produce cost probability curves or ranked lists of key risk drivers for executives. Give the finance team monthly cash flow profiles that illustrate seasonal risk patterns. Help project managers understand which aspects of the project most impact success with granular sensitivity analyses that can rank risk factors:
Image caption and alt. Text: @RISK tornado chart showing the impact of specific risk items on project cost, ranked by highest contribution to risk.
Keep everyone informed with the latest information, communicate in formats that help them make better decisions, and enable everyone to perform better in their roles.
Transform your budgets and investment decisions
Shifting to probabilistic cost-risk analysis is the smarter way to plan for large-scale projects. With the right preparation of your data—and with simulation and analysis tools provided by @RISK—it’s possible to break out of the estimate-and-contingency model of the past to create models that are more dynamic, accurate, and useful.
To take your risk management end-to-end, combine @RISK with Predict! to track, consolidate, and compare risk management efforts at scale. Together, they provide a seamless connection between detailed project-level risk analysis and broad, portfolio-wide risk visibility—giving you clear insight across all your risk registers and enabling smarter, more confident capital project decisions.
