Uncertainty results from unknown factors, complexity of interrelations, conflicting perspectives, and tradeoffs between options. Analysis, which involves the ranking and selection of approaches from alternatives, requires understanding of the opportunity or problem, the nature of the decision, areas of uncertainty that affect decisions, and the consequences of possible outcomes.
The International Institute of Business Analysis’ (IIBA) Business Analysis Body of Knowledge (BABOK) describes decision analysis as an approach that “examines and models the possible consequences of different decisions… [and] assists in making an optimal decision under conditions of uncertainty.” Mathematical models are often used to help assess alternatives.
In this article, we explore two financial valuation techniques that fall under the category of decision analysis—discount cash flow (DCF) and real options analysis (ROA)—not from the perspective of investment in risky assets, but as project management/business analysis tools. In turn, we compare DCF and ROA to predictive and adaptive projects.
Traditional projects and discount cash flow
Plan driven projects such as waterfall engender reliance on upfront analysis and planning. To help evaluate alternatives, traditional valuation techniques such as discounted cash flow (DCF) are widely accepted as a means to make cost-benefit decisions. “If benefits outweigh the costs, do it!”
For waterfall projects the irony is that requirements and cost-benefit analysis is performed when the least is known about the best possible solution and how to implement the desired outcome. To counter-balance this risk, project managers institute a change control process to manage scope.
As is the case with predictive project methods, valuation techniques such as DCF are deterministic and assume at the outset that business conditions remain fixed. In effect, there is little room for flexibility in addition to the issue that projected outcomes are simply subjective best-guess forecasts.
Figure 1 — Example Discounted Cash Flow Analysis 
Figure 1 shows a simple example of applying discounted cash flow analysis. It assumes a project that costs $1,000 to implement at Year 0 will generate the following stream of positive cash flows for five years: $500, $600, $700, $800, and $900. The timeline shows these cash flows and their respective discounted present values. The model assumes a 20 percent risk-adjusted discount rate using a weighted average cost of capital (WACC), and a long term growth rate of 5 percent.
IRR is derived as 54.97% using Excel. NPV is calculated as $985.92. Utilizing the Gordon Growth Model, terminal value of the project’s cash flow based on the fifth year is $6,300. Discounting this figure for five years at the risk-adjusted discount rate and adding it to the NPV calculates an implied enterprise value (i.e., estimated fair market value) of $3,517.75.
The problem with this static discount cash flow analysis is that it produces a single-point estimated result. As a result, confidence in its accuracy suffers due to the reason that forecasts by their nature are uncertain.
No doubt there are advantages to using DCF—it is a consistent, relatively simple, widely utilized, and economically rational decision criteria. Nevertheless, as with predictive projects, DCF does not account for managerial flexibility over the life cycle of a project. Table 1 reveals many concerns similar to that which project managers face when managing plan driven projects.
Table 1 — Assumptions versus Realities of DCF 
Project considerations and ROI
In a plan driven project, there is a tendency toward excessive planning, locked requirements, and one shot at “big bang” implementation, which in turn leads to “all or nothing” consequences. DCF valuation may contribute to failed project expectations because such analysis is treated as certain.
The “cone of uncertainty” describes the evolution of the amount of uncertainty during a project. The further a project progresses, the more accurate estimates for the remaining effort and time becomes.
Figure 2 — Cone of Uncertainty 
Change driven projects, on the other hand, recognize that the business environment is in reality highly fluid. Adaptive methods such as scrum utilize an iterative and incremental development framework. At the same time, frequent inspection and adaptation produces an empirical control process. Indeed, backlog grooming essentially involves the exercise of “real options”.
As Figure 3 shows, in plan driven projects, requirements are fixed while cost and duration estimates are subject to risk; in change driven projects, cost and time are fixed and scope is the variable.
Figure 3 — Iron Triangle Waterfall / Agile Paradigm Shift
Interestingly, the quality of return on investment (ROI) in waterfall projects (Tables 2 and 3) is fundamentally different than on agile projects (Tables 4 and 5). Adaptive projects have the potential to generate income with the first release as well as increase revenue in subsequent releases.
Table 2 — Payback Period Analysis of a Waterfall Project 
Note that the project beaks even in October, ten months after project start. Admittedly, payback period analysis suffers from several flaws. For instance, it ignores time value of money, financing, risk and other important considerations such as opportunity cost. Below is a DCF calculation:
Table 3 — DCF Calculations on a Waterfall Project
As the comparison above shows, projects evaluated using net present value (NPV) often provide a valuation that understates the fair market value of the asset. This is because projects may provide a low or zero cash flow in the near term but nonetheless be valuable to the firm over the long term. A more informed view is that projects are akin to owning an option to implement, abandon, shrink, delay, or expand a project should the opportunity costs outweigh the benefits.
To highlight the effect of generating income earlier within the lifecycle of a project, Tables 4 and 5 present an agile project that releases features at the end of January near the beginning of the project, and then follows up with new releases which add revenue until the project ends in June. The impact on results is material.
Note that the agile project breaks even in August, two months earlier than the waterfall project. In addition, the project generated 25 percent more revenue than the waterfall project.
Table 5 — DCF Calculations on an Agile Project
An alternative to the DCF approach for modeling cost-benefit decisions is real options analysis or real options valuation (ROA). Generating models, however, is only part of the usefulness of ROA.
The advantage of ROA is that it promotes thinking about the various conditions and scenarios that may evolve as aspects of a project’s uncertainty become known. As with agile projects, ROA frames alternatives and helps redirect efforts to maximize project value based on actual business dynamics. Thus, ROA promotes managerial flexibility over a project’s lifecycle.
Introduction to real options analysis
The concept of ROA is based on Merton’s (1973) definition of options as “simple contingent-claim assets” whereby the underlying asset can be practically anything. Organizations engage in projects for a multitude of reasons, not just for competitive advantage. Real options may be related to adverse impacts a problem is causing, or compliance with regulations. In addition, real options need not be analogous to a “call” option, but can also be employed as a “put” option to evaluate divestments.
There is a variety of uses for ROA as a tool for project management. Modeling techniques include the choice to choose between alternative solutions; abandon, shrink, delay, or expand a project; switch resources; or provide for phased or sequential investments. The incorporation of Monte Carlo simulation also provides a method for sensitivity/stress testing.
Project characteristics where use of ROA is most applicable are those that are expensive, long term, and affected by multiple risks (e.g., irreversible costs, market risk, regulatory risk, political risk). Mining of minerals, pharmaceutical industry, bio-technology, aeronautics, and energy production and transmission are all sectors where use of ROA has found adherents.
What makes ROA special compared to other valuation techniques is that it is the only approach that gives consideration to risk’s upside potential. To understand why this is the case, consider Figure 4, a simple example of equal probabilities involving up down movements. Given the larger potential loss, the expected value is negative, that is, expected value = 0.5(100) + 0.5(-120) = -10.
Figure 4 — One Stage Decision Tree
Now contrast the decision tree in Figure 4 with the two stage decision tree in Figure 5. Expected value = (2/3)(-10) + (1/3)(10 + (2/3)(90) + (1/3)(-110)) = 4.44. Note that the cumulative potential profits and potential losses are the same, as well as the fact that cumulative probabilities remain at 50%.
Figure 5 — Two Stage Decision Tree
What turns a potentially bad investment in the first branch into a good investment in the second is learning. Note that in Figure 5 one can make a decision based on observed results. If the outcome is negative, one can abandon the investment. The value of real options is from observing, then adapting behavior to increase upside potential and decrease downside potential.
In other words, “inspect and adapt,” a major theme in agile/scrum project methodology…
Not surprisingly, ROA views risk much differently than DCF. Whereas DCF discounts expected cash flows using a risk-adjusted discount rate, when using ROA the rate will vary depending upon which decision path is being analyzed. The reason why is because holders of options learn from observing and adapt behavior. Consequently, optimal decisions occur at each stage conditioned on outcomes from prior stages. In other words, cash flows will tend to increase because the option to abandon reduces downside risk; likewise, the option to expand will increase upside potential.
As an example, assume you are valuing an oil company based on estimated clash flows derived from multiplying the expected number of barrels produced in a year and the expected oil price per barrel. While you may have reasonable and unbiased estimates, what is missing is the interplay between these numbers and the behavior of the company. Oil companies can adjust production in response to changes in price; producing more oil when prices are high and less when prices are low. As a consequence, the company’s cash flow in response to these different scenarios will be greater than the expected cash flow calculated using traditional risk adjusted valuations such as DCF.
As Figure 6 illustrates, the sets of decisions that an oil company makes to produce, sell or store oil is actually a complex decision making process. For example, a positive forward price curve may be a factor in the decision to add to production, store the increase, and establish a carry trade even when prices are low.
Figure 6 — Decision Tree Involving Multiple Variables
To better estimate the value of complex projects as in Figure 6 where there are multiple variables such as market price, Monte Carlo simulation is employed. Key variables that drive values such as market prices are called critical success drivers. These drivers are prime candidates for Monte Carlo simulation. In addition, correlated simulations may provide closer approximation of variables’ real-life behaviors.
Regrettably, exploring the mechanics of incorporating Monte Carlo simulation into an integrated ROA process is beyond the scope of this article. We’ll leave that investigation for future discussion.
To conclude, we’ve made a case that ROA has similarities to agile project methodology and thinking, and that DCF’s deterministic approach is akin to prescriptive methods employed by traditional plan driven projects. What this implies is that agile is not just for software development projects. ROA extends change driven project framework to complex initiatives involving long term planning and multiple risks. With that in mind, ROA should not be thought of as simply about modeling, it is an entire decision-making process that complements project management/business analysis.
 Reference: IIBA BABOK Guide 2.0, Section 5.3 Determine Solution Approach and Section 9.8 Decision Analysis.
 “Plan driven projects” are also referred to in this article as “predictive method” or “waterfall” approach.
 Source: Jonathan Mun (2006)
 The WACC is the minimum return that a company must earn to satisfy its investors. The WACC is calculated taking into account the relative weights of each component of the capital structure. In a DCF model WACC is applied as a firm level discount/hurdle rate for all types of projects, and is subject to manipulation (e.g., wild guesses). It should be noted that NPV valuation is highly sensitive to WACC.
 To support this assertion, substitute the terms “cash flows” or “cash flow streams” with “requirements” or “dependencies”. Note: this exercise is not applicable to all of the comparisons, but sufficiently illustrates.
 Source: Jonathan Mun (2006)
 B. W. Boehm (1981). Software engineering economics. Englewood Cliffs, N.J: Prentice-Hall.
 Source: GAO Cost Estimating and Assessment Guide, Best Practices for Developing and Managing Capital Program Costs, GAO-09-3SP, March 2009, Pg 38.
 “Change driven projects” are also referred to in this article as the “adaptive method,” or “agile” or “scrum” approach.
 Source : Mark Layton (2012)
 Source : Mark Layton (2012)
 Twenty-five percent is calculated as the difference between cumulative FCF between the agile and waterfall projects ($930,000 - $700,000) divided into the cumulative FCF for the agile project ($930,000).
 See IIBA BABOK Guide 2.0, Section 5.1 Define Business Need.
 Jonathan Mun (2006)
 Copeland and Antikarov (2003) provide proof that the value of a risk asset will be the same under ROA and decision trees, if one allows for path-dependent discount rates.
International Institute of Business Analysis, (2009). A Guide to the Business Analysis Body of Knowledge (BABOK guide), version 2.0. Toronto, Ont: International Institute of Business Analysis.
J. Imai and M. Nakajima (2000). “A Real Option Analysis of an Oil Refinery Project” Financial Practice and Education 10(2), 78—91.
Jana Chvalkovská and Zdeněk Hrubý. “The Real Option Model – Evolution and Applications” Institute of Economic Studies, Faculty of Social Sciences, Charles University, Prague.
Johnathan Mun (2006). “Real Options Analysis versus Traditional DCF Valuation in Layman’s Terms” Real Options Valuation, Inc.
Mark Layton (2012). Agile project management for dummies. Hoboken, N.J: Wiley.
T. E. Copeland and V. Antikarov (2003). Real Options: A Practitioner’s Guide. New York, Texere.