Technical Terms Used in Project Portfolio Management (Continued)

A

B

C

D

E

F

G

H

I

J

K

L

M

N

O

P

Q

R

S

T

U

V

W

X

Y

Z

governance

The means by which an organization regulates and controls organizational behavior in accordance with its goals and objectives. A governance structure establishes accountability by implementing systems to monitor and record what people do, includes steps for ensuring compliance with policies, and provides for corrective action in cases where rules have been not been followed.

grid analysis

A decision-aiding technique that involves creating a table with options listed as rows the factors that need to be considered as columns. Each option/factor combination is then scored, the scores are weighted, and the results added to provide an overall score for the option. The approach, although simplistic, is used by some project portfolio management tools to rank projects.

group think

A dynamic of groups that promotes faulty decision making. Group think, a term coined by psychologist Irving Janis in the 1970s, has been widely studied. Consequences of group think include the tendency of groups to overlook alternatives, selectively collect information, fail to anticipate adverse consequences of choices, assume agreement among members when it does not exist, and fail to develop contingency plans. In addition, groups tend to possess a sense of invulnerability which promotes risk taking, stereotype the views of those outside the group, self-censor member views that go against group views, and believe in their inherent moral superiority over those outside the group.

GUI

Pronounced GOO-ee, GUI stands for graphic user interface. A GUI is the common method used for enabling humans to interact with a computer program. A GUI is graphic-based and relies at least in part on icons, pictures, and menus (which the user designates with a mouse or other pointing method) in addition to text (entered through a keyboard) to obtain user inputs.

H

hedonic price method

A method for inferring the value of something not traded directly in the market place based on market prices. The premise is that the prices of goods traded in the market, such as houses, depend on internal factors (e.g., house size, age, appearance) and external factors (distance to schools, level of local air and water pollution). Price data for the marketed good is analyzed (typically, via regression analysis) to determine the contribution of each factor to price, thereby inferring a monetary value for changes in that factor. The approach is often used to estimate the monetary value of environmental outcomes for cost benefit analysis.

hierarchical model

A model with a topdown, tree-like structure, such that each subsystem is linked to at most one "parent" subsystem. A hierarchical model is distinct from a network model, wherein each subsystem may have links to multiple parents.

human resource management

See resource management.

hurdle rate

A specified rate of return for a project or other investment intended to represent the minimum return that the organization will consider. The hurdle rate is also often referred to as the required rate of return.

Typically, the hurdle rate is the discount rate to be applied to the cash flows anticipated from the project. If the net present value (NPV) of cash flows using the hurdle rate as the discount rate is negative, the project is rejected. Alternatively, the hurdle rate may refer to the minimum acceptable internal rate of return (IRR) for projects—If a project's IRR is less than the hurdle rate, the project is rejected.

According to investment theory, the hurdle rate should be set equal to the rate of return that the organization could obtain by investing in alternative investment opportunities having similar risks. If the project generates a return greater than what the organization could earn elsewhere (i.e., greater than the opportunity cost of the required investment), the project will add value. Because the opportunity cost is difficult to compute, in practice, hurdle rates for projects are often specified by adding or subtracting a risk premium to the organization's marginal cost of capital, so that a higher rate is specified for project considered more risky.

Hurdle rates are generally a poor way to account for risks when prioritizing projects. As explained in one of the papers, hurdle rates tend to produce a bias toward short-term, quick payoff projects relative to projects of similar risks whose benefits accrue over longer periods of time. Also, while hurdle rates can be used to decrease the value of projects whose benefits are more uncertain, they do not achieve the converse effect of increasing the value of projects that, if not conducted or if delayed, would tend to increase risk.

I

influence diagram

An influence diagram is type of decision model that has a graphical representation composed of nodes denoting model variables connected by arrows. A rectangular node represents a decision variable, a choice that a decision maker can make among various alternatives. An elliptical node represents an uncertain variable, something relevant that is not directly under the decision-maker's control. Sometimes another shape (e.g., a rectangle with rounded corners, as shown below) is used to represent a performance measure computed from other model variables indicating the degree to which a decision objective is achieved. The arrows represent influences (not cause and effect). Specifically, an arrow from an ellipse (an uncertainty) or a rectangle (a decision) pointing to an ellipse (an uncertainty) means that the probabilities describing the possible outcomes for the latter uncertainty depend on the choice made or the outcome of the former variable. An arrow pointing from an ellipse or a rectangle to a rectangle means that the latter decision can be made after the outcome to the uncertainty is known or the prior choice is made.

A sample influence diagram (new drug project decision model)

The basic concepts underlying influence diagrams were developed at SRI International in 1973, by Allen Miller, to analyze US intelligence gathering strategies for the Persian Gulf. The diagrams allow complicated decision situations involving many variables to be represented by a compact graph. Because they are intuitive and can be constructed using a top-down model building approach (What variable do you want to know? What variables would you need to know in order to estimate what you want to know?) influence diagrams support an efficient model building process that can be conducted as a facilitated, team exercise. Also, software tools are available that allow influence diagrams to be linked to other models (e.g., to a spreadsheet model for computing business outcomes) and analyzed to calculate optimal decision strategies, risk, and the value of information. Along with decision trees, influence diagrams are the most commonly used model forms for decision analysis.

integer programming

Similar to linear programming—A method for minimizing or maximizing a linear objective function subject to linear constraints. The difference is that in the case of integer programming one or more of the decision variables must be integers. Unlike linear programming, which can be solved quickly and efficiently, finding the solution to an integer programming problem can be very difficult. The major algorithms for solving integer programming problems are branch and bound, the cutting plan method, and Bender's algorithm.

internal rate of return (IRR)

The effective interest rate that equates the present value of the income stream generated by a project to the cost of that project. In other words, the IRR is the value that satisfies the equation:

Equivalently, the IRR is the discount rate that causes the project to have a zero net present value (NPV).

The IRR is sometimes misunderstood to be the annual profitability of a project investment. However, for this to be true, the cash inflows derived from the project would have to be reinvested in opportunities that produced a return equal to the project's IRR, which is typically not the case.

Like NPV, the IRR is a commonly used criterion for project-by-project selection decisions—all projects that have IRR's greater than the cost of capital are recommended for funding.

When used as a ranking metric, IRR has an advantage over NPV, it does not depend on the size or scale of the project (therefore, its use is more consistent with the concept of ranking based on "bang-for-the buck"). However, at best, the IRR is only a heuristic or approximate logic for prioritizing projects. It creates predictable and significant biases in project rankings and cannot be used to identify project portfolios that create maximum value.

IRR analysis is popular because it compares projects based on the familiar concept of rate of return (a metric analogous to interest rates charged on capital markets). For this reason, it is for many one of the easiest project evaluation methods to understand. Given two investment alternatives of comparable costs, the investment with the higher IRR should be selected. When used as a selection rule for project-by-project decision making, IRR and NPV recommend the same projects (provided that a unique IRR exists, see below). The IRR has a perceived advantage over NPV and other methods that quantify project value in dollars in that it downplays reliance on what might appear to be overly-precise dollar values.

For several reasons, the IRR cannot be used to reliably prioritize projects. First, like other purely financial metrics, the IRR ignores the non-financial components of project value. Furthermore, IRR cannot be used to correctly prioritize based on financial return alone. Ranking projects based on IRR undervalues cash flows that occur late in a project's life (assuming that the IRR is greater than the cost of capital). It therefore creates a bias for projects with early positive returns relative to projects whose returns tend to occur later. The significance of this bias is greater the longer the duration of project cash flows and the more severely constrained the capital budget (if the budget is highly constrained then the IRR's of the projects being compared will tend to be significantly higher than the cost of capital, meaning that future cash flows will be very heavily discounted).

There are several additional disadvantages with IRR. There is no specific formula that can be used to calculate the IRR; it must be found by interpolation. Thus, computing the IRR takes more computation than does the NPV (most spreadsheets do, however, provide an IRR function). The IRR of the portfolio of projects cannot be calculated from the IRR's of the individual projects, so there is no easy way to quantify the performance of the portfolio based on the analysis of the individual projects that make up the portfolio. IRR analysis cannot be extended to account for consideration of project risks or project interdependencies. Finally, there are situations of multiple solutions (no unique IRR) when project cash flow changes direction more than once.

K

Kepner-Tregoe method

A structured decision aiding technique for collecting information and making decisions developed by Charles H. Kepner and Benjamin B. Tregoe in the 1960s. The approach consists of four basic steps, situation appraisal, problem analysis, option analysis, and potential problem and opportunity analysis.

knapsack problem

A mathematical statement of the problem of selecting projects subject to a budget constraint. The name derives from the analogy to the problem of choosing items to carry in a knapsack. To illustrate, suppose there are m items, where item i has a size c_i and, if selected, provides a benefit b_i, for i = 1, 2,,..., m. The capacity of the knapsack is C. The goal is to select items that will collectively fit in the knapsack and provide the greatest possible benefit. The problem may be expressed mathematically as:

The x_i are decision variables (1 for item acceptance and 0 for rejection). These are the same equations used to describe the project portfolio capital allocation problem.

The knapsack problem is relatively difficult to solve. It has, in fact, been used as the basis for encryption. Methods for solving the knapsack problem are computationally intensive, however, various approximate methods are available that are more efficient and that can be shown to come very close to mathematically optimal solution (see Methods for Solving the Capital Allocation Problem).

What makes the knapsack problem difficult is the "0/1 assumption"—Items must be put entirely in the knapsack or not included at all. You cannot, for example, put part of a soda pop can in the knapsack. Were it not for this requirement, you could solve the problem easily using a "greedy algorithm"—Rank the items based on benefit per unit size. Take as much of the top-ranked item as you can (or enough to fill the knapsack). Then, repeat with the next ranked item until the knapsack is full.

The knapsack problem arises in any activity, including project portfolio management (PPM), that requires allocating finite resources to items that are not infinitely divisible. Thus, a relevant consideration for choosing a PPM tool is whether it provides an algorithm for solving the knapsack problem and the quality of that algorithm.